Wednesday, July 31, 2024

New top story on Hacker News: Heavy metal analysis of dark chocolate and cocoa products in the USA

Heavy metal analysis of dark chocolate and cocoa products in the USA
27 by gnabgib | 9 comments on Hacker News.


New top story on Hacker News: Fast Multidimensional Matrix Multiplication on CPU from Scratch

Fast Multidimensional Matrix Multiplication on CPU from Scratch
9 by georgehill | 1 comments on Hacker News.


New top story on Hacker News: Ask HN: Best Tools for Monorepo?

Ask HN: Best Tools for Monorepo?
7 by bradhe | 3 comments on Hacker News.
I've got a monorepo I'm working in that has a Golang backend with a couple services and a Next.js front-end. Everything lives in a monorepo together. My tooling is super weak, though! For instance, for process management in development I'm using Goreman, which is a Foreman alternative in Goalng. Wondering what's the state of the art for managing the processes in local dev in monorepos in 2024? Or other tools for managing a monorepo I might be missing in general!

Sunday, July 28, 2024

New top story on Hacker News: Show HN: I built an open-source tool to make on-call suck less

Show HN: I built an open-source tool to make on-call suck less
18 by aray07 | 2 comments on Hacker News.
Hey HN, I am building an open source platform to make on-call better and less stressful for engineers. We are building a tool that can silence alerts and help with debugging and root cause analysis. We also want to automate tedious parts of being on-call (running runbooks manually, answering questions on Slack, dealing with Pagerduty). Here is a quick video of how it works: https://youtu.be/m_K9Dq1kZDw I hated being on-call for a couple of reasons: * Alert volume: The number of alerts kept increasing over time. It was hard to maintain existing alerts. This would lead to a lot of noisy and unactionable alerts. I have lost count of the number of times I got woken up by alert that auto-resolved 5 minutes later. * Debugging: Debugging an alert or a customer support ticket would need me to gain context on a service that I might not have worked on before. These companies used many observability tools that would make debugging challenging. There are always a time pressure to resolve issues quickly. There were some more tangential issues that used to take up a lot of on-call time * Support: Answering questions from other teams. A lot of times these questions were repetitive and have been answered before. * Dealing with PagerDuty: These tools are hard to use. e.g. It was hard to schedule an override in PD or do holiday schedules. I am building an on-call tool that is Slack-native since that has become the de-facto tool for on-call engineers. We heard from a lot of engineers that maintaining good alert hygiene is a challenge. To start off, Opslane integrates with Datadog and can classify alerts as actionable or noisy. We analyze your alert history across various signals: 1. Alert frequency 2. How quickly the alerts have resolved in the past 3. Alert priority 4. Alert response history Our classification is conservative and it can be tuned as teams get more confidence in the predictions. We want to make sure that you aren't accidentally missing a critical alert. Additionally, we generate a weekly report based on all your alerts to give you a picture of your overall alert hygiene. What’s next? 1. Building more integrations (Prometheus, Splunk, Sentry, PagerDuty) to continue making on-call quality of life better 2. Help make debugging and root cause analysis easier. 3. Runbook automation We’re still pretty early in development and we want to make on-call quality of life better. Any feedback would be much appreciated!

Saturday, July 27, 2024

New top story on Hacker News: Show HN: Semantic Grep – A Word2Vec-powered search tool

Show HN: Semantic Grep – A Word2Vec-powered search tool
13 by arunsupe | 2 comments on Hacker News.
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.

New top story on Hacker News: An experiment in UI density created with Svelte

An experiment in UI density created with Svelte
107 by 11001100 | 15 comments on Hacker News.


Thursday, July 25, 2024

New top story on Hacker News: Show HN: A personalised AI tutor with < 1s voice responses

Show HN: A personalised AI tutor with < 1s voice responses
30 by za_mike157 | 6 comments on Hacker News.
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://ift.tt/l2AEaMJ Blog: https://ift.tt/dqEGx5n... Github Repo: https://ift.tt/vXPlxgd... 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!

New top story on Hacker News: OpenAI Announces SearchGPT

OpenAI Announces SearchGPT
187 by notyouralias | 121 comments on Hacker News.


Tuesday, July 23, 2024

New top story on Hacker News: Show HN: Zerox – document OCR with GPT-mini

Show HN: Zerox – document OCR with GPT-mini
14 by themanmaran | 5 comments on Hacker News.
This started out as a weekend hack with gpt-4-mini, using the very basic strategy of "just ask the ai to ocr the document". But this turned out to be better performing than our current implementation of Unstructured/Textract. At pretty much the same cost. I've tested almost every variant of document OCR over the past year, especially trying things like table / chart extraction. I've found the rules based extraction has always been lacking. Documents are meant to be a visual representation after all. With weird layouts, tables, charts, etc. Using a vision model just make sense! In general, I'd categorize this solution as slow, expensive, and non deterministic. But 6 months ago it was impossible. And 6 months from now it'll be fast, cheap, and probably more reliable!

Wednesday, July 17, 2024

New top story on Hacker News: Show HN: VisCircuit – A Note-Taking Website for Electronics and Circuits

Show HN: VisCircuit – A Note-Taking Website for Electronics and Circuits
8 by darrenyaoyaoyao | 0 comments on Hacker News.
Hi, everyone. I created a note-taking website for electronics and circuits where you can draw circuit diagrams and write text notes at the same time. I am a Digital IC designer, and I self-study different types of analog and digital circuits a lot. However, I found a problem. Circuits have many different architectures and are hard to memorize due to numerous experiential tips. I want to document what I learn in my note app, but I found there is no method for me to easily draw circuit and block diagrams alongside text notes. This issue has bothered me for a long time, from my master's school to my current working life. I decided to solve it, so I created a note-taking website specifically for electronics and circuits, called VisCircuit. With VisCircuit, you can easily draw circuit diagrams, block diagrams, and write text notes simultaneously. I have already used it for two weeks and have noted down things I find hard to remember, such as SRAM, amplifier circuits, and PCB components of Arduino and Raspberry Pi. I found this tool really useful for memorizing knowledge about electronics and circuits. Currently, I have opened VisCircuit for alpha testing, and I want to let some people use it and give me feedback. Feel free to try it, and I will really appreciate what you think about this project. Leave any suggestions for improvement. Thank you very much.

Sunday, July 7, 2024

New top story on Hacker News: Show HN: A modern Jupyter client for macOS

Show HN: A modern Jupyter client for macOS
27 by jackhodkinson | 4 comments on Hacker News.
I love Jupyter – it's how I learned to code back when I was working as a scientist. But I was always frustrated that there wasn't a simple and elegant app that I could use with my Mac. I made do by wrapping JupyterLab in a chrome app, and then more recently switching to VS Code to make use of Copilot. I've always craved a more focused and lighter-weight experience when working in a notebook. That's why I created Satyrn. It starts up really fast (faster time-to-execution than VS Code or JupyterLab), you can launch notebooks right from the Finder, and the design is super minimalist. It's got an OpenAI integration (use your own API key) for multi-cell generation with your notebook as context (I'll add other LLMs soon). And many more useful features like a virtual environment management UI, Black code formatting, and easy image/table copy buttons. Full disclosure: it's built with Electron. I originally wrote it in Swift but couldn't get the editor experience to where I wanted it. Now it supports autocomplete, multi-cursor editing, and moving the cursor between cells just like you'd expect from JupyterLab or VS Code. Satyrn sits on top of the jupyter-server, so it works with all your existing python kernels, Jupyter configuration, and ipynb files. It only works with local files at the moment, but I'm planning to extend it to support remote servers as well. I'm an indie developer, and I will try to monetize at some point, but it's free while in alpha. If you're interested, please try it out! I'd love your feedback in the comments, or you can contact me at jack-at-satyrn-dot-app.