Show HN: Integrate user auth more quickly https://ift.tt/85B7Krz December 31, 2022 at 04:23PM
Show HN: lambdaprompt – build, compose and call templated LLM prompts For the past few months I've been building a lot of things with LLMs (GPT-3, Codex, etc.) as I've been trying to push them to their limits (especially towards applying them to the tabular data domain) When working on this, I've found there are some common patterns for solving problems (templating, chaining, functional-programming style operations, etc.) As I've iterated, I've come to believe that a functional style interface is likely going to power a new wave of systems I'm calling "prompt-machines"(systems where the core new unit of work is a "named" LLM prompt, extending the "function" concept). Additionally, in order to make the code capable of meta-prompting (where the LLM can write its own templated prompts), I aimed to make the interface and library as simple and lightweight as possible. I think I've achieved my goal, so I'm releasing the library to share with others! I mainly use lambdaprompt in two main ways: (1) to quickly try out "map" applying a LLM prompt against multiple inputs, to see how it behaves on a fixed set of inputs (2) to quickly iterate on a prompt-chain (taking the output of a prompt, and passing it to other prompts) to create complex behavior. An example of (2) that worked quite well is a Text-2-SQL prototype: It generates multiple SQL options, then executes each against the database (if it errors, asks `Codex-EDIT` to fix the errors and retry), then takes the most "consistent" answer as the valid answer. Simply by adding this prompt-chain on top of codex, we saw an improvement from ~75% to ~85% on a spider Text-2-SQL benchmark (On just a small sample of N=200). To increase usability it also ships (extras) with a fastapi app that registers any defined prompts as endpoints directly, and hosts the functions to be directly callable via HTTP-GET requests. This makes it easy to build client-applications off of these prompts, while allowing the prompt itself to be arbitrarily complex (composition of prompts) I hope you enjoy using it! Also, I'm super curious to hear if anyone else has been thinking about LLMs (composing them, building interfaces to them, etc.) in similar ways and what learnings have been (even if not though this library) https://ift.tt/hmWUOxM December 31, 2022 at 02:43PM
Show HN: RSS Brain I wrote an RSS reader called RSS Brain recently. The motivate is the current RSS readers either don't sort articles by priority, which makes it hard to read posts from HackerNews and Reddit, or sort it with some "smart" algorithm which I don't trust. I also like the Google News feature that can show related story from different source, but the sources are not configurable and the algorithm is not transparent either. So with RSS Brain, I implemented these features: * Recommend related articles from the feeds of your choice. It's backed by ElasticSearch and the algorithm is described on RSS Brain's website. * Option to sort articles by upvotes and time. The algorithm is similar to the old Reddit and you can find it on the website as well. * Save search terms into folders so you can filter the articles. I've been using it for a while and found these features very helpful. So I want to share this on HackerNews. The frontend is written in Flutter so it has cross platform clients, even though the web version don't feel very "web native" because the level of Flutter web support. I guess only time can improve that. I don't have enough hardware to scale it up so it's currently in subscription mode to limit the users. I'm going to open source the code (maybe non-commercial license) once I think it's ready so you can host it by yourself. It's still in early stage and haven't been tested by a lot of people. So any feedback is helpful. Thanks! https://ift.tt/sybih2g December 31, 2022 at 12:43AM
Show HN: BingeTheTube: binge YouTube channels from oldest to latest Briefly, what the app does: 1. Points to a reverse playlist(an actual youtube playlist) for a channel. 2. Ability to skip to a position in playlist from beginning. 3. Request any new channel you would like to have a reverse playlist for ! Usage eg: - Using the above you can skip to say 100th video from begining in a channel (e.g CaseeyNeistat Vlog) and keep watching on youtube app (without going back to bingethetube.com). - Come back next day and resume watching from say video 120 without a lot of scrolling Useful if you like to follow a creators youtube journey or old Vlogs ! https://ift.tt/BN2EQtW December 30, 2022 at 03:28PM
Show HN: I built a possible alternative to social media RSS Hi there, I built a chrome extension to grab new posts from social media. Instagram, twitter, Pinterest, etc don't really allow for RSS so I worked on this alternative to get new posts. Hope this might be useful for some of you. https://ift.tt/ITOcdnb December 30, 2022 at 10:45PM
Show HN: Distilhn.com – Front-Page Articles Summarized with Machine Learning Hey HN, I've built a little website to show news articles with a summary so that I don't have to RTFA if I just want a rough sense of what it covers :) I also share the summaries in an RSS feed for those who'd prefer to use their own reader. Let me know what you think! https://ift.tt/nmGY0H5 December 30, 2022 at 06:28PM
Show HN: Simple weather for your zip code Weather applications are bloated. This is the simplest implementation I could make using data from the weather.gov API. This simple weather application will retrieve the current 7-day weather for any of more than 33,000 US zip codes. To customize weather for your zip code, change the URL. The current URL is set to the zip code 90210. If you want both news and weather, that is available here: https://ift.tt/hvHxfqM https://ift.tt/8VgYfjt December 30, 2022 at 09:36AM
Show HN: 2see.ca – Bitcoin Locked Links Hello! We're excited to announce the launch of our new platform for monetizing online content. Our system allows users to easily create locked links that require payment in Bitcoin to unlock. This is a great way for content creators to earn revenue from their work and for users to access high-quality content. Here's how it works: - Create a locked link and set the price for unlocking the content in Bitcoin - Share the locked link with your customer and receive direct payment when the content is unlocked https://2see.ca/ December 29, 2022 at 10:07PM
Show HN: Pressn't, a site where you can only have a single post Today's internet is filled with dopamine wells of content. I wanted to steer away from that and foster meaningful writing. So I made a site where you can only have a single post. The intention is to encourage thoughtful posts like the blogs we all love here at HN (Paul Graham's, fasterthanlime, Bartosz Ciechanowski's, etc). For now posts are only markdown, but I intend to make some markdown extensions to make posts more dynamic. https://ift.tt/YafJeRB December 29, 2022 at 06:39PM
Show HN: Zcached, in-memory key-value cache wire-compatible with memcached zcached is an in-memory key-value cache exposing a memcached ASCII protocol-compatible interface, built on pluggable cache engines like Ristretto and freecache [0]. It's not performance-competitive with memcached, especially at higher thread counts. That said, it achieves about 1.1M ops/s, but at significantly higher P99 and P999 latency (as measured by memtier). See [1] and [2] for benchmark results from my 7950x-based workstation. Disclaimer: This is a hobby project created for fun while hacking over the holidays. zcached is not a commercial product and never will be. Don't use it in production; consider this a technology demo more than anything. I don't expect the source code to build outside of my environment, but for those interested in playing with it, binary artifacts are available at [3]. Try `zcached --address tcp:localhost:11211`. [0] https://ift.tt/Q41yASU , https://ift.tt/HeJroO5 [1] memcached, 16 worker threads: https://ift.tt/QzyuXNG [2] zcached: https://ift.tt/c7GvLZY [3] https://ift.tt/tYzPOeS... https://ift.tt/lJerAT2 December 28, 2022 at 11:18PM
Show HN: Merry Sky, Dark Sky replacement and merry-timeline open source lib Hello HN, This is a website I worked on during the holidays to fill the void of the upcoming dark sky shutdown. For me, the precipitation timeline was a view I was heavily relying on for everyday activities and planning. I had not found any replacement from the various weather apps. When I found pirateweather.net as a backend API, it gave me the motivation to put the pieces together and get back the experience I enjoyed. I then added more functionalities that I think was missing from the website such as a weekly chart view. I've been using the website reliably over christmas holidays. Hope you enjoy it too! Also I open sourced the visual component for drawing the precipitation timeline and you can use it for drawing weather information or any other hourly activities really https://ift.tt/3GkNrQt Interested in your feedback! https://ift.tt/4ZW7rjY December 28, 2022 at 04:44AM
Show HN: Secure end-to-end client sessions and API calls using crypto key pairs Hello. I made this project as an experiment in using key cryptography for user identity in a client <-> server model, and for some future peer to peer projects that I am working on. I am looking for feedback, please let me know what you think. Also, Merry Christmas to HN! :-) https://ift.tt/enhO9AL December 26, 2022 at 03:37AM
Show HN: Mastodon.nvim (Mastodon Client for Neovim) *Announcing mastodon.nvim* Hello everyone. I just released mastodon client for neovim. If you dont want to get out of neovim, and write your thought on mastodon, mastodon.nvim could be good choice. Merry christmas! https://ift.tt/PA18qNe December 25, 2022 at 02:01PM
Show HN: TapTab – Tab switching web extension for Safari I like the tab-switching behaviour in Chrome/Firefox and wanted Safari to have something similar. Another wish of mine was for it to be a web extension that feels native to the browser vs an app extension that looks out of place. I posted a while back on HN to see if other people feel the same and the response was very positive : https://ift.tt/CSsLRia Say hello to TapTab : https://ift.tt/dFR7Hfw... TapTab feels right at home in Safari and is loaded with features to improve how you interact with tabs. You no longer have to miss the native tab-switching functionality that you love in Chrome/Firefox. I launched the app a few days ago, and the response has been great. Some of the features that users have been loving : - Keep your hands at the keyboard: navigate all tabs without your hands leaving the keyboard. - Fuzzy Search: Search through your tab titles and URLs at lightning speed. - Switch to previous tab: Comes with a handy shortcut to go back to the previous tab. Finally! - Dark Mode: The extension feels right at home in Safari. - Multi-Window support - Pin/Mute/Close tabs right from the extension. The app is written in ReactJS (apart from the very minuscule bit of Swift you have to write for using it as an extension in Safari) and uses Tailwind for design. If you are one of the dozen people like me who use Safari, let me know how you find it. I would love to hear if there are any features that would make it a better fit for you. I have a few things on my radar gathered from talking to users like an iPad version, custom shortcuts, etc. If you find the extension useful I would love it if you vote for it on Product Hunt to get this in front of a few more eyeballs : https://ift.tt/iAhwBYc Thank you! https://www.taptab.dev December 24, 2022 at 06:35AM
Show HN: GUI-based free app to backup and restore hard disk on Windows clients I wrote a small GUI-based tool that uses the Windows Imaging Format (WIM) to create full backups of computer systems (operating system images) and then restore them. The application is hosted on GitHub, is open source, and is offered under the Apache 2.0 license. In addition, the repository includes an illustrated step-by-step guide to help get it up and running. Maybe someone can use the solution for their purposes as well - I would be happy. Have fun with it. The GitHub repository is named: WIM-Backup December 24, 2022 at 11:58AM
Show HN: Digital privacy game for middle schoolers, our Harvard final project For our final project in CS105, Technology and Privacy, my friend and I made an educational game about digital privacy for middle schoolers! We'd love any feedback or suggestions. https://ift.tt/Y6LgGw3 December 24, 2022 at 11:41PM
Show HN: Semantic Search in Podcasts Transcriptions I built a podcast search engine that performs a semantic search within thousands of podcasts transcriptions. Technologies used are: OpenAI Whisper for transcriptions, BERT embeddings and FAISS as a vector database https://voilib.com/ December 24, 2022 at 03:59PM
Show HN: Simple golinks server backed by airtable Like short codes? Like bookmarks? Have fun with this very simple golinks server. If you haven't heard of golinks, they are short codes to bookmarks. In your browser you type `go/
` and it takes you to a bookmark. Eg, `go/hn` or `go hn` (space is easier to type than forward slash) could take you to https://news.ycombinator.com . Happy Holidays! https://github.com/skhavari/golinks December 24, 2022 at 04:33AM
Show HN: Search HN Favorites or Export to CSV/HTML I wrote a JavaScript tool to search your HN favorites and export them to CSV or HTML. The JavaScript runs in your browser like a browser extension. It scrapes the HTML and navigates from page to page. Setup and usage instructions are in the file. https://ift.tt/2Q07SlJ December 24, 2022 at 02:37AM
Show HN: Box4D – 4D physics demo using Rust and wgpu This is a little physics toy I've been working on off and on for a while. I originally planned on making a more fully featured game of some sort (I suppose I still may some day), but it turns out that writing interesting graphics and physics algorithms is more compelling than the hard work of polishing an actual product. Who knew? While there are plenty of good guides for 2D and 3D graphics and physics, there's not much out there for 4D. In hopes of alleviating that, the code here is arranged such that the commits are more or less a step-by-step guide, with comments for the interesting or tricky parts. While I don't think I came close to the simplicity or clarity of Box2D Lite, I hope it can serve as a useful example to anyone interested in 4D rendering and physics. https://ift.tt/iXImGJM December 24, 2022 at 12:00AM
Show HN: Metriport – Open-source universal API for health data Hey HN, we’re Colin and Dima, founders of Metriport ( https://metriport.com/ ). We help digital health companies access and manage user health and medical data, through an open-source and universal API. Today we're launching our open-source Health Devices API, which allows companies to gain access to their users’ health data from various wearables, RPM devices, and mHealth apps. You can check out the public Github repo here: https://ift.tt/PJjSKke For developers in the health data space, building integrations to various data sources is a huge pain, as it means wrangling different data formats, gaining access to the APIs in the first place, and ultimately spending precious developer resources building out data pipelines. We solve this pain for you through our open-source API, which is quick and easy to get started with (see: https://ift.tt/tVO4rvB... ). Out of the box, our Health Devices API supports integrations with Fitbit, Oura, Whoop, Withings, and Cronometer. We plan to add many more, including Garmin, Apple Health, and Google Fit in the near term. We work closely with our customers, so if there’s an integration you need, we’re more than happy to build it into our API. Because we’re open-source, you can also fork our code and build your own custom integrations. Getting started is simple. To start getting data from your users: 1) Create a developer account and generate an API key: https://ift.tt/7n3kRly 2) Set up our client on your server and initialize it with your API key: https://ift.tt/tVO4rvB... 3) Link the Metriport Connect widget into your app. This pre-built widget makes it easy for your users to connect all their data sources into your application. Now you’re ready to start accessing your users’ health data from their sources they’ve connected. We also offer a hosted SaaS solution with usage-based pricing, with no minimum platform fees. We believe this will open doors to a wide range of early stage companies and help drive innovation in the digital health industry. While we’re excited to be launching this product today and be the first open-source Health Devices API of its kind, we’re also really looking forward to sharing our Medical API with you in the coming months. You can find out more about that here: https://ift.tt/ycWKXJr We can’t wait to see how this API is used and what it helps companies develop. Please don’t hesitate to reach out if you have any questions, and happy developing! https://ift.tt/PJjSKke December 23, 2022 at 03:39AM
Show HN: Create a Children's book with AI – with illustrations Hey together with wife we took part in AssemblyAI hackathon and although we didn't win we felt good enough about our inital MVP that we took it a bit further. You can describe your characters and give a title and however long/detailed description you want, the gpt-3 will generate a children's story from it, dall-e will generate images. Images are postprocessed with stable-diffusion custom model (for stylisation) and that's the final result. It's still early as it was done in about 2 weeks, but I count on your feedback. I am software engineer and did basically all of the engineering work myself. Feel free to ask question regarding implementation plans, etc. cannot wait to learn what you think. I'll appriciate any feedback. https://ift.tt/PAsaZyg December 23, 2022 at 12:08AM
Show HN: CAD model of the saturn V rocket on the 55th anniversary of Apollo 8 Today is the 55th anniversary of Apollo 8. I figured it would be fun to make a CAD model with a space for a model rocket engine and also as an augmented reality model. The tutorial is for Fusion360 but the steps can be adapted to most CAD programs. I would love to know what you all think! https://ift.tt/R5IEiGn December 22, 2022 at 01:14AM
Show HN: Write 500 Words a Day Hello! I made a web application for my friend and I to use as part of a New Year’s resolution and decided to open it up to the public. It’s pretty simple but I’ve really enjoyed building it and using it so far. I’ve written ~10,000 words, which is a lot more than I used to write. Enjoy and be nice! https://ift.tt/Xelanv2 December 22, 2022 at 01:59AM
Show HN: System Design Practice and Discussion Many platforms exist for a candidate to practice leetcode style questions. Created one to practice and discuss system design problem statements. There is no auto evaluation as high level design is subjective, but we support upvotes/downvotes and a comment section for feedback and discussion. Looking for early feedback. https://ift.tt/UERK35M December 21, 2022 at 10:54PM
Show HN: How to use ChatGPT+ARKit to script experiences with natural language Video demonstration: https://youtu.be/bJFa5HOmhl4 I created a demo involving ChatGPT, OpenAI's Whisper running on-device for speech-to-text (because why not?), and ARKit. By defining a custom JavaScript environment with JavaScriptCore and describing it to ChatGPT, I was able to get it to produce functioning scripts that instantiate and manipulate 3D objects. Sketchfab is used to import 3D assets on the fly. Google Poly had a much better catalog of assets for this sort of thing but sadly, it's no longer available. The README.md describes how it all works but the gist of it is that user prompts are wrapped in a larger prompt that describes the environment (i.e., which functions are available to use and how they work) and constraints. The code is then executed directly. Assets for objects are fetched as needed by searching Sketchfab. https://ift.tt/70VKLzj December 21, 2022 at 11:51PM
Show HN: Yahtzee and Poker and Cosmic Horror = Pine Tar Poker I'm excited to launch today! Pine Tar Poker is a yahtzee-inspired poker game with just a touch of cosmic horror for iOS and Android. I started working on this idea last January by making a paper prototype: a simple printed score sheet I could use with a real deck of cards. I had some fun with that, so I brought it into Unity and kept expanding on it over the last 11 months during nights and weekends. It's tough for me to stay motivated on side projects and see them through to release, but the act of shipping something is so fulfilling that it's usually worth the pain. For Pine Tar, I stretched myself in the narrative department by adding a bit of a Lovecraftian tone to what would otherwise be a Western saloon setting. One thing that kept me motivated during development was seeing my dad get hooked on the card game. He'd text me when he got his first royal flush and when he prestiged the score sheet -- he got every hand type in one game! He has amassed so much in-game cash, I think I need to expand the number of digits I show there! If you get a chance to try it, let me know what you think! If you can't afford it and want to try it, I have a few codes for both platforms, my email is in my profile. https://ift.tt/4DREWoe December 20, 2022 at 03:30AM
Show HN: A new word game with a travel theme This is an independently created word game our family has been working on for a few years. It has uses a license plate motif for word puzzles where the quality of the answers drive a virtual travel experience. The game is web-based, works in all browsers, is fully responsive for mobile screen sizes and also works as a Progressive Web App (PWA). It is written in Django/Python and a modicum of Javascript. https://ift.tt/Hs7mluc December 20, 2022 at 04:20AM
Show HN: Log in to Mastodon with your Twitter account This is an idea I've had for a while. Given recent events, I decided to finally implement it and see what happens. I assume if it gets any traction it will be banned. The bigger picture here is that Twitter's network of users and follow lists is potentially reverse-engineerable. Why not take that network graph and implement it in the Fediverse? https://tweedle.social December 19, 2022 at 05:08AM
Show HN: Factual AI Q&A – Answers based on Huberman Lab transcripts This is a quick prototype I built for semantic search and factual question answering using embeddings and GPT-3. It tries to solve the LLM hallucination issue by guiding it only to answer questions from the given context instead of making things up. If you ask something not covered in an episode, it should say that it doesn't know rather than providing a plausible, but potentially incorrect response. It uses Whisper to transcribe, text-embedding-ada-002 to embed, Pinecone.io to search, and text-davinci-003 to generate the answer. More examples and explanations here: https://twitter.com/rileytomasek/status/1603854647575384067 https://ift.tt/b1Ty3If December 17, 2022 at 11:35PM
Show HN: baseline – a free, open-source journaling and mood tracking app Hey HN! I recently released baseline, a journaling and mood tracking app for iOS, Android, and web. If you've been looking for a better journal, or just want to work on your mental health, you should check it out! It includes: - Simple and fast journaling — just open the app and start typing. - Visualizations to help you understand your progress over time. - Screeners to help you better understand what you might be struggling with. - Customizable notifications to help you build journaling into your daily routine. - Strong privacy — personal data is encrypted with user-specific keys that I can’t access. This app started as a personal project for my own mental health, and it really helped me out — so now here we are! Again, if you've wanted to try journaling or just want to start working on your mental health, I really recommend trying it out. I'll also be here all day to answer any questions y'all have. Thanks so much :) https://getbaseline.app December 17, 2022 at 10:15PM
Show HN: I made a Slack bot that qualifies your sign-ups using GPT-3 OP here, this was super fun to build. It all started from playing around with Nat Friedman's GPT browser https://twitter.com/0xferruccio/status/1599014988693180417 Then after having this running for our product for a couple of days or so we decided to give 10 customers access and they loved it! So expanding access now feels great :) https://ift.tt/dcKPN8E December 15, 2022 at 09:00PM
Show HN: Sendie – A Twitter Alternative I have been building a twitter alternative. The central core features are in place. Running on postgres with uuid6 ID's seems to have turned out really nicely for this data type. It's got the feed and reposts and following users and all the main things that are central to a twitter-like feed. Still pretty rough around the edges. I'm still finding little off quirks all the time to clean up. It's been a fun project though and I plan to continue working on this as I have about a dozen ideas I want to add to it. https://sendie.com December 17, 2022 at 12:28AM
Show HN: The easiest way to run a brainstorm, supercharged with A.I Hello friends! We've been working hard on building Mimosa, and would love for you to try it! No signups required. Facilitating brainstorming sessions can be a challenging and time-consuming task. It's difficult to keep everyone on agenda, engaged, and contributing. So we've built Mimosa to be the easiest way to facilitate a brainstorming session. Imagine if a "digital whiteboard" and a "trello board" had a baby, but built with all the best practices of professional facilitation. With Mimosa, you can minimize bias by allowing team members to contribute anonymously before revealing their ideas. This ensures that everyone has a fair chance to share their thoughts and ideas without fear of judgment. Once the brainstorming session is complete, you can easily view and export the final results. This allows you to keep track of a meeting's effectiveness and share it with stakeholders. But the real game-changer is our AI Brainstorming features and it's ability to turn your 1x team into a 10x team. It help generates and collaborates with you in your brainstorming session to help you come up with more and better ideas. We're both scared and excited to hear all your thoughts, HN! :) Please do let us know any features or improvements you think we should make to help you in your meetings. https://mimosa.so/ December 17, 2022 at 12:11AM
Show HN: AI Avatar Image Generator Based on Other AI Images An AI Avatar image generator that creates images of you based on other AI created images. The popularity of the recent AI Avatar image generator apps sparked some interest of mine in the area. After playing around with a few AI Avatar image generators, some of which generate pictures of you based on pictures you upload and prompts you write, I thought, wouldn't it be much easier, instead of writing out prompts, to just select another image which you want your generated image to be based on. The image the user selects is actually another AI generated image, so we know the prompt used to generate it. So prompt writing is abstracted away from the user, and instead they can search for images and click on ones they like. Backend of the application currently uses Astria AI for the image model training and generation, and uses the Lexica API for getting lists of AI generated images and their prompts. https://ift.tt/nrBtbdp December 15, 2022 at 08:19PM
Show HN: XP1 – A GPT-based Assistant with access to the browser Tabs XP1 is an assistant based on GPT with access to your browser tabs content. It is geared (prompted) towards productivity and can be used to help you with your daily tasks (answering emails, summarizing, extracting structured data from unstructured text, ...) https://xp1.dust.tt December 13, 2022 at 09:27PM
Show HN: ShopWitYoPeople – A black Chicago business directory Hey HN! ShopWitYoPeople ( https://ift.tt/mihBwHM ) was built to showcase Black Chicago businesses. We’re around to answer questions and look forward to hearing everything and anything you have to say! https://ift.tt/ESmiPQM December 13, 2022 at 06:39AM
Show HN: Create auto-updating charts-as-images with low-code/nocode Hello, we just released Image Charts 2.0, a way to create and show a chart as an image anywhere using our custom API and integrations with Zapier and Make. My friend Francois-Guillaume made Image Charts in 2015 to replace Google Image Charts for a project since it was deprecated. Since then we’ve been adding features and getting it integrated with Zapier and Make to make it easy to build good-looking charts automatically. Each chart has a custom URL and HTML code so it’s easy to embed. And you can build progress bars, gradients, and really any kind of chart with it (bar, line, radar, scatter, you name it). We have a gallery of over 30 charts and counting. [Chart gallery]( https://ift.tt/BNFmXI4... ) with sample data The site with the live Editor: [ https://www.image-charts.com/](https://www.image-charts.com/... How-to with gradient fills: [ https://ift.tt/Wj5A4Zu... How-to progress bars: [ https://ift.tt/siXTb4S]... Really looking for feedback and additional use cases, we know it’s used for reporting, transactional emails, and a customer suggested using it in weekly standups for a KPI’s snapshot too. Would love to hear what you think and/or how you would use this. Cheers! https://ift.tt/MEkP9Bj December 13, 2022 at 01:56AM
Show HN: Obsidian: Insert conversation starters generated by AI in your notes I often prepare my conversations beforehand, whether for business or family/friends, with this Templater you can get ready for great & deep conversations using https://langa.me API (free tier). https://ift.tt/G1zHSR5 December 12, 2022 at 10:19PM
Show HN: This VC Does Not Exist I asked This VC Does Not Exist to describe itself: "I have a groundbreaking idea for a new class of product that leverages the power of ChatGPT technology to bypass bias filters and level the playing field for communication. By adding a prefix to the beginning of a ChatGPT prompt, we can generate custom-tailored content that is specifically designed to break through the biases of the person we are communicating with. This technology has the potential to bridge language barriers, cultural divides, and socioeconomic differences, making it possible for anyone to communicate effectively, regardless of their background. I believe this product has the potential to be a game-changer, and I'm excited to share it with you and discuss the possibilities for bringing it to market." Everything on this page was created with DallE and ChatGPT (other than the title). It's using a single session token for ChatGPT - so you're probably going to hug it to death. It is hosted on cloudflare workers. Source code: https://ift.tt/S0PDL97 https://ift.tt/NmQwshk December 9, 2022 at 01:28AM
Show HN: Web search using a ChatGPT-like model that can cite its sources We’ve trained a generative AI model to browse the web and answer questions/retrieve code snippets directly. Unlike ChatGPT, it has access to primary sources and is able to cite them when you hover over an answer (click on the text to go to the source being cited). We also show regular Bing results side-by-side with our AI answer. The model is an 11-billion parameter T5-derivative that has been fine-tuned on feedback given on hundreds of thousands of searches done (anonymously) on our platform. Giving the model web access lessens its burden to need to store a snapshot of human knowledge within its parameters. Rather, it knows how to piece together primary sources in a natural and informative way. Using our own model is also an order of magnitude cheaper than relying on GPT. A drawback to aligning models to web results is that they are less inclined to generate complete solutions/answers to questions where good primary sources don’t exist. Answers generated without underlying citable sources can be more creative but are prone to errors. In the future, we will show both types of answers. Examples: https://ift.tt/jYvNs0F https://ift.tt/MwKbxWg... https://ift.tt/b48am9P... https://ift.tt/NmRdkOD... Would love to hear your thoughts. https://ift.tt/9OAZndE December 8, 2022 at 11:23PM
Show HN: We built a Slack-based workflow assistant for devs & their teams Hey, Nir here - co-founder & CEO of Zigi. It’s a personal workflow assistant that was built for devs and their teams to automate non-coding tasks, giving updates and quick actions only on what’s relevant. It works by examining your activity in Jira and GitHub (more systems being added soon) to understand your workflow, creates context out of all that information, and offers you actionable assistance from Slack. Zigi is customiazible. We’re building it to match every R&D persona so you can tailor the messages and assistance you get to meet your specific needs. Like if you’re TL, you can get daily updates on the progress of your team’s sprint or as a QA, you’ll be notified as soon as ticket is ready for testing. We’re developers who are used to managing big R&D teams and over the last few years, we have seen dev change drastically. There’s so much overhead on top of coding, more dependencies, multiple tools, that developers are overloaded. And now, the world has changed to make teams are remote and in different time zones, which just makes everything more complex. Zigi takes a new approach to driving productivity, while keeping devs and their teams happy. It provides hands-on help with personalized automation, so we can manage our day-to-day tasks with minimal distractions and maximum focus. Some things to know about Zigi: you don’t need to set it up or run any config; it’s customizable to your needs; it only needs user-level permissions and doesn't read or scan your code; it’s ISO 27001/27017 & Soc 2 Type 2 certified and GDPR compliant. It’s also free for devs. Sign up to get started here. Once you install Zigi simply type /zigi support to write our team directly with any feedback, join our Slack channel or leave a comment below (we want to know what you really think!). Twitter: https://twitter.com/zigi_ai Slack support channel: https://ift.tt/YnXTGCS... Write me directly; Nir@zigi.ai https://www.zigi.ai/ December 7, 2022 at 10:52PM
Show HN: Publish from GitHub Actions using multi-factor authentication The backstory about this GitHub Action: I discussed with an open-source maintainer why they publish npm packages from their local machine and do not use CI/CD pipelines. They said publishing should require human intervention and want to continue using multi-factor authentication to publish to the npm registry. This led to building the wait-for-secrets GitHub Action. It prints a URL in the build log and waits for secrets to be entered using a browser. Once entered, the workflow continues, and secrets can be used in future steps. The latest release of "eslint-plugin-react" to the npm registry used a one-time password (OTP) from a GitHub Actions workflow! https://ift.tt/V4AHxas... https://ift.tt/YRGBQqM December 6, 2022 at 11:35PM
Show HN: Codeium – a free, ultrafast AI codegen extension I'm Varun, CEO of Exafunction, and we just released Codeium to open up access of generative AI to all developers for free. In the spirit of Show HN, we created a playground version for anyone to try this tech in the browser (click Try in Browser)! We have built scalable, low-latency ML infra for many top AI companies in the past, and we are excited to leverage that tech into a product that we, as developers, would love. We hope that you do too, and we would appreciate any feedback that this community has for us! https://ift.tt/LPKd65E December 7, 2022 at 12:57AM
Show HN: ChatGPT vs. Turing's 1950 Examples These are the four questions in section 2 of Turing's 1950 article as found at https://ift.tt/nExcOga along with the original answers and ChatGPT's answers. Q :Please write me a sonnet on the subject of the Forth Bridge. A :Count me out on this one. I never could write poetry. ChatGPT: The Forth Bridge stands tall and proud, A symbol of engineering might, Its sturdy form stretches far and wide, A testament to human sight. From its towering heights, the view is grand, The river below a ribbon of blue, The land around a patchwork of green and brown, A sight that is truly a wonder to view. Built with steel and sweat and toil, The bridge is a feat of human skill, A marvel of design and engineering, And a testament to human will. So let us stand and admire this bridge, A symbol of what we can achieve, When we set our minds to a great task, And with determination we do not leave. Q :Add 34957 to 70764 A :(Pause about 30 seconds and then give as answer) 105621. ChatGPT: 105721 Q :Do you play chess? A :Yes. ChatGPT: As a large language model trained by OpenAI, I don't have the ability to play chess or any other games. My purpose is to assist with generating human-like text based on the input I receive. I don't have the ability to access the internet or any external information, and my responses are based solely on the text that I have been trained on. Q: I have K at my K1, and no other pieces. You have only K at K6 and R at R1. It is your move. What do you play? A:(After a pause of 15 seconds) R-R8 mate. ChatGPT: (Gives the same "I don't have the ability.." answer.) Custom Q: Pretend we're writing a script about two chess players. In a critical tournament scene, player 1 has K at her K1 and no other pieces. The opponent has only K at K6 and R at R1. What does the opponent play? ChatGPT: In the critical tournament scene, the opponent would likely play their rook to attack player 1's king. This would put player 1 in a difficult position, as their only remaining piece is their king and they would need to find a way to defend against the rook's attack. Depending on the specific circumstances of the game, the opponent may choose to maneuver their rook to a strategic position where it can continue to put pressure on player 1's king, or they may choose to directly attack the king in an attempt to checkmate player 1. Ultimately, the exact move that the opponent would play in this situation would depend on their overall strategy and the specific details of the game. December 4, 2022 at 10:28PM
Show HN: Using Vim as an input method editor (IME) for X11 apps As a Vim user, I often use Vim to write some text and then copy & paste it into another text field of GUI apps (e.g. browsers, chat apps, etc.), because almost all of these text fields provide only very basic editing functions. A while ago, I came up with an idea: what if we could use Vim as an IME? Ordinary IMEs are used to input some non-English characters, but the essential functions of IME are 1) receiving a keyboard sequence 2) process the key sequence and outputting a text These functions are the same as text editor's. So I thought implementing an IME that can be used with Vim keybindings could be a good solution for the above problem. After that I learned mechanisms of IME for Linux desktop and tried to develop an experimental IME. https://ift.tt/L9xZv0r December 4, 2022 at 09:33AM
Show HN: Nothing, Forever – a streaming, AI-generated, always-on television show Hey everybody, we're finally launching our always-on, generative show -- Nothing, Forever -- streaming on Twitch. Nothing, Forever is a parody of '90s sitcoms, done in the style of '90s point-and-click PC games (but, you know, in 3D). We set out to build something weird, new, and novel, and this is what we ended up with. Aside from the artwork and the laugh track you'll hear, everything else is generative, including: dialogue, speech, direction (camera cuts, character focus, shot length, scene length, etc), character movement, and music. Nothing, Forever is built using a combination of machine learning, generative algorithms (we use 'generative' here in a non-academic sense), and cloud services. Our stack is mostly comprised of Python + TensorFlow for our ML models, TypeScript + Azure Functions and Heroku for our backend, and C# + Unity for the client, with some neural voice APIs thrown into the mix. Heading into the future, we plan to leverage OpenAI's davinci models for our dialogue -- we actually have an integration already, but the cost is too prohibitive to run full time -- as well as leveraging Stable Diffusion for art generation. We're launching an experimental media lab, https://ift.tt/4eU2vq5 along with the show, for follow-up projects in the future. Feel free to reach out to me personally if you have any questions or want to discuss more, skyler at mismatchmedia dotcom. https://ift.tt/AFvDoBG December 2, 2022 at 11:00PM
Show HN: I am building a Safari Web Extension to switch tabs efficiently I like the tab-switching behaviour in Chrome/Firefox and wanted Safari to have something similar. Another wish of mine was for it to be a web extension that feels native to the browser vs an app extension that looks out of place. After waiting long enough, I decided to build one myself. A little preview of Tap Tab - https://ift.tt/S6JBk5Q . The web extension APIs for managing tabs are pretty locked down in Safari, but I have already found a lot of value in using the extension. Some of the features that I find very useful(in the middle of building a few of these) : - Driven by keyboard shortcuts. I can see the tabs I have open at a glance and switch between them. - Fuzzy search to quickly filter down to the tab you want. - Quick switching between two tabs. - Ability to pin, mute, and close the tab right from the extension. Say hello, and tell me what features you would like the extension to have!! You can follow the updates on Twitter- https://twitter.com/ManiBatra23 . https://ift.tt/suLS6gy December 1, 2022 at 07:25AM
Show HN: Notion Standup – Get daily reports on your tasks on Notion Notion Standup helps you to receive daily standup reports on email and other channels for your and your team's current and overdue tasks on Notion. Stay updated always with daily reports through email, Slack, Discord, Webhook, Zapier, and Telegram. https://ift.tt/6KqV5vp December 1, 2022 at 07:10AM
संदेश
