Show HN: Semantic Grep – A Word2Vec-powered search tool https://ift.tt/dLit0kQ

Show HN: Semantic Grep – A Word2Vec-powered search tool Much improved new version. Search for words similar to the query. For example, "death" will find "death", "dying", "dead", "killing"... Incredibly useful for exploring large text datasets where exact matches are too restrictive. https://ift.tt/28Ivf6i July 27, 2024 at 11:32PM

Show HN: Symbols > We are building Figma for developers https://ift.tt/0OdZWje

Show HN: Symbols > We are building Figma for developers What is Symbols? “The ultimate platform for developers & teams to build, test and document fully functional & reusable UI libraries on a canvas. Publish as a website or export to your existing tech stack with open-source development” Lifetime offers: We are currently running $100 one off offers for the first 100 users, which includes unlimited projects or users for 1 project. Eventually charging $49 per month per user/project, as design system platforms charge a similar amount. So grab a deal! Landing: Here is a quick landing put together, with a better one in the works :) https://ift.tt/7nKS1iA Any questions let me know! https://ift.tt/7nKS1iA July 27, 2024 at 01:47AM

Show HN: Check Your Mouthbreathing with MediaPipe https://ift.tt/nO2N4JM

Show HN: Check Your Mouthbreathing with MediaPipe While scrolling through Reddit, I read about the negative side effects of mouth breathing (which are quite debatable) and wondered if I was doing it myself. Since I wanted to experiment with MediaPipe anyway, I built a little website to check. It's just an index HTML file which you can download and modify. It only works on desktop and takes a few seconds to load MediaPipe. https://ift.tt/TKt0BwG July 27, 2024 at 02:56AM

Show HN: Ray Tracing in One Weekend v4.0.0 https://ift.tt/lIyRcMv

Show HN: Ray Tracing in One Weekend v4.0.0 Since this is a major new release (three and a half years in development), I think this should be ok for Show HN. This release has lots of new material, fixes, and updates across the three books in this series. All three books are online and free, with accompanying code available from GitHub. Enjoy! https://ift.tt/YOzj3lQ July 27, 2024 at 02:55AM

Show HN: Patchwork – Open-source framework to automate development gruntwork https://ift.tt/tjy9N8W

Show HN: Patchwork – Open-source framework to automate development gruntwork Hi HN! We’re Asankhaya and Rohan and we are building Patchwork. Patchwork tackles development gruntwork—like reviews, docs, linting, and security fixes—through customizable, code-first 'patchflows' using LLMs and modular code management steps, all in Python. Here's a quick overview video: https://youtu.be/MLyn6B3bFMU From our time building DevSecOps tools, we experienced first-hand the frustrations our users faced as they built complex delivery pipelines. Almost a third of developer time is spent on code management tasks[1], yet backlogs remain. Patchwork lets you combine well-defined prompts with effective workflow orchestration to automate as much as 80% of these gruntwork tasks using LLMs[2]. For instance, the AutoFix patchflow can resolve 82% of issues flagged by semgrep using gpt-4 (or 68% with llama-3.1-8B) without fine-tuning or providing specialized context [3]. Success rates are higher for text-based patchflows like PR Review and Generate Docstring, but lower for more complex tasks like Dependency Upgrades. We are not a coding assistant or a black-box GitHub bot. Our automation workflows run outside your IDE via the CLI or CI scripts without your active involvement. We are also not an ‘AI agent’ framework. In our experience, LLM agents struggle with planning and rarely identify the right execution path. Instead, Patchwork requires explicitly defined workflows that provide greater success and full control. Patchwork is open-source so you can build your own patchflows, integrate your preferred LLM endpoints, and fully self-host, ensuring privacy and compliance for large teams. As devs, we prefer to build our own ‘AI-enabled automation’ given how easy it is to consume LLM APIs. If you do, try patchwork via a simple 'pip install patchwork-cli' or find us on Github[4]. Sources: [1] https://ift.tt/uYesSFO... [2] https://ift.tt/MT2Wldg... [3] https://ift.tt/olt1TC8 [4] https://ift.tt/hHfJmen [Sample PRs] https://ift.tt/GXHsYao https://ift.tt/hHfJmen July 27, 2024 at 02:04AM

Show HN: A personalised AI tutor with < 1s voice responses https://news.ycombinator.com/item?id=41057030

Show HN: A personalised AI tutor with < 1s voice responses TLDR: We created a personalised Andrej Karpathy tutor that can response to questions about his Youtube videos in sub 1 second responses (voice-to-voice). We do this using a voice enabled RAG agent. See later in the post for demo link, Github Repo and blog write up. A few weeks ago we released the worlds fastest voice bot, achieving 500ms voice-to-voice response times, including a 200ms delay waiting for a user to stop speaking. After reaching the front page of HN, we thought about how we could take this a step further based on feedback we were getting from the community. Many companies were looking for a way to implement function calling and RAG with voice interfaces while retaining a low enough latency. We couldn’t find many resources about how to do this online that: 1. Allowed us to achieve sub-second voice-to-voice latency 2. Was more flexible than existing solutions. Vapi, Retell, [Bland.ai]( http://Bland.ai ) are too opinionated plus since they just orchestrate API’s which incur network latency at every step. See requirement above 3. The unit economics actually work at scale. So we decided to create a implementation of our own. Process: As we mentioned in our previous release, if you want to achieve response times this low you need to make everything as local as possible. So below was our setup - Local STT: Deepgram model - Local Embedding model: Nomic v1.5 - Local VectorDB: Turso - Local LLM: Llama 3B - Local TTS: Deepgram model From our previous example, the only new components where: - Local Embedding model: We chose Nomic Embed text v1.5 model that gave a processing time of roughly ~200ms - Turso offers local embedded replicas combined with edgeDB’s which meant we were able to achieve 0.01 second read times. Pinecone also gave us good times of 0.043 seconds. The above changes led us to achieve sub 1 second voice-to-voice response times Application: With Andrej Karpathy’s announcement around [Eureka Labs]( https://eurekalabs.ai/ ), a new AI+Education company we thought we would create our very own personalised Andrej tutor. Listen to anyone of his Youtube lectures, as soon as your start specking, the video will pause and he will reply. Once your question has been answered you can then tell him to continue with the lecture and the video will automatically start playing. Demo: https://educationbot.cerebrium.ai/ Blog: https://www.cerebrium.ai/blog/creating-a-realtime-rag-voice-... Github Repo: https://github.com/CerebriumAI/examples/tree/master/19-voice... For demo purposes: - We used OpenAI for GPT-4-mini and embeddings (its cheaper to run on a CPU than GPU’s when running demos at scale. These changes add about ~1 second to the response time - We used Eleven labs to clone his voice to make replies sound more realistic. This adds about 300ms to the response time. The improvements that can be made which we would like the community to contribute to are: - Embed the video screens as well that when you ask certain questions it can show you the relevant lecture slide for the same chuck that it got context from to answer. - Insert the timestamps in the vectorDB timestamps so that if a question will be answered later in the lecture he can let you know This unlocks so many use cases in education, employee training, sales etc that it would be great to see what the community builds! https://educationbot.cerebrium.ai/ July 24, 2024 at 07:11PM

Show HN: Magic Color Picker https://ift.tt/Er2F1Q0

Show HN: Magic Color Picker https://ift.tt/RbSlQg3 July 26, 2024 at 12:26AM

Show HN: Wat – Deep inspection of Python objects https://ift.tt/ioHguj5

Show HN: Wat – Deep inspection of Python objects https://ift.tt/lmZ9Rxf July 25, 2024 at 09:48PM

Show HN: Hooper – AI-driven stats and highlights for basketball play https://ift.tt/HQ6L9ZD

Show HN: Hooper – AI-driven stats and highlights for basketball play Hey everyone, OP here. Wanted to share a bit more about Hooper — I started building it with a good friend of mine six months ago. We play a lot of pickup together and were arguing about who has a better jump shot and ended up hacking together an app to settle it The way Hooper works is you can record yourself using the app and ideally a tripod (optional). The app will track everyone, whether its a solo practice, a 3v3, or a 5v5. We think there’s a lot of stuff out there for basketball drills but what we really wanted Hooper to be for is actual game play. That means, it can do things like track multiple players, sync two half court recordings, and differentiate 2s vs 3s. Once you finish recording, it’ll process for a bit and then ask you to tag yourself (and optionally other players). Then it spits back out a few things: you can watch the full footage as well as a clipped version that's only the interesting plays; you can see highlights and offensive box stats for every player. When you sign up, you get a Hooper profile that tracks your overall stats across the sessions. You can also do things like build a mixtape from your highlights for insta. You can friend other players on Hooper to see their profile & comment on their games (also add us “grub” and “kangexpress”!) We’ve been in a closed beta for about 3 months now, fixing bugs and getting things to work with a set of early adopters. We are now starting our open beta ! If anyone here wants to try it out, you can just download the app on https://ift.tt/yXFpV7E . We are early in our journey, and lots of improvements to be made in the next few months, but we would love your feedback and ideas! https://www.hooper.gg July 25, 2024 at 03:09AM

Show HN: Voluntarily add warning labels to social media websites https://ift.tt/dQUKeHY

Show HN: Voluntarily add warning labels to social media websites A Chrome extension to voluntarily add informative warning labels to social media websites. There is a significant association between social media use and depression. Recently the surgeon general of the United States called on congress to require warning labels. While warning labels are not a sure-fire way to curb use, they can be effective provided that: - The warning must provide new information to users, and - The user must find the information credible https://ift.tt/qInQ7ja July 25, 2024 at 02:24AM

Show HN: Unsearch, an open source alternative to Google Chrome Sync https://ift.tt/BopPRWq

Show HN: Unsearch, an open source alternative to Google Chrome Sync I use several browsers between college, work and home and I like to have them synced so I can find websites I've searched for in a browser when I don't have access to it or access my bookmarks. I don't like the idea of using any of the sync services offered by the major browsers because of how they might use my data, because they lack some features I'd like to have, and because they all try to lock you into their ecosystem. So I'm building an open source cross-browser manager for browsing activity. I would love to hear your thoughts on this concept, and if you have any questions or suggestions, please let me know. https://ift.tt/E4BoPft July 24, 2024 at 10:48PM

Show HN: I built a client side Chrome extension to manage your ChatGPT prompts https://ift.tt/OByex1o

Show HN: I built a client side Chrome extension to manage your ChatGPT prompts https://www.promptzilla.xyz/ July 24, 2024 at 06:55AM

Show HN: Finetune Llama-3.1 2x faster in a Colab https://ift.tt/ojHv43E

Show HN: Finetune Llama-3.1 2x faster in a Colab Just added Llama-3.1 support! Unsloth https://ift.tt/oCArEX3 makes finetuning Llama, Mistral, Gemma & Phi 2x faster, and use 50 to 70% less VRAM with no accuracy degradation. There's a custom backprop engine which reduces actual FLOPs, and all kernels are written in OpenAI's Triton language to reduce data movement. Also have an 2x faster inference only notebook in a free Colab as well! https://ift.tt/A4QLxbC... https://ift.tt/GhqBKvy July 24, 2024 at 02:00AM

Show HN: React Native Finisher Kit That'll Automate All the Complex Stuff https://ift.tt/9EmHA5i

Show HN: React Native Finisher Kit That'll Automate All the Complex Stuff Hey HN, I've been automating mobile builds since 2017, publishing a course on mobile automation back then. Over the years, I’ve configured Mobile CI/CD for various companies using Bitrise, Codemagic, GitHub Actions – over and over again. To simplify this repetitive process, I’ve created a React Native "finisher kit" that automates Mobile CI/CD setup. Here’s what it offers: - Initial setup with monorepo support, multiple environments, unified configuration, testing set up, features like Push or iOS Live Activities… - Easy app creation for App Stores + Fully automated code signing, building, deployment, - Single place for configuration (AWS), reusable for local development, builds and deployment either locally or in the cloud, - A seamless system to manage configurations for Apple, Fastlane, Google, and more… It allows you to create your Mobile CI/CD with any of these providers with a single CLI command and keep it up-to-date. Switch between providers just by changing the config. It supports both React Native CLI projects and Expo. I built it to save time and reduce the complexity of mobile app development, hoping it will be as helpful for you as it has been for me. I’d love to get your feedback and thoughts. Thanks! Igor https://ift.tt/8NUaCzs July 23, 2024 at 11:12PM

Show HN: Formula 1, 2, 3 and E Weather updates https://ift.tt/iUszQ0m

Show HN: Formula 1, 2, 3 and E Weather updates Get up to date predicted weather for the upcoming Formula 1, 2, 3 and E sessions! https://ift.tt/auNWeph July 23, 2024 at 11:57PM

Show HN: Easily map CSV data with lat/lon to H3 for enrichment or aggregation https://ift.tt/hdyGXOk

Show HN: Easily map CSV data with lat/lon to H3 for enrichment or aggregation https://ift.tt/6N2AJmn July 23, 2024 at 04:08AM

Show HN: I made a task manager / calendar app to use it with my wife https://ift.tt/c29EIsi

Show HN: I made a task manager / calendar app to use it with my wife Hi HN, Two years ago, I started building Jinear as a side project for fun. Then my wife started using it to plan her PhD thesis. I added features as we needed them. After a while, my best friend started using it in his small company, so I thought maybe I could productize it. You can create tasks, set reminders, attach files, link your Google Calendar, etc. It can be used as a PWA. We've been using it daily. I would be happy to receive your feedback. Thanks. https://jinear.co July 22, 2024 at 10:04PM

Show HN: I packaged all of the productivity advice into one product https://ift.tt/K1IFziv

Show HN: I packaged all of the productivity advice into one product Hey hackers, Like many of you, I'm always trying to optimize my productivity, and I have tried a lot of apps out there to do so. I didn't find the exact one that I could use, so I built one for myself. Check it out and let me know if you have any thoughts! https://www.focusmax.io July 22, 2024 at 08:54PM

Show HN: Shade/Bs – Modern Web UIs Without Node.js https://ift.tt/skNq37b

Show HN: Shade/Bs – Modern Web UIs Without Node.js https://ift.tt/YX1lJ8s July 22, 2024 at 04:58AM

Show HN: Create how-to videos and guides fast https://ift.tt/UXI5GeY

Show HN: Create how-to videos and guides fast Hey HN, I'm Kamal, a 20 year old builder from India, with a team of 4. I was building an AI-powered course builder initially, and once we started getting some paying customers biggest problem that we encountered was teaching them how to use the product. I wanted to create help centre but I couldn't, probably the procrastination, because I felt like it was a huge task. So, me and my co-founder decided to make creating videos and guides for this help centre fast and easy. This is where Kroto comes in, you just record a product/process walkthrough and it generates studio quality how-to videos with zoom-in and transition effects along with a step-by-step guide with GIFs for every action. Here's the demo: https://www.youtube.com/watch?v=JmeeNmpNepY I want to know if you guys also face similar issues, and want to get some feedback on the product. Biggest issues right now: Publishing time way too long, editor not optimised, no way to remove or add more zoom-ins in the video. https://www.kroto.one July 21, 2024 at 10:13PM