Show HN: Program ESP32s in Nim
4 by elcritch | 1 comments on Hacker News.
Saturday, 30 September 2023
New Show Hacker News story: latest news
Show HN: A map that tells you if a NYC cafe has WiFi, a restroom, and an outlet
13 by asimova | 3 comments on Hacker News.
I am slowly adding more locations now. This is intended to be a crowdsourced map. Everyone is welcome to add more locations and provide comments/votes here. Free people from going to a cafe for work only to leave because there's no wifi, restroom, or outlet!! Demo: https://ift.tt/GgMBWib
13 by asimova | 3 comments on Hacker News.
I am slowly adding more locations now. This is intended to be a crowdsourced map. Everyone is welcome to add more locations and provide comments/votes here. Free people from going to a cafe for work only to leave because there's no wifi, restroom, or outlet!! Demo: https://ift.tt/GgMBWib
New Show Hacker News story: latest news
Show HN: Noisy Nest Free white/pink/brown noise generator
3 by kuczmama | 0 comments on Hacker News.
Hi All, i was playing around with GPT-4 today, and threw together a simple noise generator. It can play pink, brown, and white noise indefinitely. I spent maybe an hour creating this. It's quite amazing what you can do with such little time with AI. Let me know if you think of any features to add, and i'll consider adding them to make this more useful.
3 by kuczmama | 0 comments on Hacker News.
Hi All, i was playing around with GPT-4 today, and threw together a simple noise generator. It can play pink, brown, and white noise indefinitely. I spent maybe an hour creating this. It's quite amazing what you can do with such little time with AI. Let me know if you think of any features to add, and i'll consider adding them to make this more useful.
New Show Hacker News story: latest news
Show HN: Stargazers Reloaded – LLM-Powered Analyses of Your GitHub Community
4 by jarulraj | 1 comments on Hacker News.
Hey friends! We have built an app for getting insights about your favorite GitHub community using large language models. The app uses LLMs to analyze the GitHub profiles of users who have starred the repository, capturing key details like the topics they are interested in. It takes screenshots of the stargazer's GitHub webpage, extracts text using an OCR model, and extracts insights embedded in the extracted text using LLMs. This app is inspired by the “original” Stargazers app written by Spencer Kimball (CEO of CockroachDB). While the original app exclusively used the GitHub API, this LLM-powered app built using EvaDB additionally extracts insights from unstructured data obtained from the stargazers’ webpages. Our analysis of the fast-growing GPT4All community showed that the majority of the stargazers are proficient in Python and JavaScript, and 43% of them are interested in Web Development. Web developers love open-source LLMs! We found that directly using GPT-4 to generate the “golden” table is super expensive — costing $60 to process the information of 1000 stargazers. To maintain accuracy while also reducing cost, we set up an LLM model cascade in a SQL query, running GPT-3.5 before GPT-4, that lowers the cost to $5.5 for analyzing 1000 GitHub stargazers. We’ve been working on this app for a month now and are excited to open source it today :) Some useful links: * Blog Post - https://ift.tt/w0JcrRp... * GitHub Repository - https://ift.tt/xb7EJpw * EvaDB - https://ift.tt/EU8GKBb Please let us know what you think!
4 by jarulraj | 1 comments on Hacker News.
Hey friends! We have built an app for getting insights about your favorite GitHub community using large language models. The app uses LLMs to analyze the GitHub profiles of users who have starred the repository, capturing key details like the topics they are interested in. It takes screenshots of the stargazer's GitHub webpage, extracts text using an OCR model, and extracts insights embedded in the extracted text using LLMs. This app is inspired by the “original” Stargazers app written by Spencer Kimball (CEO of CockroachDB). While the original app exclusively used the GitHub API, this LLM-powered app built using EvaDB additionally extracts insights from unstructured data obtained from the stargazers’ webpages. Our analysis of the fast-growing GPT4All community showed that the majority of the stargazers are proficient in Python and JavaScript, and 43% of them are interested in Web Development. Web developers love open-source LLMs! We found that directly using GPT-4 to generate the “golden” table is super expensive — costing $60 to process the information of 1000 stargazers. To maintain accuracy while also reducing cost, we set up an LLM model cascade in a SQL query, running GPT-3.5 before GPT-4, that lowers the cost to $5.5 for analyzing 1000 GitHub stargazers. We’ve been working on this app for a month now and are excited to open source it today :) Some useful links: * Blog Post - https://ift.tt/w0JcrRp... * GitHub Repository - https://ift.tt/xb7EJpw * EvaDB - https://ift.tt/EU8GKBb Please let us know what you think!
New Show Hacker News story: latest news
Show HN: A tool to help optimise your system prompt
2 by rcushen | 0 comments on Hacker News.
I needed a way to choose between different system prompts for a ChatGPT-powered side project, so I built Prompt Picker. The basic idea is: specify some candidate system prompts and a few example user inputs, and then run a blind experiment where you choose the better output from pairs of generations. If you need to optimise the system prompt for your LLM-powered application, give it a try and let me know what you think.
2 by rcushen | 0 comments on Hacker News.
I needed a way to choose between different system prompts for a ChatGPT-powered side project, so I built Prompt Picker. The basic idea is: specify some candidate system prompts and a few example user inputs, and then run a blind experiment where you choose the better output from pairs of generations. If you need to optimise the system prompt for your LLM-powered application, give it a try and let me know what you think.
Friday, 29 September 2023
New Show Hacker News story: latest news
Show HN: Insomnium – 100% local and privacy-focus fork of Insomnia API client
13 by archibaldJ | 1 comments on Hacker News.
13 by archibaldJ | 1 comments on Hacker News.
New Show Hacker News story: latest news
Show HN: Botanist, a composer plugin that fixes file permissions in your repo
5 by hchinchilla | 1 comments on Hacker News.
5 by hchinchilla | 1 comments on Hacker News.
New Show Hacker News story: latest news
Show HN: Open-Source Free App for Talking with AI Using Audio+Text
2 by reluct35-MATT80 | 0 comments on Hacker News.
Open-Source Free App for Talking with AI Using Audio+Text
2 by reluct35-MATT80 | 0 comments on Hacker News.
Open-Source Free App for Talking with AI Using Audio+Text
New Show Hacker News story: latest news
Show HN: Config-file-validator – CLI tool to validate all your config files
2 by treebeard5440 | 1 comments on Hacker News.
2 by treebeard5440 | 1 comments on Hacker News.
Thursday, 28 September 2023
New Show Hacker News story: latest news
Show HN: Generative Fill with AI and 3D
31 by olokobayusuf | 4 comments on Hacker News.
Hey all, You've probably seen projects that add objects to an image from a style or text prompt, like InteriorAI (levelsio) and Adobe Firefly. The prevalent issue with these diffusion-based inpainting approaches is that they don't yet have great conditioning on lighting, perspective, and structure. You'll often get incorrect or generic shadows; warped-looking objects; and distorted backgrounds. What is Fill 3D? Fill 3D is an exploration on doing generative fill in 3D to render ultra-realistic results that harmonize with the background image, using industry-standard path tracing, akin to compositing in Hollywood movies. How does it work? 1. Deproject: First, deproject an image to a 3D shell using both geometric and photometric cues from the input image. 2. Place: Draw rectangles and describe what you want in them, akin to Photoshop's Generative Fill feature. 3. Render: Use good ol' path tracing to render ultra-realistic results. Why Fill 3D? + The results are insanely realistic (see video in the github repo, or on the website). + Fast enough: Currently, generations take 40-80 seconds. Diffusion takes ~10seconds, so we're slower, but for the level of realism, it's pretty good. + Potential applications: I'm thinking of virtual staging in real estate media, what do you think? Check it out at https://fill3d.ai + There's API access! :D + Right now, you need an image of an empty room. Will loosen this restriction over time. Fill 3D is built on Function ( https://fxn.ai ). With Function, I can run the Python functions that do the steps above on powerful GPUs with only code (no Dockerfile, YAML, k8s, etc), and invoke them from just about anywhere. I'm the founder of fxn. Tell me what you think!! PS: This is my first Show HN, so please be nice :)
31 by olokobayusuf | 4 comments on Hacker News.
Hey all, You've probably seen projects that add objects to an image from a style or text prompt, like InteriorAI (levelsio) and Adobe Firefly. The prevalent issue with these diffusion-based inpainting approaches is that they don't yet have great conditioning on lighting, perspective, and structure. You'll often get incorrect or generic shadows; warped-looking objects; and distorted backgrounds. What is Fill 3D? Fill 3D is an exploration on doing generative fill in 3D to render ultra-realistic results that harmonize with the background image, using industry-standard path tracing, akin to compositing in Hollywood movies. How does it work? 1. Deproject: First, deproject an image to a 3D shell using both geometric and photometric cues from the input image. 2. Place: Draw rectangles and describe what you want in them, akin to Photoshop's Generative Fill feature. 3. Render: Use good ol' path tracing to render ultra-realistic results. Why Fill 3D? + The results are insanely realistic (see video in the github repo, or on the website). + Fast enough: Currently, generations take 40-80 seconds. Diffusion takes ~10seconds, so we're slower, but for the level of realism, it's pretty good. + Potential applications: I'm thinking of virtual staging in real estate media, what do you think? Check it out at https://fill3d.ai + There's API access! :D + Right now, you need an image of an empty room. Will loosen this restriction over time. Fill 3D is built on Function ( https://fxn.ai ). With Function, I can run the Python functions that do the steps above on powerful GPUs with only code (no Dockerfile, YAML, k8s, etc), and invoke them from just about anywhere. I'm the founder of fxn. Tell me what you think!! PS: This is my first Show HN, so please be nice :)
New Show Hacker News story: latest news
Show HN: Semiform.ai – A new kind of web form, powered by LLMs
2 by npilk | 0 comments on Hacker News.
2 by npilk | 0 comments on Hacker News.
New Show Hacker News story: latest news
Show HN: Abuse inflight WiFi APIs to track your flight
83 by notmysql_ | 14 comments on Hacker News.
83 by notmysql_ | 14 comments on Hacker News.
New Show Hacker News story: latest news
Show HN: Launching a new home-grown embedding LLM for RAG
5 by eskibars | 0 comments on Hacker News.
Hi HN! Vectara is a "batteries included" retrieval augmented generation platform. You can upload your rich text documents like PDFs, HTML pages, word docs, etc, or semi-structured JSON and Vectara handles the text and metadata extraction, segmentation, vector embedding, and vector storage, and keyword storage. You can ask a question or perform a search in the UI or via our APIs and Vectara will automatically handle the vectorization, structured metadata filtering, vector+keyword retrieval, hybrid blending, and generative summarization of the results. We're focusing on building and operationalizing the complex infrastructure for vector storage, hybrid retrieval, and generative summarization so you can use fairly high-level APIs and focus on building your own applications. We know that retrieval accuracy is incredibly important for RAG: garbage in, garbage out. We've seen a lot of projects not spend enough time on really getting the retrieval model right and wasting a lot of time/money with poor outcomes. We've spent about the past 6 months working on a new embedding model named Boomerang and just released it on the Vectara platform. We've run it through standard evaluations like BEIR (though we know many models over-fit against BEIR) as well as multi-domain evaluations. We've published the details of our tests for those that really want to dive in, but the TL;DR is that Boomerang beats most/all publicly available models in many/most situations and is particularly strong at cross-lingual and multi-domain tests. We'd love any and all feedback!
5 by eskibars | 0 comments on Hacker News.
Hi HN! Vectara is a "batteries included" retrieval augmented generation platform. You can upload your rich text documents like PDFs, HTML pages, word docs, etc, or semi-structured JSON and Vectara handles the text and metadata extraction, segmentation, vector embedding, and vector storage, and keyword storage. You can ask a question or perform a search in the UI or via our APIs and Vectara will automatically handle the vectorization, structured metadata filtering, vector+keyword retrieval, hybrid blending, and generative summarization of the results. We're focusing on building and operationalizing the complex infrastructure for vector storage, hybrid retrieval, and generative summarization so you can use fairly high-level APIs and focus on building your own applications. We know that retrieval accuracy is incredibly important for RAG: garbage in, garbage out. We've seen a lot of projects not spend enough time on really getting the retrieval model right and wasting a lot of time/money with poor outcomes. We've spent about the past 6 months working on a new embedding model named Boomerang and just released it on the Vectara platform. We've run it through standard evaluations like BEIR (though we know many models over-fit against BEIR) as well as multi-domain evaluations. We've published the details of our tests for those that really want to dive in, but the TL;DR is that Boomerang beats most/all publicly available models in many/most situations and is particularly strong at cross-lingual and multi-domain tests. We'd love any and all feedback!
New Show Hacker News story: latest news
Show HN: UI Shots – Speed up your design research process
2 by mubeen138 | 0 comments on Hacker News.
UI Shots help designers explore 1000+ pages from 100+ SaaS apps. This helps them getting inspiration for their next design. Websites are divided into categories like Finance, CRM, Email, Website Builder, AI, Automation and so on, with a total of 18 categories (more coming sooner) Same goes for pages, they are divided into page types e.g. Landing Page, Pricing Page and so on. Users can also filter by colors and can check the fonts used on a page. Further, they can bookmark their favorite designs to have all the inspiration at one place. Few days ago, I launched it on Product Hunt, 50 users have signed up already, with 1 premium subscription. I am looking for feedback to help improve my app. Thank You. Here is the link: https://ift.tt/3KGT1Xn
2 by mubeen138 | 0 comments on Hacker News.
UI Shots help designers explore 1000+ pages from 100+ SaaS apps. This helps them getting inspiration for their next design. Websites are divided into categories like Finance, CRM, Email, Website Builder, AI, Automation and so on, with a total of 18 categories (more coming sooner) Same goes for pages, they are divided into page types e.g. Landing Page, Pricing Page and so on. Users can also filter by colors and can check the fonts used on a page. Further, they can bookmark their favorite designs to have all the inspiration at one place. Few days ago, I launched it on Product Hunt, 50 users have signed up already, with 1 premium subscription. I am looking for feedback to help improve my app. Thank You. Here is the link: https://ift.tt/3KGT1Xn
New Show Hacker News story: latest news
Show HN: Diarupt AI – Converse with AI over video in realtime
2 by marvinified | 0 comments on Hacker News.
Diarupt AI is a platform that makes it possible to integrate real-time video AI conversations into your product. With Diarupt AI, you can build: Dynamic video and audio-based AI interactions, with human-like avatars and realistic voices, that can have long and fluid conversations, exposed through easy-to-use APIs & SDKs so you can integrate in minutes. At Diarupt, we believe user experience is an essential part of any technology, and the same is true for AI. We aim to take our everyday experience and interaction with AI to the next level by making it seamless and as natural as possible. Diarupt is in Public Alpha, and we will be looking forward to your feedback as we build the future of AI Interactions together. We’ll be granting access in batches over the coming weeks, so join the waitlist and watch out for your invite. We are so excited to see what you’ll build. Cheers Useful links: - Website: https://diarupt.ai - Demo: https://youtu.be/eROf7SCdqcQ?si=B2nyqxBeTssGN-3E - Quickstart: https://ift.tt/xVpWrdn
2 by marvinified | 0 comments on Hacker News.
Diarupt AI is a platform that makes it possible to integrate real-time video AI conversations into your product. With Diarupt AI, you can build: Dynamic video and audio-based AI interactions, with human-like avatars and realistic voices, that can have long and fluid conversations, exposed through easy-to-use APIs & SDKs so you can integrate in minutes. At Diarupt, we believe user experience is an essential part of any technology, and the same is true for AI. We aim to take our everyday experience and interaction with AI to the next level by making it seamless and as natural as possible. Diarupt is in Public Alpha, and we will be looking forward to your feedback as we build the future of AI Interactions together. We’ll be granting access in batches over the coming weeks, so join the waitlist and watch out for your invite. We are so excited to see what you’ll build. Cheers Useful links: - Website: https://diarupt.ai - Demo: https://youtu.be/eROf7SCdqcQ?si=B2nyqxBeTssGN-3E - Quickstart: https://ift.tt/xVpWrdn
Wednesday, 27 September 2023
New Show Hacker News story: latest news
Show HN: The Tomb of Ramesses I in the Valley of the Kings
15 by lukehollis | 5 comments on Hacker News.
Hey all, I’m coding up a new tour system for my 3d Egyptian work. After the last feedback, I focused on building more interactive content and fx into the guided tours with sound and telling the mythology in the wall art. I’d love feedback with the new version – this is built with vanilla Three.js and footage captured on my iPhone 12. For various fx, I coded many of my own shaders based on work by https://twitter.com/akella and others on ShaderToy, so I’m keen to test on more devices. As the hacker, so the (ancient) painter.
15 by lukehollis | 5 comments on Hacker News.
Hey all, I’m coding up a new tour system for my 3d Egyptian work. After the last feedback, I focused on building more interactive content and fx into the guided tours with sound and telling the mythology in the wall art. I’d love feedback with the new version – this is built with vanilla Three.js and footage captured on my iPhone 12. For various fx, I coded many of my own shaders based on work by https://twitter.com/akella and others on ShaderToy, so I’m keen to test on more devices. As the hacker, so the (ancient) painter.
New Show Hacker News story: latest news
Show HN: A simple API/CLI for scheduling HTTP requests
3 by areichert | 0 comments on Hacker News.
Hi HN! This is something I've been tinkering on for the past couple months. It's basically just an API/CLI for scheduling delayed or recurring jobs as HTTP requests. I initially built it as a personal tool to save myself a bit of time on little side projects where I've needed scheduled/recurring alerts, but decided it could be a good opportunity to practice building out a nice landing page [0] and documentation [1]. And who knows, maybe someone else will find it useful ¯\_(ツ)_/¯ The tool relies heavily on Elixir's Oban [2] library for managing jobs, and Mintlify [3] for documentation. I also shamelessly stole most of the frontend design from Resend [4] because I'm a fan of the aesthetic and thought it would be good for my design chops to use their design as a guide. I also discovered Radix [5] UI while working on this, which ended up being immensely helpful for moving quickly on the frontend. Anyways, I almost certainly spent a bit too much time on small UX details that are most likely utterly inconsequential, but it was a fun exercise in polish :) All feedback is welcome! [0] https://www.booper.dev/ [1] https://ift.tt/9l7n3WC [2] https://ift.tt/ha4I8cR [3] https://mintlify.com/ [4] https://resend.com/ [5] https://ift.tt/ic24gCq
3 by areichert | 0 comments on Hacker News.
Hi HN! This is something I've been tinkering on for the past couple months. It's basically just an API/CLI for scheduling delayed or recurring jobs as HTTP requests. I initially built it as a personal tool to save myself a bit of time on little side projects where I've needed scheduled/recurring alerts, but decided it could be a good opportunity to practice building out a nice landing page [0] and documentation [1]. And who knows, maybe someone else will find it useful ¯\_(ツ)_/¯ The tool relies heavily on Elixir's Oban [2] library for managing jobs, and Mintlify [3] for documentation. I also shamelessly stole most of the frontend design from Resend [4] because I'm a fan of the aesthetic and thought it would be good for my design chops to use their design as a guide. I also discovered Radix [5] UI while working on this, which ended up being immensely helpful for moving quickly on the frontend. Anyways, I almost certainly spent a bit too much time on small UX details that are most likely utterly inconsequential, but it was a fun exercise in polish :) All feedback is welcome! [0] https://www.booper.dev/ [1] https://ift.tt/9l7n3WC [2] https://ift.tt/ha4I8cR [3] https://mintlify.com/ [4] https://resend.com/ [5] https://ift.tt/ic24gCq
New Show Hacker News story: latest news
Show HN: Using LLMs and Embeddings to classify application errors
12 by vadman97 | 0 comments on Hacker News.
Hi Hacker News! We’re Vadim and Chris from Highlight.io [1]. We do web app monitoring and are working on using LLMs/embeddings to add new functionality to our error monitoring product. Given that there’s a lot of founders/engineers using LLMs in their products, we figured we’d share how we built the new functionality, their impact on our workflows, and how you can try it out. Our goal was to build two features: (1) tagging errors (e.g. deeming an error as “authentication error” or a “database error”); and (2) grouping similar errors together (e.g. two errors that have a different stacktrace and body, but are semantically not very different). Each of these rely heavily on comparing text across our application. After some experimentation with the OpenAI embeddings API [3], we went ahead and hosted a private model instance of thenlper/gte-large (an open-source MIT licensed model), which is a 1024-dimension model running on an Intel Ice Lake 2 vCPU machine on Hugging face [4]. Our general approach for classifying/comparing text is as follows. As each set of tokens (i.e a string) comes in, our backend makes a request to an inference endpoint and receives a 1024-dimension float vector as a response (see the code here [5]). We then store that vector using pgvector [6]. To compare any two sets for similarity, we simply look at the Euclidian distance between their respective embeddings using the ivfflat index implemented by pgvector (example code here [7]). To tag errors, we assign an error its most relevant tag from a predetermined set decided by us. For example, if we tag an error as an "authentication error" or a "database error", we can allow developers to have a starting point before inspecting an issue.(see the logic here [8]). Anecdotally, this approach seems to work very well. For example, here are two authentication errors that got tagged as “Authentication Error”: * Firebase: A network AuthError has occurred * Error retrieving user from firebase api for email verification: cannot find user from uid. We also use these error embeddings to group similar errors. To decide whether an error joins a group or starts a new one, we decide on a distance threshold (using the euclidean distance) ahead of time. An interesting thing about this approach, compared to using a text-based heuristic, is that two errors with different stack traces can still be grouped together. Here’s an example: * github.com/highlight-run/highlight/backend/worker.(*Worker).ReportStripeUsage * github.com/highlight-run/highlight/backend/private-graph/graph.(*Resolver).GetSlackChannelsFromSlack.func1 Both reported as `integration api error` as they involve the Stripe and Slack integrations respectively. The neat thing is that the LLM can use the full context of an error and match based on the most relevant details about the error. We have rolled out a first version of the error grouping logic to our cloud product [9], and there’s a demo of all the functionality at [2]. Long-term, if the HN community has other ideas of what we could build with LLM tooling in observability, we’re all ears. Let us know what you think! Links [1] https://ift.tt/duflCLT [2] https://ift.tt/8dFuyEU [3] https://ift.tt/xXtHm30 [4] https://ift.tt/Wc0wfN7 [5] https://ift.tt/KA3mLu7... [6] https://ift.tt/Vq3oGIx... [7] https://ift.tt/aNwRVo8... [8] https://ift.tt/VEUhYDQ... [9] https://ift.tt/6Xzm2JI
12 by vadman97 | 0 comments on Hacker News.
Hi Hacker News! We’re Vadim and Chris from Highlight.io [1]. We do web app monitoring and are working on using LLMs/embeddings to add new functionality to our error monitoring product. Given that there’s a lot of founders/engineers using LLMs in their products, we figured we’d share how we built the new functionality, their impact on our workflows, and how you can try it out. Our goal was to build two features: (1) tagging errors (e.g. deeming an error as “authentication error” or a “database error”); and (2) grouping similar errors together (e.g. two errors that have a different stacktrace and body, but are semantically not very different). Each of these rely heavily on comparing text across our application. After some experimentation with the OpenAI embeddings API [3], we went ahead and hosted a private model instance of thenlper/gte-large (an open-source MIT licensed model), which is a 1024-dimension model running on an Intel Ice Lake 2 vCPU machine on Hugging face [4]. Our general approach for classifying/comparing text is as follows. As each set of tokens (i.e a string) comes in, our backend makes a request to an inference endpoint and receives a 1024-dimension float vector as a response (see the code here [5]). We then store that vector using pgvector [6]. To compare any two sets for similarity, we simply look at the Euclidian distance between their respective embeddings using the ivfflat index implemented by pgvector (example code here [7]). To tag errors, we assign an error its most relevant tag from a predetermined set decided by us. For example, if we tag an error as an "authentication error" or a "database error", we can allow developers to have a starting point before inspecting an issue.(see the logic here [8]). Anecdotally, this approach seems to work very well. For example, here are two authentication errors that got tagged as “Authentication Error”: * Firebase: A network AuthError has occurred * Error retrieving user from firebase api for email verification: cannot find user from uid. We also use these error embeddings to group similar errors. To decide whether an error joins a group or starts a new one, we decide on a distance threshold (using the euclidean distance) ahead of time. An interesting thing about this approach, compared to using a text-based heuristic, is that two errors with different stack traces can still be grouped together. Here’s an example: * github.com/highlight-run/highlight/backend/worker.(*Worker).ReportStripeUsage * github.com/highlight-run/highlight/backend/private-graph/graph.(*Resolver).GetSlackChannelsFromSlack.func1 Both reported as `integration api error` as they involve the Stripe and Slack integrations respectively. The neat thing is that the LLM can use the full context of an error and match based on the most relevant details about the error. We have rolled out a first version of the error grouping logic to our cloud product [9], and there’s a demo of all the functionality at [2]. Long-term, if the HN community has other ideas of what we could build with LLM tooling in observability, we’re all ears. Let us know what you think! Links [1] https://ift.tt/duflCLT [2] https://ift.tt/8dFuyEU [3] https://ift.tt/xXtHm30 [4] https://ift.tt/Wc0wfN7 [5] https://ift.tt/KA3mLu7... [6] https://ift.tt/Vq3oGIx... [7] https://ift.tt/aNwRVo8... [8] https://ift.tt/VEUhYDQ... [9] https://ift.tt/6Xzm2JI
Tuesday, 26 September 2023
New Show Hacker News story: latest news
Show HN: Magentic – Use LLMs as simple Python functions
20 by jackmpcollins | 4 comments on Hacker News.
This is a Python package that allows you to write function signatures to define LLM queries. This makes it easy to mix regular code with calls to LLMs, which enables you to use the LLM for its creativity and reasoning while also enforcing structure/logic as necessary. LLM output is parsed for you according to the return type annotation of the function, including complex return types such as streaming an array of structured objects. I built this to show that we can think about using LLMs more fluidly than just chains and chats, i.e. more interchangeably with regular code, and to make it easy to do that. Please let me know what you think! Contributions welcome. https://ift.tt/Gh9zvOS
20 by jackmpcollins | 4 comments on Hacker News.
This is a Python package that allows you to write function signatures to define LLM queries. This makes it easy to mix regular code with calls to LLMs, which enables you to use the LLM for its creativity and reasoning while also enforcing structure/logic as necessary. LLM output is parsed for you according to the return type annotation of the function, including complex return types such as streaming an array of structured objects. I built this to show that we can think about using LLMs more fluidly than just chains and chats, i.e. more interchangeably with regular code, and to make it easy to do that. Please let me know what you think! Contributions welcome. https://ift.tt/Gh9zvOS
New Show Hacker News story: latest news
Show HN: Ollama for Linux – Run LLMs on Linux with GPU Acceleration
49 by jmorgan | 9 comments on Hacker News.
Hi HN, Over the last few months I've been working with some folks on a tool named Ollama ( https://ift.tt/tqNLpTC ) to run open-source LLMs like Llama 2, Code Llama and Falcon locally, starting with macOS. The biggest ask since then has been "how can I run Ollama on Linux?" with GPU support out of the box. Setting up and configuring CUDA and then compiling and running llama.cpp (which is a fantastic library and runs under the hood) can be quite painful on different combinations of linux distributions and Nvidia GPUs. The goal for Ollama's linux version was to automate this process to make it easy to get up and running. The is the first Linux release! There's still lots to do, but I wanted to share it here for to see what everyone thinks. Thanks for anyone who has given it a try and sent feedback!
49 by jmorgan | 9 comments on Hacker News.
Hi HN, Over the last few months I've been working with some folks on a tool named Ollama ( https://ift.tt/tqNLpTC ) to run open-source LLMs like Llama 2, Code Llama and Falcon locally, starting with macOS. The biggest ask since then has been "how can I run Ollama on Linux?" with GPU support out of the box. Setting up and configuring CUDA and then compiling and running llama.cpp (which is a fantastic library and runs under the hood) can be quite painful on different combinations of linux distributions and Nvidia GPUs. The goal for Ollama's linux version was to automate this process to make it easy to get up and running. The is the first Linux release! There's still lots to do, but I wanted to share it here for to see what everyone thinks. Thanks for anyone who has given it a try and sent feedback!
New Show Hacker News story: latest news
Show HN: MyGPT a toy LLM which can be trained on Project Gutenberg and dad jokes
2 by disconnection | 0 comments on Hacker News.
My puny version of ChatGPT. This was based on the excellent LLM lecture series by Andrej Karpathy: https://www.youtube.com/watch?v=kCc8FmEb1nY The main points of differentiation are that my version is token-based (tiktoken) with code to load up multiple text files as a trining set. Plus, it has a minimal server which is a drop-in replacement for the OpenAI REST API. So you can train the default tiny 15M parameter model, and use that in your projects instead of ChatGPT. I trained it on 20Mb of Project Gutenberg encyclopaedias, then fine-tuned it on 120 dad jokes, to get a Q: A: prompt format. This model + training set is so small that the results are basically a joke; it's for entertainment purposes only. The code is also very rough, and the server only has the minimum functionality filled in. I embodied this model in my talking LLM-driven hexapod robot, and it could give very silly answers to spoken questions.
2 by disconnection | 0 comments on Hacker News.
My puny version of ChatGPT. This was based on the excellent LLM lecture series by Andrej Karpathy: https://www.youtube.com/watch?v=kCc8FmEb1nY The main points of differentiation are that my version is token-based (tiktoken) with code to load up multiple text files as a trining set. Plus, it has a minimal server which is a drop-in replacement for the OpenAI REST API. So you can train the default tiny 15M parameter model, and use that in your projects instead of ChatGPT. I trained it on 20Mb of Project Gutenberg encyclopaedias, then fine-tuned it on 120 dad jokes, to get a Q: A: prompt format. This model + training set is so small that the results are basically a joke; it's for entertainment purposes only. The code is also very rough, and the server only has the minimum functionality filled in. I embodied this model in my talking LLM-driven hexapod robot, and it could give very silly answers to spoken questions.
New Show Hacker News story: latest news
Show HN: Chat with your favourite celebrity – AI Project
2 by ankitnair06 | 2 comments on Hacker News.
2 by ankitnair06 | 2 comments on Hacker News.
New Show Hacker News story: latest news
Show HN: I recorded nature sounds for a year, then released an ambient sound app
2 by disconnection | 0 comments on Hacker News.
There are lots of ambient sound apps on the App Store, but they all seem to be paywalled. I thought I should make one where all the content was free. I bought a secondhand professional audio recorder, then took it to national parks in Australia and Germany, in case I heard anything interesting. It was a bit of a learning experience, but I now have a library of half a dozen passable recordings. The biggest problem was noise pollution: humans are noisy and everywhere; I never noticed until I started listening carefully for background noise. Traffic and building noise carries for huge distances. I wrapped all the sounds up in a swiftUI interface, built around a curved video carousel. I wanted to pay special attention to vision impaired users, so it all has VoiceOver support, and extra UI elements that can be turned on for vision impaired users. It's free on the App Store if you want to take a look.
2 by disconnection | 0 comments on Hacker News.
There are lots of ambient sound apps on the App Store, but they all seem to be paywalled. I thought I should make one where all the content was free. I bought a secondhand professional audio recorder, then took it to national parks in Australia and Germany, in case I heard anything interesting. It was a bit of a learning experience, but I now have a library of half a dozen passable recordings. The biggest problem was noise pollution: humans are noisy and everywhere; I never noticed until I started listening carefully for background noise. Traffic and building noise carries for huge distances. I wrapped all the sounds up in a swiftUI interface, built around a curved video carousel. I wanted to pay special attention to vision impaired users, so it all has VoiceOver support, and extra UI elements that can be turned on for vision impaired users. It's free on the App Store if you want to take a look.
New Show Hacker News story: latest news
Show HN: Adding dynamic library loading to my pet programming language
2 by nbittich | 0 comments on Hacker News.
2 by nbittich | 0 comments on Hacker News.
Monday, 25 September 2023
New Show Hacker News story: latest news
Show HN: Zero-dependency Java framework out of beta
2 by 7ep | 0 comments on Hacker News.
I am happy to announce my minimalist zero-dependency web framework, Minum, is out of beta. https://ift.tt/2pFaQwr You will be hard-pressed to find another modern project as obsessively minimalistic. Other frameworks will claim simplicity and minimalism and then, casually, mention they are built on a multitude of libraries. This follows self-imposed constraints, predicated on a belief that smaller and lighter is long-term better. Caveat emptor: This is a project by and for developers who know and like programming (rather than, let us say, configuring). It is written in Java, and presumes familiarity with the HTTP/HTML paradigm. Driving paradigms of this project: * ease of use * maintainability / sustainability * simplicity * performance * good documentation * good testing It requires Java 21, for its virtual threads (Project Loom)
2 by 7ep | 0 comments on Hacker News.
I am happy to announce my minimalist zero-dependency web framework, Minum, is out of beta. https://ift.tt/2pFaQwr You will be hard-pressed to find another modern project as obsessively minimalistic. Other frameworks will claim simplicity and minimalism and then, casually, mention they are built on a multitude of libraries. This follows self-imposed constraints, predicated on a belief that smaller and lighter is long-term better. Caveat emptor: This is a project by and for developers who know and like programming (rather than, let us say, configuring). It is written in Java, and presumes familiarity with the HTTP/HTML paradigm. Driving paradigms of this project: * ease of use * maintainability / sustainability * simplicity * performance * good documentation * good testing It requires Java 21, for its virtual threads (Project Loom)
New Show Hacker News story: latest news
Show HN: An Admin-Friendly, User Management Server (With Passkeys and JWTs)
4 by okhuman | 0 comments on Hacker News.
4 by okhuman | 0 comments on Hacker News.
Sunday, 24 September 2023
New Show Hacker News story: latest news
Show HN: Bots Playing Online Cash Bingo on an iPad
3 by admtal | 0 comments on Hacker News.
Hey HN, I really enjoy playing "Bing King" from time to time. I had a little downtime, and thought it would be fun to try and write a bot to play it. Sharing it here: https://ift.tt/JO0tBIS Demo videos in the repo
3 by admtal | 0 comments on Hacker News.
Hey HN, I really enjoy playing "Bing King" from time to time. I had a little downtime, and thought it would be fun to try and write a bot to play it. Sharing it here: https://ift.tt/JO0tBIS Demo videos in the repo
New Show Hacker News story: latest news
Show HN: Hichatbot.ai – Chat with your documents, video transcripts
6 by ngvan | 0 comments on Hacker News.
Hi HN, I created this site recently and need som feedbacks. Please give it a try, let me know if you see anything to need to be improved or added. Below is the summary of https://hichatbot.ai What is HiChatbot? - HiChatbot is an AI-powered chatbot that can answer your questions about documents, text, and videos. You can upload a document, text, or provide a video link to HiChatbot and have a Q&A chat with the content. What are some of the things I can do with HiChatbot? You can use HiChatbot to: - Summarize the main points of a document - Get more details about a specific topic in a document, text, or video - Based on the provided document, generate creative text formats of text content, like executive summaries, reports, poems, code, scripts, musical pieces, email, letters, etc.
6 by ngvan | 0 comments on Hacker News.
Hi HN, I created this site recently and need som feedbacks. Please give it a try, let me know if you see anything to need to be improved or added. Below is the summary of https://hichatbot.ai What is HiChatbot? - HiChatbot is an AI-powered chatbot that can answer your questions about documents, text, and videos. You can upload a document, text, or provide a video link to HiChatbot and have a Q&A chat with the content. What are some of the things I can do with HiChatbot? You can use HiChatbot to: - Summarize the main points of a document - Get more details about a specific topic in a document, text, or video - Based on the provided document, generate creative text formats of text content, like executive summaries, reports, poems, code, scripts, musical pieces, email, letters, etc.
New Show Hacker News story: latest news
Show HN: pzip- blazing fast concurrent zip archiver and extractor
2 by exposition | 0 comments on Hacker News.
2 by exposition | 0 comments on Hacker News.
New Show Hacker News story: latest news
Show HN: Krestomatio – managed e-learning platforms in public beta
2 by jobcespedes | 0 comments on Hacker News.
Public beta available for our managed service that takes care of all the technical aspects of hosting open source Moodle™ e-learning platforms
2 by jobcespedes | 0 comments on Hacker News.
Public beta available for our managed service that takes care of all the technical aspects of hosting open source Moodle™ e-learning platforms
Saturday, 23 September 2023
New Show Hacker News story: latest news
Show HN: Basketball Play Designer
4 by kmcgraw | 0 comments on Hacker News.
In a company hackathon this weekend, I thought it would be fun to build a basketball play designer. Essentially, you enter some information about the type of play you want, and it generates a play following these requirements. I was imagining it might be useful for stuff like: 1. Designing personalized plays for first-time coaches: "Give me a play that gets my shooting guard an open three-pointer. My shooting guard shoots the best from the right wing, and my point guard has a weak left hand. My center sets great screens, so have my center set an off-ball screen for my shooting guard." 2. Brainstorming novel plays for long-time coaches: "Generate a novel play out of the UCLA set." It is loaded in with an example play, so make sure to hit the buttons to generate new plays with the entered information. The animation generation is based on the play-by-play and is a bit finicky, so you can try to generate new play-by-plays/animations if the first ones don't work well. It's still a bit rough around the edges, but I'd love to know what people think!
4 by kmcgraw | 0 comments on Hacker News.
In a company hackathon this weekend, I thought it would be fun to build a basketball play designer. Essentially, you enter some information about the type of play you want, and it generates a play following these requirements. I was imagining it might be useful for stuff like: 1. Designing personalized plays for first-time coaches: "Give me a play that gets my shooting guard an open three-pointer. My shooting guard shoots the best from the right wing, and my point guard has a weak left hand. My center sets great screens, so have my center set an off-ball screen for my shooting guard." 2. Brainstorming novel plays for long-time coaches: "Generate a novel play out of the UCLA set." It is loaded in with an example play, so make sure to hit the buttons to generate new plays with the entered information. The animation generation is based on the play-by-play and is a bit finicky, so you can try to generate new play-by-plays/animations if the first ones don't work well. It's still a bit rough around the edges, but I'd love to know what people think!
New Show Hacker News story: latest news
Show HN: A “CRM” for your personal relationships
2 by Ritepaw | 1 comments on Hacker News.
Android: https://ift.tt/Ci0PKYg... iOS: https://ift.tt/SZ1AndB We know life can be busy. And we often struggle to make time for the people that matter most to us. According to a Harvard study from 2021, "36% of all Americans [...] feel “serious loneliness.”" [1] We created an app to help people focus on their most personal relationships and guide them through forming more meaningful connections with them, in the hopes to combat the "Loneliness Epidemic". We just launched the first iteration - call it an MVP if you like - something to validate our ideas and get the conversation started. We are looking for people to try out what we have built and share their thoughts. The functionality is very basic for now, but we are planning to expand based on our users' feedback. Here is what you can do today: * Tell us who the people are that matter most to you * Get daily reminders to reach out to them * Send them virtual postcards with over 50 handcrafted designs [1] https://ift.tt/4Ec7pN5
2 by Ritepaw | 1 comments on Hacker News.
Android: https://ift.tt/Ci0PKYg... iOS: https://ift.tt/SZ1AndB We know life can be busy. And we often struggle to make time for the people that matter most to us. According to a Harvard study from 2021, "36% of all Americans [...] feel “serious loneliness.”" [1] We created an app to help people focus on their most personal relationships and guide them through forming more meaningful connections with them, in the hopes to combat the "Loneliness Epidemic". We just launched the first iteration - call it an MVP if you like - something to validate our ideas and get the conversation started. We are looking for people to try out what we have built and share their thoughts. The functionality is very basic for now, but we are planning to expand based on our users' feedback. Here is what you can do today: * Tell us who the people are that matter most to you * Get daily reminders to reach out to them * Send them virtual postcards with over 50 handcrafted designs [1] https://ift.tt/4Ec7pN5
New Show Hacker News story: latest news
Show HN: Vanya – FHIR data viewer for developers working with healthcare APIs
2 by DarrenDev | 0 comments on Hacker News.
FHIR => Fast Healthcare Interoperability Resources I began working with healthcare data using FHIR APIs a few years ago, and was immediately struck by the lack of useful tools for developers. Everyone was using Postman. Development and testing was slooooow. QA was a nightmare. Sprint and business demos involved running Postman queries and browsing JSON results. Painful! So I built Vanya — an app for developers in the style of a database client, but with extras for FHIR. It’s a quick viewer that shows the most recently updated resources, and allows the developer to drill down into connected resources (Patient, Procedure, Medication). The raw JSON is still there to see, but the more common fields are presented in a results grid. The backend is C# / .NET. The UI is Electron, with builds for Windows and Mac. (The Mac was a real pain to build and notarize — a post here on HN a few weeks ago helped with that.) It’s being used right now by the teams I’m working with on a big real world project, and despite its lack of features (alpha version) is delivering real value and saving devs a lot of time. I’d love any feedback from developers who work with FHIR or in the healthcare space. And anyone who’s able to test out the Mac versions. Website: https://vanyalabs.com/ Screenshot 1: https://ift.tt/HrXsKtc Screenshot 2: https://ift.tt/s5vSTxX....
2 by DarrenDev | 0 comments on Hacker News.
FHIR => Fast Healthcare Interoperability Resources I began working with healthcare data using FHIR APIs a few years ago, and was immediately struck by the lack of useful tools for developers. Everyone was using Postman. Development and testing was slooooow. QA was a nightmare. Sprint and business demos involved running Postman queries and browsing JSON results. Painful! So I built Vanya — an app for developers in the style of a database client, but with extras for FHIR. It’s a quick viewer that shows the most recently updated resources, and allows the developer to drill down into connected resources (Patient, Procedure, Medication). The raw JSON is still there to see, but the more common fields are presented in a results grid. The backend is C# / .NET. The UI is Electron, with builds for Windows and Mac. (The Mac was a real pain to build and notarize — a post here on HN a few weeks ago helped with that.) It’s being used right now by the teams I’m working with on a big real world project, and despite its lack of features (alpha version) is delivering real value and saving devs a lot of time. I’d love any feedback from developers who work with FHIR or in the healthcare space. And anyone who’s able to test out the Mac versions. Website: https://vanyalabs.com/ Screenshot 1: https://ift.tt/HrXsKtc Screenshot 2: https://ift.tt/s5vSTxX....
New Show Hacker News story: latest news
Show HN: I rewrote the 1990's LambdaMOO server from scratch
2 by cmrdporcupine | 0 comments on Hacker News.
I got my start on the Internet in the very early 90s playing with, authoring in, and programming on LambdaMOO ( https://ift.tt/3a4gky0 ) and similar systems. Shared virtual social spaces, with a persistent object oriented authoring / scripting language. They can be classified as MUDs (depending on who you talk to) but the focus is social, creative / authoring, and shared programming not RPG gaming. I've always wanted to see this kind of thing modernized and further developed. Over the last 25 years or so I've worked on similar but novel & improved things, but never finished. So I decided to just re-implement LambdaMOO and use that as a base, instead and keep compatibility as a goal, but build it out on a more modern foundation that takes advantage of multiple core machines, newer network protocols, newer connectivity methods, uses MVCC transactions for the shared database etc. LambdaMOO is a somewhat extensive system in that it is composed of compiler, a virtual machine, an object database, user permissions system, network runtime. In some ways it's kind of like a shared, text-based Smalltalk image/runtime... So quite a bit to implement and get right before it all works together. The big challenge throughout has been slavishly maintaining backwards compatibility so existing "cores" (databases) work. It's not done, but it's darn close. Would like for people who are into this kind of thing to check it out, and maybe even help. Many of the technical aspects here are still provisional, but this is the start. Constructive assistance welcome. (Yes, it's a rewrite in Rust, but that's not really the point, even though that's a cliche that's fun.)
2 by cmrdporcupine | 0 comments on Hacker News.
I got my start on the Internet in the very early 90s playing with, authoring in, and programming on LambdaMOO ( https://ift.tt/3a4gky0 ) and similar systems. Shared virtual social spaces, with a persistent object oriented authoring / scripting language. They can be classified as MUDs (depending on who you talk to) but the focus is social, creative / authoring, and shared programming not RPG gaming. I've always wanted to see this kind of thing modernized and further developed. Over the last 25 years or so I've worked on similar but novel & improved things, but never finished. So I decided to just re-implement LambdaMOO and use that as a base, instead and keep compatibility as a goal, but build it out on a more modern foundation that takes advantage of multiple core machines, newer network protocols, newer connectivity methods, uses MVCC transactions for the shared database etc. LambdaMOO is a somewhat extensive system in that it is composed of compiler, a virtual machine, an object database, user permissions system, network runtime. In some ways it's kind of like a shared, text-based Smalltalk image/runtime... So quite a bit to implement and get right before it all works together. The big challenge throughout has been slavishly maintaining backwards compatibility so existing "cores" (databases) work. It's not done, but it's darn close. Would like for people who are into this kind of thing to check it out, and maybe even help. Many of the technical aspects here are still provisional, but this is the start. Constructive assistance welcome. (Yes, it's a rewrite in Rust, but that's not really the point, even though that's a cliche that's fun.)
New Show Hacker News story: latest news
Show HN: LLM Agent Paper List
2 by Anil1331 | 0 comments on Hacker News.
The most comprehensive repo out there with all the updated must-read papers for LLM Agents From the repo - "We start by the general conceptual framework for LLM-based agents: comprising three main components: brain, perception, and action, and the framework can be tailored to suit different applications."
2 by Anil1331 | 0 comments on Hacker News.
The most comprehensive repo out there with all the updated must-read papers for LLM Agents From the repo - "We start by the general conceptual framework for LLM-based agents: comprising three main components: brain, perception, and action, and the framework can be tailored to suit different applications."
New Show Hacker News story: latest news
Show HN: A Social Bookmarking tool – Pinterest for knowledge
9 by askwhyharsh | 2 comments on Hacker News.
I spent few months building this tool after i lost my job earlier this year, it's something like a bookmark manager but super simple to use. I also built a explore page, where people upvote and view collections people curate using linkcollect, save and share their profile. It has around 1000 public collections and 1100 users
9 by askwhyharsh | 2 comments on Hacker News.
I spent few months building this tool after i lost my job earlier this year, it's something like a bookmark manager but super simple to use. I also built a explore page, where people upvote and view collections people curate using linkcollect, save and share their profile. It has around 1000 public collections and 1100 users
Friday, 22 September 2023
New Show Hacker News story: latest news
Show HN: Tinyfeed – CLI to generate static webpage from a collection of feeds
2 by TheBigRoomXXL | 0 comments on Hacker News.
This project started because I wanted to consume my RSS/Atoms feeds the same way I consume news on HN: a lightweight webpage with a minimalist user interface where I can comfortably browse interesting links. I didn't like existing solutions for feeds reader. Many of them seemed overly complex, most of them require a database and don't event support sqlite. For a simple feed reader, it felt overkill. What I wanted was an easy setup that didn't rely on external databases or configurations. My solution was to build a small CLI tool that print an HTML page from a list of feed URLs. I can just set it up in Cron, point NGINX to the generate HTML file and that's it, I have a lovely webpage with my feeds. This was my first real project in Go and I'm still learning so I'm eager to receive feedback on the code. Checkout the demo here: https://ift.tt/oNAjDV5
2 by TheBigRoomXXL | 0 comments on Hacker News.
This project started because I wanted to consume my RSS/Atoms feeds the same way I consume news on HN: a lightweight webpage with a minimalist user interface where I can comfortably browse interesting links. I didn't like existing solutions for feeds reader. Many of them seemed overly complex, most of them require a database and don't event support sqlite. For a simple feed reader, it felt overkill. What I wanted was an easy setup that didn't rely on external databases or configurations. My solution was to build a small CLI tool that print an HTML page from a list of feed URLs. I can just set it up in Cron, point NGINX to the generate HTML file and that's it, I have a lovely webpage with my feeds. This was my first real project in Go and I'm still learning so I'm eager to receive feedback on the code. Checkout the demo here: https://ift.tt/oNAjDV5
New Show Hacker News story: latest news
Show HN: Charset Normalizer library – port from Python to Rust
2 by editor678 | 0 comments on Hacker News.
I've just ported a popular library charset-normalizer from Python to Rust. I would appreciate for any feedback :)
2 by editor678 | 0 comments on Hacker News.
I've just ported a popular library charset-normalizer from Python to Rust. I would appreciate for any feedback :)
New Show Hacker News story: latest news
Show HN: Claws – A visual interface for the AWS CLI
2 by jeffgardnerdev | 0 comments on Hacker News.
I've been developing on AWS for many years now. I like the idea of using the AWS CLI but I very much dislike the developer experience of clicking through the CLI doc website or building out the long complex arguments. So I created a visual interface for the AWS CLI where you can add & edit profiles, and explore & experiment with API calls via a user-friendly interface. https://clawsapp.com I also added a natural language search to help identify the CLI command you need and fill in arguments for you based on your description. I would love to get your feedback & suggestions. Thank you in advance!
2 by jeffgardnerdev | 0 comments on Hacker News.
I've been developing on AWS for many years now. I like the idea of using the AWS CLI but I very much dislike the developer experience of clicking through the CLI doc website or building out the long complex arguments. So I created a visual interface for the AWS CLI where you can add & edit profiles, and explore & experiment with API calls via a user-friendly interface. https://clawsapp.com I also added a natural language search to help identify the CLI command you need and fill in arguments for you based on your description. I would love to get your feedback & suggestions. Thank you in advance!
New Show Hacker News story: latest news
Show HN: Learn piano without sheet music
2 by jacobp100 | 0 comments on Hacker News.
I always found sheet music way too hard to read - and I literally spent a year at a company building a sheet music rendering engine. I wanted an app that would display music like the tutorials on YouTube, but not be focused on upselling lessons etc. like most current apps, and also would let me import my own files This works on MIDI files. If it’s a valid midi it probably plays. Since releasing, I did add a subscription for classical music - on a theory that most normal users don’t know what a midi file is. It changed about a month ago from an up front price to in app purchases and/or a subscription - which has absolutely tanked revenue so far - but maybe it will pick up Would love to hear your thoughts and if you have any suggestions!
2 by jacobp100 | 0 comments on Hacker News.
I always found sheet music way too hard to read - and I literally spent a year at a company building a sheet music rendering engine. I wanted an app that would display music like the tutorials on YouTube, but not be focused on upselling lessons etc. like most current apps, and also would let me import my own files This works on MIDI files. If it’s a valid midi it probably plays. Since releasing, I did add a subscription for classical music - on a theory that most normal users don’t know what a midi file is. It changed about a month ago from an up front price to in app purchases and/or a subscription - which has absolutely tanked revenue so far - but maybe it will pick up Would love to hear your thoughts and if you have any suggestions!
New Show Hacker News story: latest news
Show HN: FormsLab – Open-source form builder
5 by Magiqon | 3 comments on Hacker News.
Hello guys, I wanted to share with you and ask your opinion about my first open-source project. FormsLab is a tool that can be used as a feedback app to collect feedback from customers, create polls for voting, or as a survey creator. I would like to ask for your feedback and opinions. Thank you!
5 by Magiqon | 3 comments on Hacker News.
Hello guys, I wanted to share with you and ask your opinion about my first open-source project. FormsLab is a tool that can be used as a feedback app to collect feedback from customers, create polls for voting, or as a survey creator. I would like to ask for your feedback and opinions. Thank you!
Thursday, 21 September 2023
New Show Hacker News story: latest news
Show HN: My Single-File Python Script I Used to Replace Splunk in My Startup
81 by eigenvalue | 12 comments on Hacker News.
My immediate reaction to today's news that Splunk was being acquired was to comment in the HN discussion for that story: "I hated Splunk so much that I spent a couple days a few months ago writing a single 1200 line python script that does absolutely everything I need in terms of automatic log collection, ingestion, and analysis from a fleet of cloud instances. It pulls in all the log lines, enriches them with useful metadata like the IP address of the instance, the machine name, the log source, the datetime, etc. and stores it all in SQlite, which it then exposes to a very convenient web interface using Datasette. I put it in a cronjob and it's infinitely better (at least for my purposes) than Splunk, which is just a total nightmare to use, and can be customized super easily and quickly. My coworkers all prefer it to Splunk as well. And oh yeah, it's totally free instead of costing my company thousands of dollars a year! If I owned CSCO stock I would sell it-- this deal shows incredibly bad judgment." I had been meaning to clean it up a bit and open-source it but never got around to it. However, someone asked today in response to my comment if I had released it, so I figured now would be a good time to go through it and clean it up, move the constants to an .env file, and create a README. This code is obviously tailored to my own requirements for my project, but if you know Python, it's extremely straightforward to customize it for your own logs (plus, some of the logs are generic, like systemd logs, and the output of netstat/ss/lsof, which it combines to get a table of open connections by process over time for each machine-- extremely useful for finding code that is leaking connections!). And I also included the actual sample log files from my project that correspond to the parsing functions in the code, so you can easily reason by analogy to adapt it to your own log files. As many people pointed out in responses to my comment, this is obviously not a real replacement for Splunk for enterprise users who are ingesting terabytes a day from thousands of machines and hundreds of sources. If it were, hopefully someone would be paying me $28 billion for it instead of me giving it away for free! But if you don't have a huge number of machines and really hate using Splunk while wasting thousands of dollars, this might be for you.
81 by eigenvalue | 12 comments on Hacker News.
My immediate reaction to today's news that Splunk was being acquired was to comment in the HN discussion for that story: "I hated Splunk so much that I spent a couple days a few months ago writing a single 1200 line python script that does absolutely everything I need in terms of automatic log collection, ingestion, and analysis from a fleet of cloud instances. It pulls in all the log lines, enriches them with useful metadata like the IP address of the instance, the machine name, the log source, the datetime, etc. and stores it all in SQlite, which it then exposes to a very convenient web interface using Datasette. I put it in a cronjob and it's infinitely better (at least for my purposes) than Splunk, which is just a total nightmare to use, and can be customized super easily and quickly. My coworkers all prefer it to Splunk as well. And oh yeah, it's totally free instead of costing my company thousands of dollars a year! If I owned CSCO stock I would sell it-- this deal shows incredibly bad judgment." I had been meaning to clean it up a bit and open-source it but never got around to it. However, someone asked today in response to my comment if I had released it, so I figured now would be a good time to go through it and clean it up, move the constants to an .env file, and create a README. This code is obviously tailored to my own requirements for my project, but if you know Python, it's extremely straightforward to customize it for your own logs (plus, some of the logs are generic, like systemd logs, and the output of netstat/ss/lsof, which it combines to get a table of open connections by process over time for each machine-- extremely useful for finding code that is leaking connections!). And I also included the actual sample log files from my project that correspond to the parsing functions in the code, so you can easily reason by analogy to adapt it to your own log files. As many people pointed out in responses to my comment, this is obviously not a real replacement for Splunk for enterprise users who are ingesting terabytes a day from thousands of machines and hundreds of sources. If it were, hopefully someone would be paying me $28 billion for it instead of me giving it away for free! But if you don't have a huge number of machines and really hate using Splunk while wasting thousands of dollars, this might be for you.
New Show Hacker News story: latest news
Show HN: Scrum Poker Hub – A Simple Tool for Effort Estimation Voting
2 by nirvanist | 0 comments on Hacker News.
2 by nirvanist | 0 comments on Hacker News.
New Show Hacker News story: latest news
Show HN: Odin – the integration of LLMs with Obsidian note taking
2 by AlexIchenskiy | 1 comments on Hacker News.
2 by AlexIchenskiy | 1 comments on Hacker News.
Wednesday, 20 September 2023
New Show Hacker News story: latest news
Show HN: Swirl – AI Based Open-Source Search Engine Alternative to Algolia
17 by srbhr | 0 comments on Hacker News.
Swirl Is an Open-Source Search engine alternative to Algolia, and the likes. But Swirl queries anything with an API then uses Large Language Models to re-rank the unified results without copying any data! If the database has an API it can connect to and provide search functionality. You can use it to generate AI insights over a large and spread-out database. Check the website: https://swirl.today/ GitHub: https://ift.tt/sziaTWA
17 by srbhr | 0 comments on Hacker News.
Swirl Is an Open-Source Search engine alternative to Algolia, and the likes. But Swirl queries anything with an API then uses Large Language Models to re-rank the unified results without copying any data! If the database has an API it can connect to and provide search functionality. You can use it to generate AI insights over a large and spread-out database. Check the website: https://swirl.today/ GitHub: https://ift.tt/sziaTWA
New Show Hacker News story: latest news
Show HN: Breathwork App for Endurance Athletes
2 by AndrusAsumets | 0 comments on Hacker News.
Hi, I had a problem of how to get into a proper shape again after quitting smoking after 17 years, where lungs of mine were the main bottleneck. I was therefore inclined to find a solution that would help me to systematically improve my breathing pace, but since I couldn't find anything related to it, I therefore decided to solve it myself. Being a designer while also having skills of a full-stack developer, I decided to learn Swift and use Apple Watch to design one. In essence, the app is rather similar to how Strava works on Apple Watch: you start your workout session, go through with it, and then have it saved for further analysis. The real difference is that you also set your breathing pace, which will translate your heart rate to breath rate and give relevant haptic cues when to breath in and out. Since I had to build a sampler to WatchOS, I also experimented with trying the app to generate background music that would be synced to both heart and breathing rates. After scraping on domain audio samples reaching over 300GB, I then used Jupyter/Python and the process of dynamic time warping to group similar sounds to eventually create an audible journey when doing workouts. Currently it only has one track/genre for ambient, but I'm thinking of adding more styles to it soon. The app also allows for fetching all your workout data in CSV though Cloudflare's Workers, if need be. I have been mostly using it when doing my runs. I recently also went through my first half-marathon, where lungs were not the bottleneck anymore, so it seems the app seems to have some merit for people to improve their breathwork. Cheers, Andrus
2 by AndrusAsumets | 0 comments on Hacker News.
Hi, I had a problem of how to get into a proper shape again after quitting smoking after 17 years, where lungs of mine were the main bottleneck. I was therefore inclined to find a solution that would help me to systematically improve my breathing pace, but since I couldn't find anything related to it, I therefore decided to solve it myself. Being a designer while also having skills of a full-stack developer, I decided to learn Swift and use Apple Watch to design one. In essence, the app is rather similar to how Strava works on Apple Watch: you start your workout session, go through with it, and then have it saved for further analysis. The real difference is that you also set your breathing pace, which will translate your heart rate to breath rate and give relevant haptic cues when to breath in and out. Since I had to build a sampler to WatchOS, I also experimented with trying the app to generate background music that would be synced to both heart and breathing rates. After scraping on domain audio samples reaching over 300GB, I then used Jupyter/Python and the process of dynamic time warping to group similar sounds to eventually create an audible journey when doing workouts. Currently it only has one track/genre for ambient, but I'm thinking of adding more styles to it soon. The app also allows for fetching all your workout data in CSV though Cloudflare's Workers, if need be. I have been mostly using it when doing my runs. I recently also went through my first half-marathon, where lungs were not the bottleneck anymore, so it seems the app seems to have some merit for people to improve their breathwork. Cheers, Andrus
New Show Hacker News story: latest news
Show HN: SeaGOAT – local, “AI-based” grep for semantic code search
5 by kantord | 0 comments on Hacker News.
5 by kantord | 0 comments on Hacker News.
Tuesday, 19 September 2023
New Show Hacker News story: latest news
Show HN: Hydra - Open-Source Columnar Postgres
41 by coatue | 5 comments on Hacker News.
hi hn, hydra ceo here hydra is an open-source extension that adds columnar tables to Postgres for efficient analytical reporting. With Hydra, you can analyze billions of rows instantly without changing code. demo video (5 min): https://youtu.be/1yzxgb0Oyrw github repo: https://ift.tt/YAnaf9e For 1.0 GA release, aggregate queries are over *60% faster* than Hydra beta due to aggregate vectorization. Spatial indexes (gin, gist, spgist, and rum indexes) and pg_hint_plan are now enabled for performance optimization. postgres is great, but aggregates can take minutes to hours to return results on large data sets. long-running analytical queries hog database resources and degrade performance. use hydra to run much faster analytics on postgres without changing code. for testing, try the hydra free tier to create a column postgres instance on the cloud. https://ift.tt/vLNio2I
41 by coatue | 5 comments on Hacker News.
hi hn, hydra ceo here hydra is an open-source extension that adds columnar tables to Postgres for efficient analytical reporting. With Hydra, you can analyze billions of rows instantly without changing code. demo video (5 min): https://youtu.be/1yzxgb0Oyrw github repo: https://ift.tt/YAnaf9e For 1.0 GA release, aggregate queries are over *60% faster* than Hydra beta due to aggregate vectorization. Spatial indexes (gin, gist, spgist, and rum indexes) and pg_hint_plan are now enabled for performance optimization. postgres is great, but aggregates can take minutes to hours to return results on large data sets. long-running analytical queries hog database resources and degrade performance. use hydra to run much faster analytics on postgres without changing code. for testing, try the hydra free tier to create a column postgres instance on the cloud. https://ift.tt/vLNio2I
New Show Hacker News story: latest news
Show HN: Focusdoro(AI) – AI assisted Pomodoro task list
3 by supershobu | 0 comments on Hacker News.
Hi, I have always struggled with changing priorities and figuring out what to focus on a particular day. The actual priorities get missed amongst all other things in life. With the launch of Gen AI in almost all aspects of our lives, i decided to build something that can assist me in finding the right priorities and staying on them. This has helped me out for the past couple of days. I am using Openai apis for task breakdown and Daily planning bot and the rest of the site is built using Django, JS, and Tailwind. I have tried a lot of productivity tools throughout the years and Pomodoro and timer-based focused work has worked out best for me. If there are other techniques that have worked out for you, do let me know, I will try to incorporate those :) Would love honest feedback on this so that I can make this more useful to others and to myself. Looking forward to your feedback, thanks in advance :)
3 by supershobu | 0 comments on Hacker News.
Hi, I have always struggled with changing priorities and figuring out what to focus on a particular day. The actual priorities get missed amongst all other things in life. With the launch of Gen AI in almost all aspects of our lives, i decided to build something that can assist me in finding the right priorities and staying on them. This has helped me out for the past couple of days. I am using Openai apis for task breakdown and Daily planning bot and the rest of the site is built using Django, JS, and Tailwind. I have tried a lot of productivity tools throughout the years and Pomodoro and timer-based focused work has worked out best for me. If there are other techniques that have worked out for you, do let me know, I will try to incorporate those :) Would love honest feedback on this so that I can make this more useful to others and to myself. Looking forward to your feedback, thanks in advance :)
New Show Hacker News story: latest news
Show HN: Dittofeed – 1-Click deploy, self-host Mailchimp alternative
4 by maxthegeek1 | 0 comments on Hacker News.
Hi HN, we've released Dittofeed v0.4.0, a significant update primarily focused on facilitating self-hosting. Dittofeed is an open source (MIT licensed) alternative to platforms like customer.io, mailchimp, klaviyo, iterable etc. Self-hosting is particularly useful if you: - Want to evaluate Dittofeed without a long-term commitment - Prefer fixed, non-volume based pricing models - Aim to keep all personally identifiable information within your own infrastructure In this version, we introduced 'Dittofeed Lite,' which merges the dashboard, API, and worker services, simplifying deployment, along with providing an easy auth setup. We've also configured a 1-Click Render deployment. The Render setup comes at a fixed cost of $39/month. However, by using alternative hosting solutions for the Postgres and ClickHouse instances (neon, clickhouse cloud), the monthly costs can be reduced to under $20. We personally feel that there’s a big problem with how existing platforms handle pricing, charging insanely high prices with low contact limits, while restricting access to really basic table stakes features. Excited to share this work, and hear your thoughts and feedback! Github Repo - https://ift.tt/blGrmJA Demo - https://ift.tt/2uHsi0V Docs - https://ift.tt/3ylCtY9
4 by maxthegeek1 | 0 comments on Hacker News.
Hi HN, we've released Dittofeed v0.4.0, a significant update primarily focused on facilitating self-hosting. Dittofeed is an open source (MIT licensed) alternative to platforms like customer.io, mailchimp, klaviyo, iterable etc. Self-hosting is particularly useful if you: - Want to evaluate Dittofeed without a long-term commitment - Prefer fixed, non-volume based pricing models - Aim to keep all personally identifiable information within your own infrastructure In this version, we introduced 'Dittofeed Lite,' which merges the dashboard, API, and worker services, simplifying deployment, along with providing an easy auth setup. We've also configured a 1-Click Render deployment. The Render setup comes at a fixed cost of $39/month. However, by using alternative hosting solutions for the Postgres and ClickHouse instances (neon, clickhouse cloud), the monthly costs can be reduced to under $20. We personally feel that there’s a big problem with how existing platforms handle pricing, charging insanely high prices with low contact limits, while restricting access to really basic table stakes features. Excited to share this work, and hear your thoughts and feedback! Github Repo - https://ift.tt/blGrmJA Demo - https://ift.tt/2uHsi0V Docs - https://ift.tt/3ylCtY9
New Show Hacker News story: latest news
Show HN: Graphite – Stacked Diffs on GitHub
32 by tomasreimers | 5 comments on Hacker News.
TLDR; Graphite enables a git workflow called “stacking” - the fastest way to develop and ship code, which many large tech companies have been using for years. Graphite makes stacking available to anyone with a GitHub account. Hi HN! I’m Tomas, co-founder of graphite.dev, and today we’re launching Graphite after almost two years of development in closed beta. [1] Graphite started as an internal solution to our own problem. When we (engineers from Meta, Google and Airbnb) left our previous roles, we lost access to the internal code review tools we loved. So we built our own. https://graphite.dev --- Graphite is how the fastest developers ship code - it’s a developer tool that allows you to create smaller pull requests, stay unblocked, and ship faster with “stacking” (creating a set of dependent pull requests). Stacking [2] allows developers to break up large pull requests (PRs) into smaller ones that can be reviewed & merged independently, while keeping ongoing development unblocked. Engineering best practices at Google advise that a “reasonable” PR be around 100 lines, and recommend splitting PRs in order to achieve this through stacking or other methods. [3] Unlike other tools like Phabricator, Gerrit, or Sapling, Graphite syncs seamlessly with your GitHub repositories, so that you don’t have to manage any extra infrastructure. This also means that even if your teammates don’t use Graphite yet, you still can. Here’s what you can expect when you sign in to Graphite with your GitHub account: (1) First class support for stacking: At its core, Graphite enables “stacking”—a workflow used by engineers at top companies like Meta and Google to create small, dependent sets of pull requests. The Graphite CLI, web app, and VS Code extension all come together to empower engineers to start stacking. (2) Pull request inbox: You can think of this as your home page on Graphite, where you have full visibility into the status of all your PRs and know what still needs to be done across every repo, author, and stage of review. You can also create custom inboxes that filter PRs on reviewers, authors, labels, CI status, and more. (3) Streamlined code review interface: Graphite’s pull request page removes tabs and minimizes distractions, with the aim of putting your code front and center. You can use keyboard shortcuts to navigate between files and comments or to move between PRs in your stack. You can also import custom memes and gifs to add some to your reviews too! (4) AI-powered pull requests: Auto-generate a detailed description for every PR with our OpenAI integration. You can even turn your comments into suggested code changes (coming soon!). (5) Real-time notifications: Connect Graphite to your Slack workspace to stay up-to-date on review requests, comments threads, merge status, and other activity on your PRs. For smaller PRs, you can leave a review (and even merge) directly from Slack. (6) Stack-aware merges: Since Graphite is built to support a stacking workflow, it automates the manual work of rebasing PRs when it’s time to merge. You can merge your stacks with one click from the web app, or in a single command from the CLI. Feel free to take a look at our getting started guide [4] or product tour video [5] for a tutorial on how to get started, and drop your comments to us below! [1] https://ift.tt/SfrdRCg [2] http://stacking.dev [3] https://ift.tt/TtFBnZp [4] https://ift.tt/dZ74b9S [5] https://www.youtube.com/watch?v=sBcd9uopLOY
32 by tomasreimers | 5 comments on Hacker News.
TLDR; Graphite enables a git workflow called “stacking” - the fastest way to develop and ship code, which many large tech companies have been using for years. Graphite makes stacking available to anyone with a GitHub account. Hi HN! I’m Tomas, co-founder of graphite.dev, and today we’re launching Graphite after almost two years of development in closed beta. [1] Graphite started as an internal solution to our own problem. When we (engineers from Meta, Google and Airbnb) left our previous roles, we lost access to the internal code review tools we loved. So we built our own. https://graphite.dev --- Graphite is how the fastest developers ship code - it’s a developer tool that allows you to create smaller pull requests, stay unblocked, and ship faster with “stacking” (creating a set of dependent pull requests). Stacking [2] allows developers to break up large pull requests (PRs) into smaller ones that can be reviewed & merged independently, while keeping ongoing development unblocked. Engineering best practices at Google advise that a “reasonable” PR be around 100 lines, and recommend splitting PRs in order to achieve this through stacking or other methods. [3] Unlike other tools like Phabricator, Gerrit, or Sapling, Graphite syncs seamlessly with your GitHub repositories, so that you don’t have to manage any extra infrastructure. This also means that even if your teammates don’t use Graphite yet, you still can. Here’s what you can expect when you sign in to Graphite with your GitHub account: (1) First class support for stacking: At its core, Graphite enables “stacking”—a workflow used by engineers at top companies like Meta and Google to create small, dependent sets of pull requests. The Graphite CLI, web app, and VS Code extension all come together to empower engineers to start stacking. (2) Pull request inbox: You can think of this as your home page on Graphite, where you have full visibility into the status of all your PRs and know what still needs to be done across every repo, author, and stage of review. You can also create custom inboxes that filter PRs on reviewers, authors, labels, CI status, and more. (3) Streamlined code review interface: Graphite’s pull request page removes tabs and minimizes distractions, with the aim of putting your code front and center. You can use keyboard shortcuts to navigate between files and comments or to move between PRs in your stack. You can also import custom memes and gifs to add some to your reviews too! (4) AI-powered pull requests: Auto-generate a detailed description for every PR with our OpenAI integration. You can even turn your comments into suggested code changes (coming soon!). (5) Real-time notifications: Connect Graphite to your Slack workspace to stay up-to-date on review requests, comments threads, merge status, and other activity on your PRs. For smaller PRs, you can leave a review (and even merge) directly from Slack. (6) Stack-aware merges: Since Graphite is built to support a stacking workflow, it automates the manual work of rebasing PRs when it’s time to merge. You can merge your stacks with one click from the web app, or in a single command from the CLI. Feel free to take a look at our getting started guide [4] or product tour video [5] for a tutorial on how to get started, and drop your comments to us below! [1] https://ift.tt/SfrdRCg [2] http://stacking.dev [3] https://ift.tt/TtFBnZp [4] https://ift.tt/dZ74b9S [5] https://www.youtube.com/watch?v=sBcd9uopLOY
Monday, 18 September 2023
New Show Hacker News story: latest news
Show HN: Collaborative Lisp Coding on Discord
2 by D4ckard | 0 comments on Hacker News.
This is a Discord bot I made, which lets you write and evaluate Lisp code in collaborative sessions on Discord. I wrote it in a short amount of time, partly for learning purposes, but also because I found the idea exciting and it was fun to hack around. I'd be happy to get your feedback, especially on how to improve the deletion mechanism (maybe make deletion sexpr-based?). I've tried to make the README detailed enough so that it's easy to make some changes and run your own version.
2 by D4ckard | 0 comments on Hacker News.
This is a Discord bot I made, which lets you write and evaluate Lisp code in collaborative sessions on Discord. I wrote it in a short amount of time, partly for learning purposes, but also because I found the idea exciting and it was fun to hack around. I'd be happy to get your feedback, especially on how to improve the deletion mechanism (maybe make deletion sexpr-based?). I've tried to make the README detailed enough so that it's easy to make some changes and run your own version.
New Show Hacker News story: latest news
Show HN: macOS GUI for running LLMs locally
6 by cztomsik | 0 comments on Hacker News.
Hello HN, I've been working on this project for a while, and it has been in an "open" beta for some time. I finally believe it's ready for its first release. I hope you like it. Here are some potential questions that may arise: 1. How does it compare to LM Studio? It's likely that if you're already using LM Studio, you'll continue to do so. This project is designed to be more user-friendly. 2. Is it open-source? No, it is not. 3. Does it use any open-source libraries? Yes, it uses llama.cpp and a few others, as indicated in the license information included with the application. 4. Why is not using electronjs? Two reasons, I wanted total control over the whole tech-stack and second, I wanted to be able to send this to my friends over iMessage. 5. Does it support Intel macs? It should, but I couldn't test it. 6. Does it support older macOS? 12.6 is the lowest version at the moment. 7. Is XXX a bug? Probably :)
6 by cztomsik | 0 comments on Hacker News.
Hello HN, I've been working on this project for a while, and it has been in an "open" beta for some time. I finally believe it's ready for its first release. I hope you like it. Here are some potential questions that may arise: 1. How does it compare to LM Studio? It's likely that if you're already using LM Studio, you'll continue to do so. This project is designed to be more user-friendly. 2. Is it open-source? No, it is not. 3. Does it use any open-source libraries? Yes, it uses llama.cpp and a few others, as indicated in the license information included with the application. 4. Why is not using electronjs? Two reasons, I wanted total control over the whole tech-stack and second, I wanted to be able to send this to my friends over iMessage. 5. Does it support Intel macs? It should, but I couldn't test it. 6. Does it support older macOS? 12.6 is the lowest version at the moment. 7. Is XXX a bug? Probably :)
New Show Hacker News story: latest news
Show HN: VimGPT: LLM agent/toolkit for fast file edits using Neovim
2 by nbrad | 0 comments on Hacker News.
2 by nbrad | 0 comments on Hacker News.
New Show Hacker News story: latest news
Show HN: A murder mystery game built on an open-source gen-AI agent framework
8 by broken_clock | 2 comments on Hacker News.
Hey HN, Michael and Scott here. We’re open-sourcing an interactive murder mystery featuring LLM-driven character agents. Solve the mystery by finding clues, taking notes, and interrogating agents. They all have distinct motives, personality, and can impact the game in different ways (attacking you, running away, etc). Try it out, it’s pretty fun! We’re also open-sourcing the framework that we used to make and refine the agents. The goal is to create an intuitive interface for storytellers to create, debug, and test game agents. We then take those game agents and expose an API beyond just chat - such as actions, player guardrails, emotional queries, etc. We’re not done yet - there are a lot more features coming on the way: scenario-based agent evals, agent-storyline consistency management, automatic agent generation, etc. We would love to hear your feedback. Thanks! [0] https://ift.tt/y25XRhw [1] https://ift.tt/4XziLZo [2] https://ift.tt/5wHkMnK
8 by broken_clock | 2 comments on Hacker News.
Hey HN, Michael and Scott here. We’re open-sourcing an interactive murder mystery featuring LLM-driven character agents. Solve the mystery by finding clues, taking notes, and interrogating agents. They all have distinct motives, personality, and can impact the game in different ways (attacking you, running away, etc). Try it out, it’s pretty fun! We’re also open-sourcing the framework that we used to make and refine the agents. The goal is to create an intuitive interface for storytellers to create, debug, and test game agents. We then take those game agents and expose an API beyond just chat - such as actions, player guardrails, emotional queries, etc. We’re not done yet - there are a lot more features coming on the way: scenario-based agent evals, agent-storyline consistency management, automatic agent generation, etc. We would love to hear your feedback. Thanks! [0] https://ift.tt/y25XRhw [1] https://ift.tt/4XziLZo [2] https://ift.tt/5wHkMnK
New Show Hacker News story: latest news
Show HN: HyperDX – open-source dev-friendly Datadog alternative
10 by mikeshi42 | 0 comments on Hacker News.
Hi HN, Mike and Warren here! We've been building HyperDX (hyperdx.io). HyperDX allows you to easily search and correlate logs, traces, metrics (alpha), and session replays all in one place. For example, if a user reports a bug “this button doesn't work," an engineer can play back what the user was doing in their browser and trace API calls back to the backend logs for that specific request, all from a single view. Github Repo: https://ift.tt/2v6BFiX Coming from an observability nerd background, with Warren being SRE #1 at his last startup and me previously leading dev experience at LogDNA/Mezmo, we knew there were gaps in the existing tools we were used to using. Our previous stack of tools like Bugsnag, LogRocket, and Cloudwatch required us to switch between different tools, correlate timestamps (UTC? local?), and manually cross-check IDs to piece together what was actually happening. This often made meant small issues required hours of frustration to root cause. Other tools like Datadog or New Relic come with high price tags - when estimating costs for Datadog in the past, we found that our Datadog bill would exceed our AWS bill! Other teams have had to adjust their infrastructure just to appease the Datadog pricing model. To build HyperDX, we've centralized all the telemetry in one place by leveraging OpenTelemetry (a CNCF project for standardizing/collecting telemetry) to pull and correlate logs, metrics, traces, and replays. In-app, we can correlate your logs/traces together in one panel by joining everything automatically via trace ids and session ids, so you can go from log <> trace <> replay in the same panel. To keep costs low, we store everything in Clickhouse (w/ S3 backing) to make it extremely affordable to store large amounts of data (compared to Elasticsearch) while still being able to query it efficiently (compared to services like Cloudwatch or Loki), in large part thanks to Clickhouse's bloom filters + columnar layout. On top of that, we've focused on providing a smooth developer experience (the DX in HyperDX!). This includes features like native parsing of JSON logs, full-text search on any log or trace, 2-click alert creation, and SDKs that help you get started with OpenTelemetry faster than the default OpenTelemetry SDKs. I'm excited to share what we've been working with you all and would love to hear your feedback and opinions! Hosted Demo - https://ift.tt/HGodiq4 Open Source Repo: https://ift.tt/2v6BFiX Landing Page: https://hyperdx.io
10 by mikeshi42 | 0 comments on Hacker News.
Hi HN, Mike and Warren here! We've been building HyperDX (hyperdx.io). HyperDX allows you to easily search and correlate logs, traces, metrics (alpha), and session replays all in one place. For example, if a user reports a bug “this button doesn't work," an engineer can play back what the user was doing in their browser and trace API calls back to the backend logs for that specific request, all from a single view. Github Repo: https://ift.tt/2v6BFiX Coming from an observability nerd background, with Warren being SRE #1 at his last startup and me previously leading dev experience at LogDNA/Mezmo, we knew there were gaps in the existing tools we were used to using. Our previous stack of tools like Bugsnag, LogRocket, and Cloudwatch required us to switch between different tools, correlate timestamps (UTC? local?), and manually cross-check IDs to piece together what was actually happening. This often made meant small issues required hours of frustration to root cause. Other tools like Datadog or New Relic come with high price tags - when estimating costs for Datadog in the past, we found that our Datadog bill would exceed our AWS bill! Other teams have had to adjust their infrastructure just to appease the Datadog pricing model. To build HyperDX, we've centralized all the telemetry in one place by leveraging OpenTelemetry (a CNCF project for standardizing/collecting telemetry) to pull and correlate logs, metrics, traces, and replays. In-app, we can correlate your logs/traces together in one panel by joining everything automatically via trace ids and session ids, so you can go from log <> trace <> replay in the same panel. To keep costs low, we store everything in Clickhouse (w/ S3 backing) to make it extremely affordable to store large amounts of data (compared to Elasticsearch) while still being able to query it efficiently (compared to services like Cloudwatch or Loki), in large part thanks to Clickhouse's bloom filters + columnar layout. On top of that, we've focused on providing a smooth developer experience (the DX in HyperDX!). This includes features like native parsing of JSON logs, full-text search on any log or trace, 2-click alert creation, and SDKs that help you get started with OpenTelemetry faster than the default OpenTelemetry SDKs. I'm excited to share what we've been working with you all and would love to hear your feedback and opinions! Hosted Demo - https://ift.tt/HGodiq4 Open Source Repo: https://ift.tt/2v6BFiX Landing Page: https://hyperdx.io
New Show Hacker News story: latest news
Show HN: Visualizing pipeline parallel algorithms: GPipe, 1F1B (& -2bw, & eager)
3 by sighingnow | 0 comments on Hacker News.
Visualizing how pipeline mode parallel works and how batches are scheduled under various scheduling scheme: GPipe (F-then-B), Pipedream (1F1b), Pipedrea-2BW (no flushes between iterations) and Eager-1F1B (better computation-communication overlapping).
3 by sighingnow | 0 comments on Hacker News.
Visualizing how pipeline mode parallel works and how batches are scheduled under various scheduling scheme: GPipe (F-then-B), Pipedream (1F1b), Pipedrea-2BW (no flushes between iterations) and Eager-1F1B (better computation-communication overlapping).
New Show Hacker News story: latest news
Show HN: Crystal Maiden Discord Bot
2 by anonymousd3vil | 0 comments on Hacker News.
So I made a playful discord bot in Python using my recent project horde-client. The bot impersonates as a video game character from "Dota 2" Crystal Maiden (duh). The entire project is available in replit.
2 by anonymousd3vil | 0 comments on Hacker News.
So I made a playful discord bot in Python using my recent project horde-client. The bot impersonates as a video game character from "Dota 2" Crystal Maiden (duh). The entire project is available in replit.
New Show Hacker News story: latest news
Show HN: LLM Powered Keyboard [video]
3 by da4id | 0 comments on Hacker News.
Hey HN, I'm excited to share Taikoboard, a mechanical keyboard with LLM powered autocomplete. I made this keyboard for myself so I could use LLM-powered autocomplete in apps that didn't have this feature. Although more apps have started integrating LLMs, there's a large number of software tools that don't support this. Taikoboard bridges that gap, making the AI autocomplete experience universal. Would love to hear feedback, answer questions, or discuss potential use cases you see for Taikoboard.
3 by da4id | 0 comments on Hacker News.
Hey HN, I'm excited to share Taikoboard, a mechanical keyboard with LLM powered autocomplete. I made this keyboard for myself so I could use LLM-powered autocomplete in apps that didn't have this feature. Although more apps have started integrating LLMs, there's a large number of software tools that don't support this. Taikoboard bridges that gap, making the AI autocomplete experience universal. Would love to hear feedback, answer questions, or discuss potential use cases you see for Taikoboard.
Sunday, 17 September 2023
New Show Hacker News story: latest news
Show HN: Panvimdoc – Convert pandoc Markdown to vimdoc help
2 by kdheepak | 0 comments on Hacker News.
2 by kdheepak | 0 comments on Hacker News.
New Show Hacker News story: latest news
Show HN: Live Train Webcams
2 by eskibars | 0 comments on Hacker News.
Our son is really into trains. He was home sick a few weeks ago and we found that there are several YouTube channels that have live video feeds from various train stations or railway clubs around the US. He really liked it, but he kept wanting to know when the next train was going to show up so he could see it. I built this small website at https://ift.tt/dnlFv2V which shows the time to the next train (Pacific time) and will automatically rotate channels to find the next train. It uses the Amtrak live train data today. In the future I might add other agencies, since a few of these stations do have more than one agency that services them and there are also some international stations I found feeds for. Hope this may be of some use for those of you with your own railfans in your life :)
2 by eskibars | 0 comments on Hacker News.
Our son is really into trains. He was home sick a few weeks ago and we found that there are several YouTube channels that have live video feeds from various train stations or railway clubs around the US. He really liked it, but he kept wanting to know when the next train was going to show up so he could see it. I built this small website at https://ift.tt/dnlFv2V which shows the time to the next train (Pacific time) and will automatically rotate channels to find the next train. It uses the Amtrak live train data today. In the future I might add other agencies, since a few of these stations do have more than one agency that services them and there are also some international stations I found feeds for. Hope this may be of some use for those of you with your own railfans in your life :)
New Show Hacker News story: latest news
Show HN: Type and Calculate Anything, Instantly. Numi like Calculator.
2 by prenx4x | 0 comments on Hacker News.
2 by prenx4x | 0 comments on Hacker News.
New Show Hacker News story: latest news
Show HN: Me and my buddy made $20 with a Stripe link and a Tweet
3 by Adriannnnn3 | 1 comments on Hacker News.
Hi there, So a buddy a couple of days ago came up with the idea for a new SaaS product regarding gpt and image generation. We did a quick MVP that barely works, it breaks almost 50% of the time but we still wanted to validate the idea before going all in. We have almost 1k followers on Twitter together. Did a quick post describing the idea with a stripe link in the comments with a 5$ value, no landing page, and no shipped product. In exchange, the buyers become beta users. Around 30 mins later we got our first sale next 30 mins another 5$ and so on. Do we consider the idea validated? The short answer is no, We gained some initial momentum but we can't call it a business yet, not until we are starting to get recurring revenue. What next? We created a notion file with multiple marketing channels we plan to test out and see which one is the best. We also created a list full of websites to submit our product and we are tracking the traffic we are going to get from each of them. PS: If you are curious, this is the product https://coverposts.com/
3 by Adriannnnn3 | 1 comments on Hacker News.
Hi there, So a buddy a couple of days ago came up with the idea for a new SaaS product regarding gpt and image generation. We did a quick MVP that barely works, it breaks almost 50% of the time but we still wanted to validate the idea before going all in. We have almost 1k followers on Twitter together. Did a quick post describing the idea with a stripe link in the comments with a 5$ value, no landing page, and no shipped product. In exchange, the buyers become beta users. Around 30 mins later we got our first sale next 30 mins another 5$ and so on. Do we consider the idea validated? The short answer is no, We gained some initial momentum but we can't call it a business yet, not until we are starting to get recurring revenue. What next? We created a notion file with multiple marketing channels we plan to test out and see which one is the best. We also created a list full of websites to submit our product and we are tracking the traffic we are going to get from each of them. PS: If you are curious, this is the product https://coverposts.com/
New Show Hacker News story: latest news
Show HN: ARA Records Ansible and makes it easier to understand and troubleshoot
3 by rfc2549 | 0 comments on Hacker News.
Hey HN, long time lurker here. I humbly present ara which is a project I created back at Red Hat in 2016 (before AWX was open source) to make my life easier working with distributed and large scale Ansible playbooks. It is designed to be simple and compatible with however you're already running Ansible today without needing to change much of your existing workflows. My time and skills are limited but I have learned a lot and I am happy with how the project has steadily improved over the years with the much appreciated help of contributors. If you'd like to read more about the project, the blog might be interesting: https://ift.tt/KM7leP3 If you'd like to help, contribute or chat, feel free to hop on IRC, Slack or Matrix: https://ift.tt/09e7z2C The best way to stay up to date with the project is on Mastodon: https://ift.tt/KN4eucA I will be here for a while to reply to comments and answer questions.
3 by rfc2549 | 0 comments on Hacker News.
Hey HN, long time lurker here. I humbly present ara which is a project I created back at Red Hat in 2016 (before AWX was open source) to make my life easier working with distributed and large scale Ansible playbooks. It is designed to be simple and compatible with however you're already running Ansible today without needing to change much of your existing workflows. My time and skills are limited but I have learned a lot and I am happy with how the project has steadily improved over the years with the much appreciated help of contributors. If you'd like to read more about the project, the blog might be interesting: https://ift.tt/KM7leP3 If you'd like to help, contribute or chat, feel free to hop on IRC, Slack or Matrix: https://ift.tt/09e7z2C The best way to stay up to date with the project is on Mastodon: https://ift.tt/KN4eucA I will be here for a while to reply to comments and answer questions.
New Show Hacker News story: latest news
Show HN: Infinitely Recyclable Plastic (PDK) from Berkley
2 by mojomark | 0 comments on Hacker News.
2 by mojomark | 0 comments on Hacker News.
New Show Hacker News story: latest news
Show HN: I made a browser extension for building your own custom HN themes
2 by dandrew5 | 0 comments on Hacker News.
Hi HN! I’ve spent the last few weeks writing a browser extension that lets users write their own HN themes using Handlebars and CSS. It works by converting the current Hacker News page’s content into an object that is made available to the user-defined Handlebars templates. There is a built-in code editor that makes writing themes super easy and you can see changes in real-time. If you don’t want to write a theme from scratch, there are premade themes available or you may import one made by somebody else. It is shipped with Bootstrap so you can get pretty far without needing to write a lot of custom CSS. Material icons are also included so your themes get iconography out of the box. Any theme can be toggled into dark mode with a single click. It is open-sourced under the MIT license so feel free to use any of the code in the linked repo[0]. Additionally, I’ve put together a short demo video on YouTube[1]. If you decide to check it out, let me know of any feedback and please share any themes you create! I’d be happy to answer any questions. [0] https://ift.tt/UVjylXo [1] https://youtu.be/6DxLJQrKXa0
2 by dandrew5 | 0 comments on Hacker News.
Hi HN! I’ve spent the last few weeks writing a browser extension that lets users write their own HN themes using Handlebars and CSS. It works by converting the current Hacker News page’s content into an object that is made available to the user-defined Handlebars templates. There is a built-in code editor that makes writing themes super easy and you can see changes in real-time. If you don’t want to write a theme from scratch, there are premade themes available or you may import one made by somebody else. It is shipped with Bootstrap so you can get pretty far without needing to write a lot of custom CSS. Material icons are also included so your themes get iconography out of the box. Any theme can be toggled into dark mode with a single click. It is open-sourced under the MIT license so feel free to use any of the code in the linked repo[0]. Additionally, I’ve put together a short demo video on YouTube[1]. If you decide to check it out, let me know of any feedback and please share any themes you create! I’d be happy to answer any questions. [0] https://ift.tt/UVjylXo [1] https://youtu.be/6DxLJQrKXa0
Saturday, 16 September 2023
New Show Hacker News story: latest news
Show HN: Yet another Hacker News app
2 by asim | 0 comments on Hacker News.
Thought I'd show off my CSS skills with a little cleaned up version. Often I do wonder what alternative ways are to viewing hackernews. How much more can you strip away right?
2 by asim | 0 comments on Hacker News.
Thought I'd show off my CSS skills with a little cleaned up version. Often I do wonder what alternative ways are to viewing hackernews. How much more can you strip away right?
New Show Hacker News story: latest news
Show HN: My demo for vector embeddings for the Earth's surface
2 by ckrapu | 1 comments on Hacker News.
2 by ckrapu | 1 comments on Hacker News.
New Show Hacker News story: latest news
Show HN: Every Breath You Take – Heart Rate Variability Training
2 by kbre93 | 0 comments on Hacker News.
Through controlled breathing it is possible to regulate your body's stress response. This application allows you to measure and train this effect with a Polar H10 Heart Rate monitor. With every breath you take, you can set the pace of your breathing rate, measure your breathing control with the chest accelerometer, and see how heart rate variability responds.
2 by kbre93 | 0 comments on Hacker News.
Through controlled breathing it is possible to regulate your body's stress response. This application allows you to measure and train this effect with a Polar H10 Heart Rate monitor. With every breath you take, you can set the pace of your breathing rate, measure your breathing control with the chest accelerometer, and see how heart rate variability responds.
Friday, 15 September 2023
New Show Hacker News story: latest news
Show HN: Shello – Wrangle Environment Variables
2 by shayneo | 0 comments on Hacker News.
A lightweight mac app for managing environment variables locally. No auth, no network requests, no usage tracking... just files on your machine.
2 by shayneo | 0 comments on Hacker News.
A lightweight mac app for managing environment variables locally. No auth, no network requests, no usage tracking... just files on your machine.
New Show Hacker News story: latest news
Show HN: Aesort – create a personalized ranking of aesthetics
2 by 7373737373 | 0 comments on Hacker News.
2 by 7373737373 | 0 comments on Hacker News.
New Show Hacker News story: latest news
Show HN: A functional version of my MIDI Keyboard with CSS and React
3 by grvcoelho | 0 comments on Hacker News.
I'm learning music and have bought a MIDI Keyboard. I tried to recreate it in a web version with support to: multiple instruments, volume + reverb + velocity faders, sustain and multipe octaves and keyboard support :) You can also highlight music scales on the keyboard Here's my takeaways 1. TailwindCSS is possibly the best way to prototype something with CSS at the moment. Zero configuration using Next.js and an awesome tooling around it 2. It's funny how DHH is wrong about no typings in JS. 3. React still feels like a lot of "boilerplate" to do some basic stuff (like managing state).
3 by grvcoelho | 0 comments on Hacker News.
I'm learning music and have bought a MIDI Keyboard. I tried to recreate it in a web version with support to: multiple instruments, volume + reverb + velocity faders, sustain and multipe octaves and keyboard support :) You can also highlight music scales on the keyboard Here's my takeaways 1. TailwindCSS is possibly the best way to prototype something with CSS at the moment. Zero configuration using Next.js and an awesome tooling around it 2. It's funny how DHH is wrong about no typings in JS. 3. React still feels like a lot of "boilerplate" to do some basic stuff (like managing state).
New Show Hacker News story: latest news
Show HN: I built a dataset of 25k+ Gumroad products to spot market opportunities
26 by topoftheforts | 11 comments on Hacker News.
26 by topoftheforts | 11 comments on Hacker News.
Thursday, 14 September 2023
New Show Hacker News story: latest news
Show HN: SlackRevert, a Firefox extension to revert Slack's new design
2 by apimade | 0 comments on Hacker News.
2 by apimade | 0 comments on Hacker News.
New Show Hacker News story: latest news
Show HN: Image Eval – An evaluation toolkit for image generation models
3 by nutellalover | 0 comments on Hacker News.
3 by nutellalover | 0 comments on Hacker News.
New Show Hacker News story: latest news
Show HN: A vector database with semantic SQL-like filtering
3 by mengk | 0 comments on Hacker News.
Hi HN! It’s always bothered me that there’s no real equivalent of SQL WHERE for vector content. Filtering is one of the cornerstones of a modern database — but vector DBs only support either top-k sort, which is only useful for fuzzy search, or metadata filtering, which isn’t semantic. I’ve found myself wanting all the results matching my semantic query, not just k! Aside from data analysis, it's relevant if you’re trying to do any LLM reasoning: you don’t make good decisions or reach good conclusions by considering a small subset of information. So, we’ve designed a filtering primitive on top of vectors and assembled a demo on customer reviews from Trustpilot, Yelp, App Store, etc. You can select any brand/restaurant/app, and slice the review data however you want. The filter should find all matching documents, not just the top-k. Check it out at https://ift.tt/V7EzhaW ! Not super optimized yet, and really just an exploration, but hopefully gets the point across. FAQ: - Can I try it on my own data? Sure, shoot me a message at hello [at] emberml [dot] com. - How does it work? We’ve built a custom vector-based index, and we learn a high-quality decision boundary between relevant and irrelevant vectors at query time. You can think of it as forming a few-shot classifier each time. - What’s the catch? It’s far slower and less scalable than KNN/ANN right now. But I’d rather solve quality before trying to scale up quantity; tbh I’m not satisfied with vector DB performance even at @ N=1,000. A hot take, maybe? - Why don’t you just classify the data beforehand? Unstructured data has too many degrees of freedom, so it’s hard to anticipate every search/filter a priori. Our approach is somewhat analogous to schema-on-read.
3 by mengk | 0 comments on Hacker News.
Hi HN! It’s always bothered me that there’s no real equivalent of SQL WHERE for vector content. Filtering is one of the cornerstones of a modern database — but vector DBs only support either top-k sort, which is only useful for fuzzy search, or metadata filtering, which isn’t semantic. I’ve found myself wanting all the results matching my semantic query, not just k! Aside from data analysis, it's relevant if you’re trying to do any LLM reasoning: you don’t make good decisions or reach good conclusions by considering a small subset of information. So, we’ve designed a filtering primitive on top of vectors and assembled a demo on customer reviews from Trustpilot, Yelp, App Store, etc. You can select any brand/restaurant/app, and slice the review data however you want. The filter should find all matching documents, not just the top-k. Check it out at https://ift.tt/V7EzhaW ! Not super optimized yet, and really just an exploration, but hopefully gets the point across. FAQ: - Can I try it on my own data? Sure, shoot me a message at hello [at] emberml [dot] com. - How does it work? We’ve built a custom vector-based index, and we learn a high-quality decision boundary between relevant and irrelevant vectors at query time. You can think of it as forming a few-shot classifier each time. - What’s the catch? It’s far slower and less scalable than KNN/ANN right now. But I’d rather solve quality before trying to scale up quantity; tbh I’m not satisfied with vector DB performance even at @ N=1,000. A hot take, maybe? - Why don’t you just classify the data beforehand? Unstructured data has too many degrees of freedom, so it’s hard to anticipate every search/filter a priori. Our approach is somewhat analogous to schema-on-read.
New Show Hacker News story: latest news
Show HN: Ever Been Rated by an AI Date? Try Dating Playground and Get Feedback
4 by akaushik759 | 0 comments on Hacker News.
Hey HackerNews folks! I first launched DatingAI.pro to help people improve their dating profiles, get reply suggestions, etc but to that I got feedback where people pointed out that someone could just impersonate their entire profile which made sense and I thought what if I could actually help people really get better at texting or conversations by giving them feedback and letting them improve via suggestions. I understand a human date can't be replaced but certainly we could help to an extent. NOTE - THIS IS NOT AN AI GIRLFRIEND. This is how it works: 1. Swipe from profiles of the opposite gender and then you're matched with someone. 2. Start an engaging chat & earn or lose points based on your charm (or lack thereof). 3. Simultaneously get reply suggestions to navigate the conversation. 4. Receive a final verdict: Would they like to go out on a date with you? And the best part? Those points aren't just for bragging rights they can be redeemed them for other features in the tool. Think of it as a dating simulator with tangible rewards. Would love to know your thoughts
4 by akaushik759 | 0 comments on Hacker News.
Hey HackerNews folks! I first launched DatingAI.pro to help people improve their dating profiles, get reply suggestions, etc but to that I got feedback where people pointed out that someone could just impersonate their entire profile which made sense and I thought what if I could actually help people really get better at texting or conversations by giving them feedback and letting them improve via suggestions. I understand a human date can't be replaced but certainly we could help to an extent. NOTE - THIS IS NOT AN AI GIRLFRIEND. This is how it works: 1. Swipe from profiles of the opposite gender and then you're matched with someone. 2. Start an engaging chat & earn or lose points based on your charm (or lack thereof). 3. Simultaneously get reply suggestions to navigate the conversation. 4. Receive a final verdict: Would they like to go out on a date with you? And the best part? Those points aren't just for bragging rights they can be redeemed them for other features in the tool. Think of it as a dating simulator with tangible rewards. Would love to know your thoughts
New Show Hacker News story: latest news
Show HN: Eesel AI – ChatGPT over company knowledge, without APIs
2 by amoghs | 0 comments on Hacker News.
Hey everyone! I'm Amogh and I've built eesel AI. It's like ChatGPT over your company knowledge (yet another one?!) - but with a twist. Instead of relying on APIs to integrate with your applications, you simply use our browser extension to add relevant pages as "Knowledge". Then, you can chat over that. Whether it's a Google Doc, Notion page, help docs, or any other website - you can add it as "Knowledge" and chat with it. You don't need to grant full API access to your work apps. For public pages, we have a crawler. For private pages, we use the extension, but we're exploring a crawler for that as well. While bypassing APIs might seem unusual, we see it as an interesting approach to create a universal AI teammate. One that could, in the future, operate autonomously, accessing your applications just as you would - retrieving information and even taking actions. That's the vision. Today, 100+ companies use eesel as their "oracle" to deflect recurring questions from colleagues or customers: - A DevOps team in a 600 people org added eesel to Slack to handle engineering queries. - A Support team in a 70 people org integrated eesel as a widget on their developer documentation. - A Growth Agency in a 60 people org leverages eesel, trained on Segment, Amplitude, and Mixpanel docs, to assist engineers. - And an unusual example - a 50 people Cremation Service added eesel to their Slack as a support assistant. It'd mean a lot for you to check out eesel AI and share your thoughts about our approach.
2 by amoghs | 0 comments on Hacker News.
Hey everyone! I'm Amogh and I've built eesel AI. It's like ChatGPT over your company knowledge (yet another one?!) - but with a twist. Instead of relying on APIs to integrate with your applications, you simply use our browser extension to add relevant pages as "Knowledge". Then, you can chat over that. Whether it's a Google Doc, Notion page, help docs, or any other website - you can add it as "Knowledge" and chat with it. You don't need to grant full API access to your work apps. For public pages, we have a crawler. For private pages, we use the extension, but we're exploring a crawler for that as well. While bypassing APIs might seem unusual, we see it as an interesting approach to create a universal AI teammate. One that could, in the future, operate autonomously, accessing your applications just as you would - retrieving information and even taking actions. That's the vision. Today, 100+ companies use eesel as their "oracle" to deflect recurring questions from colleagues or customers: - A DevOps team in a 600 people org added eesel to Slack to handle engineering queries. - A Support team in a 70 people org integrated eesel as a widget on their developer documentation. - A Growth Agency in a 60 people org leverages eesel, trained on Segment, Amplitude, and Mixpanel docs, to assist engineers. - And an unusual example - a 50 people Cremation Service added eesel to their Slack as a support assistant. It'd mean a lot for you to check out eesel AI and share your thoughts about our approach.
New Show Hacker News story: latest news
Show HN: ChatGPT History Explorer
2 by kuchin | 0 comments on Hacker News.
Hey HN, As a heavy ChatGPT user, I often want to explore my chat history, look for things I've already asked, keep favorite chats, and much more. ChatGPT doesn't even offer search capability (except in mobile version). So I decided to build one myself. It runs locally, so no sensitive data leaks. You only need to export your ChatGPT conversations, and put it in the `data` folder. Optionally setup OpenAI API key and it will create embeddings for semantic search. *Important* - it's not a desktop app (at least, not yet) - it's a python code with a browser interface that I've only tested on MacOS so far. Minimal dev skills are required to set it up. Current functionality: - Github-style activity graph and some other statistics - Browse and read your conversations - Quick filters by conversation title and time period - Search (semantic and simple "strict") - Add chats to favorites - Open conversations on the ChatGPT site to continue working My plans for new features: https://ift.tt/pSLmAij Hope it will be useful to people. Would love to hear your feedback.
2 by kuchin | 0 comments on Hacker News.
Hey HN, As a heavy ChatGPT user, I often want to explore my chat history, look for things I've already asked, keep favorite chats, and much more. ChatGPT doesn't even offer search capability (except in mobile version). So I decided to build one myself. It runs locally, so no sensitive data leaks. You only need to export your ChatGPT conversations, and put it in the `data` folder. Optionally setup OpenAI API key and it will create embeddings for semantic search. *Important* - it's not a desktop app (at least, not yet) - it's a python code with a browser interface that I've only tested on MacOS so far. Minimal dev skills are required to set it up. Current functionality: - Github-style activity graph and some other statistics - Browse and read your conversations - Quick filters by conversation title and time period - Search (semantic and simple "strict") - Add chats to favorites - Open conversations on the ChatGPT site to continue working My plans for new features: https://ift.tt/pSLmAij Hope it will be useful to people. Would love to hear your feedback.
Wednesday, 13 September 2023
New Show Hacker News story: latest news
Show HN: Hands on tutorial for open source contribution
2 by sadhrietshs | 0 comments on Hacker News.
2 by sadhrietshs | 0 comments on Hacker News.
New Show Hacker News story: latest news
Show HN: C++ LMAX Disruptor < 200 LOC
2 by sneilan1 | 0 comments on Hacker News.
Hi HN, I built an implementation of the LMAX disruptor described by Martin Fowler. https://ift.tt/cLnEN6x Would love to hear your comments and suggestions!
2 by sneilan1 | 0 comments on Hacker News.
Hi HN, I built an implementation of the LMAX disruptor described by Martin Fowler. https://ift.tt/cLnEN6x Would love to hear your comments and suggestions!
Tuesday, 12 September 2023
New Show Hacker News story: latest news
Show HN: HTTP Status Code Generator
2 by armiiller | 2 comments on Hacker News.
A service for generating HTTP codes. It’s useful for testing monitoring services. Just add the status code you want to the end of the URL, like this https://ift.tt/lbpvtIx
2 by armiiller | 2 comments on Hacker News.
A service for generating HTTP codes. It’s useful for testing monitoring services. Just add the status code you want to the end of the URL, like this https://ift.tt/lbpvtIx
New Show Hacker News story: latest news
Show HN: Rental data supplied by tenants in Ireland, searchable by all
26 by vinnyglennon | 14 comments on Hacker News.
I created https://ift.tt/n5t1Gsl last Friday to help bring this kind of transparency to Ireland, allowing people to submit their rents. Would love to get any HN feedback on the idea/website.
26 by vinnyglennon | 14 comments on Hacker News.
I created https://ift.tt/n5t1Gsl last Friday to help bring this kind of transparency to Ireland, allowing people to submit their rents. Would love to get any HN feedback on the idea/website.
New Show Hacker News story: latest news
Show HN: Fine-tune your own Llama 2 to replace GPT-3.5/4
12 by kcorbitt | 2 comments on Hacker News.
There has been a lot of interest on HN in fine-tuning open-source LLMs recently (eg. Anyscale's post at https://ift.tt/w1YQrLn ). I've been playing around with fine-tuning models for a couple of years, and wanted to share some insights and practical code. I’ve condensed what I’ve learned into a small set of notebooks at https://ift.tt/xMV741P... , covering labeling data, fine-tuning, running efficient inference, and evaluating costs/performance. The 7B model we train here matches GPT-4’s labels 95% of the time on the test set, and for the 5% of cases where they disagree it’s often because the correct answer is genuinely ambiguous. What is fine-tuning? You can think of it as a more-powerful form of prompting, where instead of writing your instructions in text you actually encode them in the weights of the model itself. You do this by training an existing model on example input/output pairs that demonstrate the task you want your fine-tuned model to learn. Fine-tuning can work with as few as 50 examples but I usually try to get 1000+ if possible. Prompting still has some big advantages over fine-tuning. It's way easier/faster to iterate on your instructions than label data and re-train a model. And operationally it's easier to deploy one big model and just adjust its behavior as necessary vs deploying many small fine-tuned models that will likely each get lower utilization. Fine-tuning has one huge advantage though: it is far more effective at guiding a model's behavior than prompting, so you can often get away with a much smaller model. That gets you faster responses and lower inference costs. A fine-tuned Llama 7B model is 50x cheaper than GPT-3.5 on a per-token basis, and for many use cases can produce results that are as good or better! For example, classifying the 2M recipes at https://ift.tt/WNadCk7 with GPT-4 would cost $23k. Even with GPT-3.5 it would cost over $1k. The model we fine-tuned performs similarly to GPT-4 and costs just $19 to run over the entire dataset. Disclaimer: My brother David and I are working on an open-source product called OpenPipe to help engineers adopt fine-tuning as simply as possible. But none of the information above depends on our startup. The current post is just about sharing information that we’ve learned about fine-tuning. I hope it’s useful! If you're interested in finding out more you can watch our repo at https://ift.tt/FnfWz1Y , and if you'd just like to chat about a fine-tuning project you're thinking about you can also just email me at kyle@openpipe.ai!
12 by kcorbitt | 2 comments on Hacker News.
There has been a lot of interest on HN in fine-tuning open-source LLMs recently (eg. Anyscale's post at https://ift.tt/w1YQrLn ). I've been playing around with fine-tuning models for a couple of years, and wanted to share some insights and practical code. I’ve condensed what I’ve learned into a small set of notebooks at https://ift.tt/xMV741P... , covering labeling data, fine-tuning, running efficient inference, and evaluating costs/performance. The 7B model we train here matches GPT-4’s labels 95% of the time on the test set, and for the 5% of cases where they disagree it’s often because the correct answer is genuinely ambiguous. What is fine-tuning? You can think of it as a more-powerful form of prompting, where instead of writing your instructions in text you actually encode them in the weights of the model itself. You do this by training an existing model on example input/output pairs that demonstrate the task you want your fine-tuned model to learn. Fine-tuning can work with as few as 50 examples but I usually try to get 1000+ if possible. Prompting still has some big advantages over fine-tuning. It's way easier/faster to iterate on your instructions than label data and re-train a model. And operationally it's easier to deploy one big model and just adjust its behavior as necessary vs deploying many small fine-tuned models that will likely each get lower utilization. Fine-tuning has one huge advantage though: it is far more effective at guiding a model's behavior than prompting, so you can often get away with a much smaller model. That gets you faster responses and lower inference costs. A fine-tuned Llama 7B model is 50x cheaper than GPT-3.5 on a per-token basis, and for many use cases can produce results that are as good or better! For example, classifying the 2M recipes at https://ift.tt/WNadCk7 with GPT-4 would cost $23k. Even with GPT-3.5 it would cost over $1k. The model we fine-tuned performs similarly to GPT-4 and costs just $19 to run over the entire dataset. Disclaimer: My brother David and I are working on an open-source product called OpenPipe to help engineers adopt fine-tuning as simply as possible. But none of the information above depends on our startup. The current post is just about sharing information that we’ve learned about fine-tuning. I hope it’s useful! If you're interested in finding out more you can watch our repo at https://ift.tt/FnfWz1Y , and if you'd just like to chat about a fine-tuning project you're thinking about you can also just email me at kyle@openpipe.ai!
New Show Hacker News story: latest news
Show HN: Permify Now Supports Attribute-Based Access Control
12 by freddgn | 0 comments on Hacker News.
Hey Everyone, Almost year ago we launched Permify an open-source authorization service to build fine-grained and scalable user permissions and access control systems over here( https://ift.tt/y4QCGE6 ) My co-founders and I had a enterprise solution agency where we worked with a bunch of fortune 500 companies. Each projects we made we have to re-invent the wheel for access control and authorization. Not having a easy to integrate, scalable and a granular system that can fulfill the requirements of these enterprise companies was a pain in the ass for us. As well as, platform engineers and software architects in these teams as well. So, we start working on a centralized authorization piece which makes authorization less scary for these architects. We learned a lot from google-zanzibar, OPA, and XACML. Yet we wanted make Permify unique. Permify become a true ReBAC system where you can create a structured authorization logic. But there was a missing piece in which makes you build attribute based access rights such as based on date, number of times, amount of money, etc. Today we are launching that piece with Permify ABAC. We would love to get your feedback on this!!! Feel free roast or toast :)
12 by freddgn | 0 comments on Hacker News.
Hey Everyone, Almost year ago we launched Permify an open-source authorization service to build fine-grained and scalable user permissions and access control systems over here( https://ift.tt/y4QCGE6 ) My co-founders and I had a enterprise solution agency where we worked with a bunch of fortune 500 companies. Each projects we made we have to re-invent the wheel for access control and authorization. Not having a easy to integrate, scalable and a granular system that can fulfill the requirements of these enterprise companies was a pain in the ass for us. As well as, platform engineers and software architects in these teams as well. So, we start working on a centralized authorization piece which makes authorization less scary for these architects. We learned a lot from google-zanzibar, OPA, and XACML. Yet we wanted make Permify unique. Permify become a true ReBAC system where you can create a structured authorization logic. But there was a missing piece in which makes you build attribute based access rights such as based on date, number of times, amount of money, etc. Today we are launching that piece with Permify ABAC. We would love to get your feedback on this!!! Feel free roast or toast :)
New Show Hacker News story: latest news
Show HN: Embeddable DocuSign alternative via native Web Components
3 by somery | 3 comments on Hacker News.
3 by somery | 3 comments on Hacker News.
Monday, 11 September 2023
New Show Hacker News story: latest news
Show HN: I built Wuf, mobile notifications for all your needs
4 by ikoichi2112 | 1 comments on Hacker News.
Hi, I am Luca software engineer and indie hacker. For one of my SaaS products, I had the need to add mobile push notifications to send specific messages to the users (like new messages received, support requests, and so on). But that meant building and releasing a mobile app for iOS and Android. So I thought that I could create a third-party product, to provide exactly that service. The users install the Wuf app. They send you their User Key generated upon sign-up. You send them push mobile notifications via HTTP POST. I'd love to hear your feedback on it, and if you want to become an early user, I'd be happy to welcome you Luca
4 by ikoichi2112 | 1 comments on Hacker News.
Hi, I am Luca software engineer and indie hacker. For one of my SaaS products, I had the need to add mobile push notifications to send specific messages to the users (like new messages received, support requests, and so on). But that meant building and releasing a mobile app for iOS and Android. So I thought that I could create a third-party product, to provide exactly that service. The users install the Wuf app. They send you their User Key generated upon sign-up. You send them push mobile notifications via HTTP POST. I'd love to hear your feedback on it, and if you want to become an early user, I'd be happy to welcome you Luca
New Show Hacker News story: latest news
Show HN: Which of these two wiki pages exist in the most languages?
2 by cpa | 2 comments on Hacker News.
A very simple game where you're presented two wiki (en) pages and have to guess which one exists in the most languages. It's not that easy and stringing the code together wasn't either (wikidata's sparql is a — powerful — nightmare). repo: https://ift.tt/pX8TfsQ
2 by cpa | 2 comments on Hacker News.
A very simple game where you're presented two wiki (en) pages and have to guess which one exists in the most languages. It's not that easy and stringing the code together wasn't either (wikidata's sparql is a — powerful — nightmare). repo: https://ift.tt/pX8TfsQ
New Show Hacker News story: latest news
Show HN: Mavex.ai – Your Personal AI Executive Assistant
2 by yednap868 | 2 comments on Hacker News.
Mavy is your personal AI Executive Assistant which helps in scheduling and calendar management. Looking for early adopters.
2 by yednap868 | 2 comments on Hacker News.
Mavy is your personal AI Executive Assistant which helps in scheduling and calendar management. Looking for early adopters.
New Show Hacker News story: latest news
Show HN: Slotmachine, to book and free server ports at scale
2 by cyansmoker | 0 comments on Hacker News.
This is a very specialized Go library and, frankly, somewhat of a straw man: I've found myself needing to quickly assign ports on servers with a high density of highly ephemeral inbound/outbound connections where the protocol relies on a specific port number. Obviously, running sequentially through a list of ports until I find one I have not already allocated does not perform very well. I was surprised when I did not find anything already addressing this need. So, this is the straw man part: surely I missed it! Anyway, should you ever need something like that... here it is.
2 by cyansmoker | 0 comments on Hacker News.
This is a very specialized Go library and, frankly, somewhat of a straw man: I've found myself needing to quickly assign ports on servers with a high density of highly ephemeral inbound/outbound connections where the protocol relies on a specific port number. Obviously, running sequentially through a list of ports until I find one I have not already allocated does not perform very well. I was surprised when I did not find anything already addressing this need. So, this is the straw man part: surely I missed it! Anyway, should you ever need something like that... here it is.
New Show Hacker News story: latest news
Show HN: Free High-quality TailwindCSS Components. No attribution required
3 by andyydao | 0 comments on Hacker News.
3 by andyydao | 0 comments on Hacker News.
Sunday, 10 September 2023
New Show Hacker News story: latest news
Show HN: Erlmacs – a script to update your .emacs file for Erlang development
7 by dlachausse | 0 comments on Hacker News.
erlmacs automatically configures and updates your .emacs file with support for the emacs mode that is included with Erlang/OTP. It frees you from having to locate the installation directory of Erlang/OTP and its bundled emacs mode. It is an escript that only depends upon Erlang/OTP and Emacs. Note: There is not much in the way of error checking at this moment, but it does make a backup of your .emacs files before any destructive operations.
7 by dlachausse | 0 comments on Hacker News.
erlmacs automatically configures and updates your .emacs file with support for the emacs mode that is included with Erlang/OTP. It frees you from having to locate the installation directory of Erlang/OTP and its bundled emacs mode. It is an escript that only depends upon Erlang/OTP and Emacs. Note: There is not much in the way of error checking at this moment, but it does make a backup of your .emacs files before any destructive operations.
Saturday, 9 September 2023
New Show Hacker News story: latest news
Show HN: Productonboarding.com – Mobbin for SaaS product onboarding
3 by brownrout | 0 comments on Hacker News.
Hey Hackernews, Eric here! Wanted to share a new website we just built called productonboarding.com (Next.js and RSC). The site has screenshots and videos of product onboarding from companies like Figma, Notion, Framer, and more. It’s sort of like Mobbin for web-based product onboarding. We build a lot of product onboarding at our startup Frigade, and over the last year we’ve put together an internal library of hundreds of product onboarding examples that we refer to all the time with customers. It helps them find and copy patterns that work at other companies so they don’t need to create net new experiences or A/B test their way into the best performing pattern from scratch. Given it's been so useful to us, we decided to open it up to the world. We bought productonboarding.com and have started adding examples from our collection and made them browsable and sortable. We’re planning to add new examples weekly. Hope this is helpful to any of you who are currently thinking through building new onboarding experiences. Would love any ideas and feedback. Thanks!
3 by brownrout | 0 comments on Hacker News.
Hey Hackernews, Eric here! Wanted to share a new website we just built called productonboarding.com (Next.js and RSC). The site has screenshots and videos of product onboarding from companies like Figma, Notion, Framer, and more. It’s sort of like Mobbin for web-based product onboarding. We build a lot of product onboarding at our startup Frigade, and over the last year we’ve put together an internal library of hundreds of product onboarding examples that we refer to all the time with customers. It helps them find and copy patterns that work at other companies so they don’t need to create net new experiences or A/B test their way into the best performing pattern from scratch. Given it's been so useful to us, we decided to open it up to the world. We bought productonboarding.com and have started adding examples from our collection and made them browsable and sortable. We’re planning to add new examples weekly. Hope this is helpful to any of you who are currently thinking through building new onboarding experiences. Would love any ideas and feedback. Thanks!
New Show Hacker News story: latest news
Show HN: Which is faster? Puppeteer, Playwright or Selenium
3 by eneuman | 0 comments on Hacker News.
Hey Everyone, I just ran a [rather silly] race between Puppeteer (JS), Playwright (Python) and Selenium (Python) to see which one would be fastest on a simple scrape (using Google Colab so you can also run it) Far from a comprehensive benchmark, this race is 100% free from advanced configurations, multi-threading or anything complicated. It just opens Wallapop (a second hand marketplace in Spain) and times how long it takes to extract the first 2000 results of a search. If you like this simple format, have any ideas on how to improve a race like this or have a strong urge to prove Ward Cunningham wright, let me know in the comments!
3 by eneuman | 0 comments on Hacker News.
Hey Everyone, I just ran a [rather silly] race between Puppeteer (JS), Playwright (Python) and Selenium (Python) to see which one would be fastest on a simple scrape (using Google Colab so you can also run it) Far from a comprehensive benchmark, this race is 100% free from advanced configurations, multi-threading or anything complicated. It just opens Wallapop (a second hand marketplace in Spain) and times how long it takes to extract the first 2000 results of a search. If you like this simple format, have any ideas on how to improve a race like this or have a strong urge to prove Ward Cunningham wright, let me know in the comments!
Friday, 8 September 2023
New Show Hacker News story: latest news
Show HN: AI Copilot for the Stock Market
2 by stocknear | 4 comments on Hacker News.
Any feedback is much much appreciated
2 by stocknear | 4 comments on Hacker News.
Any feedback is much much appreciated
New Show Hacker News story: latest news
Show HN: Vanna AI – Open-Sourced Text-to-SQL in Python
4 by zainhoda | 1 comments on Hacker News.
Hey there HN! We've just open-sourced Vanna – a Python package that allows you to transform questions into SQL. We've leveraged LLMs to enable you to "ask" databases what you need, bypassing the need to "write" complex SQL. Quick Overview: - "Train" using DDL statements, documentation, or known correct SQL statements. - "Ask" questions in natural language and receive SQL, tables, and charts in return. - Open Source Flexibility: Swap storage mechanisms, customize LLMs, and choose your databases. - Local or Hosted: Operate everything locally or use our hosted version for free (including complimentary LLM calls). - Use it wherever Python is applicable. We provide code examples for integration in Jupyter, Streamlit, Slack, and more. We would greatly appreciate your feedback, insights on issues, and contributions: https://ift.tt/RCkgdtm
4 by zainhoda | 1 comments on Hacker News.
Hey there HN! We've just open-sourced Vanna – a Python package that allows you to transform questions into SQL. We've leveraged LLMs to enable you to "ask" databases what you need, bypassing the need to "write" complex SQL. Quick Overview: - "Train" using DDL statements, documentation, or known correct SQL statements. - "Ask" questions in natural language and receive SQL, tables, and charts in return. - Open Source Flexibility: Swap storage mechanisms, customize LLMs, and choose your databases. - Local or Hosted: Operate everything locally or use our hosted version for free (including complimentary LLM calls). - Use it wherever Python is applicable. We provide code examples for integration in Jupyter, Streamlit, Slack, and more. We would greatly appreciate your feedback, insights on issues, and contributions: https://ift.tt/RCkgdtm
New Show Hacker News story: latest news
Show HN: Performant Rust Streaming Algorithms from Python
5 by amath | 0 comments on Hacker News.
The author also wrote a blog post on how to use it with Bytewax, a Python stream processor, to read and analyze Mastadon posts. https://ift.tt/KQhfjgb...
5 by amath | 0 comments on Hacker News.
The author also wrote a blog post on how to use it with Bytewax, a Python stream processor, to read and analyze Mastadon posts. https://ift.tt/KQhfjgb...
New Show Hacker News story: latest news
Show HN: Trickle – Let GPT-4 Understand Your Screenshots
7 by jarodxu | 2 comments on Hacker News.
Hey there HN! It took nearly 5 years for my team and me to truly find right direction. So, after introducing our work, I'm keen to share the story behind it. The following content is divided into two parts. If you're not interested in the backstory of the product, feel free to skip the content after the divider. > The Problem We're Solving: During a casual afternoon, while brainstorming what to do next on a WeWork sofa, we realized that almost everyone present had a habit of saving information via screenshots. When I opened my photo gallery, I was astonished to see that more than half were such screenshots. Given that traditional OCR and gallery apps hadn't really addressed our screenshot chaos, we decided to build something to solve our own problem. > How it works: At first glance, you might think Trickle is a manual screenshot version of Rewind. But in reality, they're vastly different. All you need to do is send your screenshots, and let Trickle handle the rest: [1] Trickle doesn't constantly record the entire desktop, so it won't consume all your Mac's storage or affect its performance. Moreover, it won't give you the unsettling feeling of being constantly watched. [2] Although we have a Mac screenshot tool, and a browser extension is on its way, you don't actually need to install them. You can easily upload your screenshots via a web page. This makes it platform-agnostic; you can browse, search, and ask about your historical screenshots at any time via a browser. Of course, it doesn't occupy any of your local storage. Last but not least, it's Windows-user friendly. [3] User-controlled screenshots mean that the embedded chunks are semantically more precise. Beyond the advanced reasoning capabilities of GPT-4, we've integrated some tricks of our own, allowing Trickle to truly comprehend your screenshots, rather than just summarizing the text. Sometimes you might be surprised when Trickle even reasons out essential information that's not present in the image. This also ensures a better experience when you try to recall information. ------------The Story Behind the Product:------------ In 2018, I left my consulting role and dove head-first into the startup world with two co-founders. Our initial venture involved creating a visual recognition model for a vending machine company, marking our first income. Yet, custom builds weren’t sustainable, prompting our first pivot. By 2019, we were deep into retail tech, winning a demo day and launching a product to automate in-store promotions. The climax seemed to be our partnership with a large multinational, but 2020 and the pandemic shifted landscapes. We then explored the realm of asynchronous video, building an alternative to Loom. By 2021, we hit 10k users, but profitability remained elusive. As workspaces evolved and people returned to offices, we sensed another opportunity. Our solution? An in-house social platform for teams, named "Trickle". In 2022, after a launch on Product Hunt, even though we garnered attention from people like Ryan Hoover and former Microsoft VP Lu Qi, we faced a stark realization: Daily team updates might not be as 'cool' as anticipated. Our anxiety steered us off-path, and soon, Trickle became a bulky hybrid, attempting to replace tools like Notion and Slack. Looking back, this detour was a misstep. The turning point came mid-2023. A series of tepid Product Hunt launches forced introspection. We stripped Trickle down, retaining only its name and began anew. Today, with the original team intact, we're addressing an everyday issue that resonated with all of us: the chaos of managing screenshots. Reflecting on our journey, it's clear that the essence of the startup spirit lies in adapting, evolving, and pursuing that 'Eureka' moment, no matter how winding the path.
7 by jarodxu | 2 comments on Hacker News.
Hey there HN! It took nearly 5 years for my team and me to truly find right direction. So, after introducing our work, I'm keen to share the story behind it. The following content is divided into two parts. If you're not interested in the backstory of the product, feel free to skip the content after the divider. > The Problem We're Solving: During a casual afternoon, while brainstorming what to do next on a WeWork sofa, we realized that almost everyone present had a habit of saving information via screenshots. When I opened my photo gallery, I was astonished to see that more than half were such screenshots. Given that traditional OCR and gallery apps hadn't really addressed our screenshot chaos, we decided to build something to solve our own problem. > How it works: At first glance, you might think Trickle is a manual screenshot version of Rewind. But in reality, they're vastly different. All you need to do is send your screenshots, and let Trickle handle the rest: [1] Trickle doesn't constantly record the entire desktop, so it won't consume all your Mac's storage or affect its performance. Moreover, it won't give you the unsettling feeling of being constantly watched. [2] Although we have a Mac screenshot tool, and a browser extension is on its way, you don't actually need to install them. You can easily upload your screenshots via a web page. This makes it platform-agnostic; you can browse, search, and ask about your historical screenshots at any time via a browser. Of course, it doesn't occupy any of your local storage. Last but not least, it's Windows-user friendly. [3] User-controlled screenshots mean that the embedded chunks are semantically more precise. Beyond the advanced reasoning capabilities of GPT-4, we've integrated some tricks of our own, allowing Trickle to truly comprehend your screenshots, rather than just summarizing the text. Sometimes you might be surprised when Trickle even reasons out essential information that's not present in the image. This also ensures a better experience when you try to recall information. ------------The Story Behind the Product:------------ In 2018, I left my consulting role and dove head-first into the startup world with two co-founders. Our initial venture involved creating a visual recognition model for a vending machine company, marking our first income. Yet, custom builds weren’t sustainable, prompting our first pivot. By 2019, we were deep into retail tech, winning a demo day and launching a product to automate in-store promotions. The climax seemed to be our partnership with a large multinational, but 2020 and the pandemic shifted landscapes. We then explored the realm of asynchronous video, building an alternative to Loom. By 2021, we hit 10k users, but profitability remained elusive. As workspaces evolved and people returned to offices, we sensed another opportunity. Our solution? An in-house social platform for teams, named "Trickle". In 2022, after a launch on Product Hunt, even though we garnered attention from people like Ryan Hoover and former Microsoft VP Lu Qi, we faced a stark realization: Daily team updates might not be as 'cool' as anticipated. Our anxiety steered us off-path, and soon, Trickle became a bulky hybrid, attempting to replace tools like Notion and Slack. Looking back, this detour was a misstep. The turning point came mid-2023. A series of tepid Product Hunt launches forced introspection. We stripped Trickle down, retaining only its name and began anew. Today, with the original team intact, we're addressing an everyday issue that resonated with all of us: the chaos of managing screenshots. Reflecting on our journey, it's clear that the essence of the startup spirit lies in adapting, evolving, and pursuing that 'Eureka' moment, no matter how winding the path.
New Show Hacker News story: latest news
Show HN: ChatGPT Powered Code Analysis
3 by llmllmllm | 0 comments on Hacker News.
I'm building an AI platform, FlowChai and found a neat use case for it today that I thought would be useful to HN readers. I use GPT4 heavily for writing / editing code, but a major downside is that it doesn't know about new projects. I made the connection today that I could upload the zip file of a Github repo to FlowChai and then write prompts just like with ChatGPT with code questions. While the original intent for this platform is more around natural language, it's neat how this works so well. It's powered underneath by pgvector and OpenAI embeddings.
3 by llmllmllm | 0 comments on Hacker News.
I'm building an AI platform, FlowChai and found a neat use case for it today that I thought would be useful to HN readers. I use GPT4 heavily for writing / editing code, but a major downside is that it doesn't know about new projects. I made the connection today that I could upload the zip file of a Github repo to FlowChai and then write prompts just like with ChatGPT with code questions. While the original intent for this platform is more around natural language, it's neat how this works so well. It's powered underneath by pgvector and OpenAI embeddings.
New Show Hacker News story: latest news
Show HN: Passport Protocol – Programmable and trustless key management network
2 by fearlessboi | 0 comments on Hacker News.
Passport Protocol is a programmable and MPC-based distributed key management network. Simply put, we make it easy for developers and companies to securely store, manage, and program keys in a distributed manner. Passport Protocol has usecases in both web3 and web2. For web2 usecase, we essentially provide a performant but trustless and programmable alternative to traditional key management infrastructure. For web3, we make it easy to easily spin up wallets based on custom authentication rules and even schedule and automate transactions. If this sounds interesting, sign up and we'll help you get started! Website - https://ift.tt/52Rh8ck
2 by fearlessboi | 0 comments on Hacker News.
Passport Protocol is a programmable and MPC-based distributed key management network. Simply put, we make it easy for developers and companies to securely store, manage, and program keys in a distributed manner. Passport Protocol has usecases in both web3 and web2. For web2 usecase, we essentially provide a performant but trustless and programmable alternative to traditional key management infrastructure. For web3, we make it easy to easily spin up wallets based on custom authentication rules and even schedule and automate transactions. If this sounds interesting, sign up and we'll help you get started! Website - https://ift.tt/52Rh8ck
Thursday, 7 September 2023
New Show Hacker News story: latest news
Show HN: Aisolarpunk – Explore thousands of AI generated SolarPunk pictures
1 by justspamjustin | 0 comments on Hacker News.
1 by justspamjustin | 0 comments on Hacker News.
New Show Hacker News story: latest news
Show HN: Dweets.com – Microblogging for Devs
2 by raymondmoay | 0 comments on Hacker News.
Built this to microblog everything I am learning as a dev. It's an MVP, lots to improve on! Would love some feedback. (: Also am looking for a job (preferably remote since I'm in SG), 1.5 years in as a Jr Dev at a local startup, self-taught by building stuff. I'll be on Twitter if you're interested! @raymondmoay
2 by raymondmoay | 0 comments on Hacker News.
Built this to microblog everything I am learning as a dev. It's an MVP, lots to improve on! Would love some feedback. (: Also am looking for a job (preferably remote since I'm in SG), 1.5 years in as a Jr Dev at a local startup, self-taught by building stuff. I'll be on Twitter if you're interested! @raymondmoay
Subscribe to:
Posts (Atom)