Show HN: Exploring the design space of binary search trees https://ift.tt/LMpZG9N

Show HN: Exploring the design space of binary search trees https://ift.tt/B2gmlYv August 15, 2023 at 08:04PM

Show HN: Make Any Video Searchable https://ift.tt/kD4Vr6n

Show HN: Make Any Video Searchable Hey HN friends! Yuval here, I'm the maker of Conversational Demos: https://ift.tt/tZSd65r This is a product with a bit of an interesting history: it evolved from my fascination with what we can do with LLMs and video. My original (engineering) thinking was "how cool would it be if we let someone search through video with natural language"? I built the original proof of concept for this over the weekend, taped together with passion, code glue, and sheer will to look away from the really shitty code I originally wrote :) When I showed it to users, I discovered that nobody was really willing to pay for video search - everyone thought it was cool, but there was no hook from a commercial standpoint. And so, as these things happen, I ended up turning it a marketing / sales tool. That's all I'll say about the commercial aspect of it, though I'm happy to expand on it if interesting to the community here. Technical stuff: --- Behind the scenes, when you upload a video (or record one - it's integrated with the not-yet-officially-released Loom SDK), it sends the video to AWS Transcribe, which works with nearly 40 languages, and at the same time converts it to an MP4 format that's tiny and viewable across devices (I do this with AWS Media Transcoder). Next up, I break the transcription down into chunks of around 30 seconds and then send it to OpenAI's embedding database. Through TONS of trial and error, I learned that to get good latency on this thing, you can't really use a (cheap) vector database, and so I'm doing something pretty hacky, and storing the index on a always-on Docker (deployed to fly.io) as opposed to a serverless function, which I was really trying to use until I realized it's not fit-for-purpose. Next up, I upload the embedding index to S3 for later use, as well as store it locally on the machine. That's because the machine has ephemeral memory, and when I re-deploy, the index goes away, and so I just re-download it from S3 if it's not there. I have no idea if this project will commercially succeed. It's live on Product Hunt now, and we'll see how people respond. To be honest, I find building the most fun, and marketing the least fun. So on that front, I did learn several important things about LLMs and building chat-bots along the way: 1. Use the simplest solution if you're optimizing for latency 2. You can't just feed a VTT transcript into an LLM and expect it to figure out the context - you have to do your own chunking 3. There are lots of open source projects out there for using LLMs (langchain, llamaindex, auto-gpt, etc), but just using OpenAI's APIs directly yielded the best results at lowest latency, and least complexity. I'm curious to get the feedback of the HN community - less so commercially, and moreso on the technical front, UX, etc. What suggestions do you have, and what would love to see? Thanks! Yuval August 15, 2023 at 04:23AM

Show HN: LLM Connected with REST APIs https://ift.tt/IxEjLRH

Show HN: LLM Connected with REST APIs Hey, folks here is a peek into Jujutsu. We at Poozle are working with hundreds of APIs and it has been always frustrating to 1. Search the API in the documentation or ask ChatGPT 2. Then copy it to the postman and understand/test the API 3. Generate code to integrate into the codebase We thought how about having all of this at one place. We currently fine-tuned LLM on public REST APIs to reduce hallucination and then combined it with ChatGPT and Postman. I look forward to feedback, feature requests and discussions! https://ift.tt/qNKmxzy August 15, 2023 at 04:02AM

Show HN: Command palette-style Git client https://ift.tt/BfPdqJo

Show HN: Command palette-style Git client https://ift.tt/vrGtuCK August 15, 2023 at 03:31AM

Show HN: Display a chess position in Anki without any addon https://ift.tt/0RJHXMS

Show HN: Display a chess position in Anki without any addon There are quite a few libraries to display a chess position in a web page, but I couldn't find one without dependencies so that I could make an Anki card template with it. There are Anki addons for that purpose but I found them too complicated. I wanted something short that would fit in one page. https://ift.tt/4lmt5jJ August 15, 2023 at 01:30AM

Show HN: Revision History – Show the History of Edits in Google Docs https://ift.tt/OZKSPc4

Show HN: Revision History – Show the History of Edits in Google Docs https://ift.tt/nCfQsih August 15, 2023 at 01:24AM

Show HN: I resurrected one of the top dead Show HNs https://ift.tt/uCKi0lc

Show HN: I resurrected one of the top dead Show HNs OneView was first posted to HN in 2017, but died sometime around late 2019. Using the web archive I cobbled together something that works. Open to hearing people's thoughts on whether this is ethical or not... https://ift.tt/pGdRCPt August 14, 2023 at 11:54PM

Show HN: A website chatbot that also uses APIs https://ift.tt/3ufhDVk

Show HN: A website chatbot that also uses APIs Hi! "Another chatbot for your content" you may ask? Not really! Let me explain. I have realized that there's an untapped potential for ChatGPT plugins: why leave them locked up in the OpenAI store? Let's embed them on websites or share using links! Let's make chatbots ~ do stuff ~, not just talk about what they know. That's how Chatwith was born. It is a chat widget, it scrapes your website, knowledge page, notion etc - but more importantly, it also understands your API. It can interact with that API on your or your visitors' behalf. Some use cases & ideas: - Shopify bot - ask it for order status and invoice - Realtor - ask for budget, provide listings - Survey - ask questions, collect answers, submit to Typeform - Airline service - get your boarding pass - Mixpanel - talk to your analytics In other words: - Talk to APIs & hot data - ChatGPT plugin on every website - Your SaaS can be a chatbot - APIs made accessible (they were reserved for devs until now) - Create an API, use chat as the interface - skip the frontend I am trying to envision a future where website visitors are empowered to just say what they want from the business - instead of navigating endless dashboards and documentation. I'm looking for some feedback about my product! lmk Rafal https://chatwith.tools August 14, 2023 at 05:38PM

Show HN: I made a tool that turns images into videos to boost our Twitter reach https://ift.tt/4mCLnWt

Show HN: I made a tool that turns images into videos to boost our Twitter reach Recently there were some changes to the Twitter algorithm and one of the most important ones was boosting video content even more in the feed, which got me thinking: "Why shouldn't we benefit from the algo changes and get a boost every time we are sharing some visuals?" So I spent the weekend building a tool that: - Turns your images into videos - Has 30 beautiful animated gradients that we can pick from - Is completely FREE I'd love to know your thoughts about it! https://ift.tt/IsgJBwM August 14, 2023 at 04:49PM

Show HN: A free course on how to write a good Midjourney/ChatGPT prompt https://ift.tt/k810uqT

Show HN: A free course on how to write a good Midjourney/ChatGPT prompt https://ift.tt/qXftQ9V August 14, 2023 at 03:07PM

Show HN: Zero-dependency minimal Java web framework https://ift.tt/aFSDkjm

Show HN: Zero-dependency minimal Java web framework I have written a from-scratch minimalist Java web framework, called Minum. I'd like to get your thoughts. The driving goal has been achieving characterics of plainness, frugality, performance, maintainability, testability, and ease of use. It has zero dependencies other than Java's SDK, yet provides a web server, routing, authentication, testing, logging, templating, HTML parsing, and a database, in 3800 lines of code, compiling to a 150k binary. It currently supports a web application in production. There are also example projects demonstrating its use, linked in the documentation. As the framework itself has no dependencies, it is built using GNU Make. The example applications that depend on it, however, use Maven as their build tool. Because Minum is still in beta, it is not yet published to Maven - but there are instructions for a workaround. The framework and example apps require Java 20. https://ift.tt/1gD0vlC August 14, 2023 at 02:45AM

Show HN: PitchPower – AI powered proposal generation for consultants https://ift.tt/XydVo06

Show HN: PitchPower – AI powered proposal generation for consultants Hey HN! I'm an indie hacker who builds SAAS products in my free time. I've got a friend who runs a small consulting firm and said this would be a useful tool for them so i thought why not release it as a standalone online tool for everyone. Basically the premise is, you enter some information about your services (or we just scrape your website and figure it out). Then you enter a description of the client brief, tweak the outline, and out comes a unique tailored proposal that you can tweak with AI. That's pretty much it! Would love to hear your thoughts. Thank you. https://ift.tt/LOWFTD5 August 14, 2023 at 01:02AM

Show HN: NotYetNews – AI-Generated News from the Future https://ift.tt/ZlpmqYv

Show HN: NotYetNews – AI-Generated News from the Future https://ift.tt/Wdw9z0O August 13, 2023 at 09:09PM

Show HN: Broken Bear, the AI teddy bear that loves your broken self https://ift.tt/1MyICLT

Show HN: Broken Bear, the AI teddy bear that loves your broken self I made a GPT-based AI Chatbot based on Carl Roger's philosophy of radical self-acceptance. Broken Bear is designed to be a kind, comforting, and quietly encouraging friend. https://brokenbear.com/ August 13, 2023 at 03:17PM

Show HN: Mixtape of 200 “futuristic” songs circa 1980 https://ift.tt/K8gl53q

Show HN: Mixtape of 200 “futuristic” songs circa 1980 I'm posting this long mix of 200 "futuristic" music recordings, in chronological order (from the mid 1970's to the mid 1980's). Some are obvious (eg: "She Blinded Me With Science") while others are long forgotten (eg: G.G. Tonet's "Dedicated To Norbert Wiener" or Jyl's "Silicon Valley") --- Circa 1980 Mixtape https://ift.tt/ZGxFEBQ --- While the theme of the linked mix (ie: tech) probably interests some here, it is likely too fluffy for many others so apologies for that. This thing took me almost two years to finish, so in a moment of weakness I am going ahead and posting it here. There actually is a tech angle to why it took so long. After I gathered the 200 audio tracks I wanted, I wound up stymied for several months because trying to match the timbre and loudness of 200 songs overwhelmed me. During that time I began writing a program in my spare time to band-split all the files and match their perceived loudness. I got around half-way through that when OpenAI released GPT4. On a whim, I asked it to write a BASH script to perform the auto-EQ: a couple evenings of minor tweaks and it was done https://ift.tt/Ig2Kzqt so I went ahead and completed the mix. The final mixing together of songs - I am not a DJ and have no experience beat-matching - was also done programmatically (wrote a program that takes start- and end- times of the regions to crossfade, and then ramps samplerate up or down for the two tracks. August 13, 2023 at 01:32PM

Show HN: Run LLaMa2 on the Browser with Ggml.js https://ift.tt/XbsjhGf

Show HN: Run LLaMa2 on the Browser with Ggml.js You can now build serverless AI inference web application with ggml.js's LM backends. https://ift.tt/q5cWzbx August 13, 2023 at 01:22AM

Show HN: Pho – Phonetic Alphabet Renderer in Rust https://ift.tt/0UijuCa

Show HN: Pho – Phonetic Alphabet Renderer in Rust https://ift.tt/1nBqrvF August 12, 2023 at 02:41PM

Show HN: Open-source proxy server for Llama2, GPT-4, Claude2 with Logging,Cache https://ift.tt/bVPzlwA

Show HN: Open-source proxy server for Llama2, GPT-4, Claude2 with Logging,Cache Hello hacker news, I’m the maintainer of liteLLM() - package to simplify input/output to OpenAI, Azure, Cohere, Anthropic, Hugging face API Endpoints: https://ift.tt/HX8j0Il We’re open sourcing our implementation of liteLLM proxy: https://ift.tt/FGrvDk8... TLDR: It has one API endpoint /chat/completions and standardizes input/output for 50+ LLM models + handles logging, error tracking, caching, streaming What can liteLLM proxy do? - It’s a central place to manage all LLM provider integrations - Consistent Input/Output Format - Call all models using the OpenAI format: completion(model, messages) - Text responses will always be available at ['choices'][0]['message']['content'] - Error Handling Using Model Fallbacks (if GPT-4 fails, try llama2) - Logging - Log Requests, Responses and Errors to Supabase, Posthog, Mixpanel, Sentry, Helicone - Token Usage & Spend - Track Input + Completion tokens used + Spend/model - Caching - Implementation of Semantic Caching - Streaming & Async Support - Return generators to stream text responses You can deploy liteLLM to your own infrastructure using Railway, GCP, AWS, Azure Happy completion() ! https://ift.tt/iyDoJuk August 12, 2023 at 05:38AM

Show HN: Weaviate – Multi Tenancy Notebook https://ift.tt/gNxV7FT

Show HN: Weaviate – Multi Tenancy Notebook https://ift.tt/DA8Qbdv August 12, 2023 at 04:20AM

Show HN: Clop – optimise screenshots and recordings automatically https://ift.tt/MH1vwWY

Show HN: Clop – optimise screenshots and recordings automatically https://ift.tt/tLsfu59 August 12, 2023 at 01:41AM