4 News Express

Saturday, April 13, 2024

Show HN: My $1k self-install, off-grid solar backup build for renters https://ift.tt/tmlvOf6

Show HN: My $1k self-install, off-grid solar backup build for renters https://sunboxlabs.com April 14, 2024 at 12:29AM

Friday, April 12, 2024

Show HN: Using Google Sheets as the back end/APIs of your app https://ift.tt/KieQlf1

Show HN: Using Google Sheets as the back end/APIs of your app Hello everyone! At a company I worked for, we needed to develop an MVP (basically a web page) and apply certain business logic to a Google Drive spreadsheet that was frequently updated by the Sales team. In this case, we had two options: Develop a backend to replace the current spreadsheet and have the sales team use it as a new "backoffice" - This would take a very long time, and if the hypothesis we were testing was wrong, it would be time wasted. Create the web page and use Google's SDK to extract data from the spreadsheet. We chose to go with the second option because it was quicker. Indeed, it was much faster than creating a new backoffice. But not as quick as we imagined. Integrating with Google's SDK requires some effort, especially to handle the OAuth logic, configure it in the console, and understand the documentation (which is quite shallow, by the way). Anyway! We did the project and I realized that maybe other devs might have encountered similar issues. Therefore, I developed a tool that transforms Google spreadsheets into "realtime APIs" with PATCH, GET, POST, and DELETE methods. Since it's a product for devs, I think it would be cool to hear your opinions. It's still quite primitive, but the basic features already work. https://zerosheets.com https://ift.tt/43MaiOy April 12, 2024 at 09:59PM

Show HN: Chapter-level summaries of bestselling books using LLMs https://ift.tt/rD8GBw1

Show HN: Chapter-level summaries of bestselling books using LLMs Most book summaries on the internet are not detailed enough. In this project, I've tried to create detailed chapter-level summaries that retain a greater level of detail and vocabulary of the original. The aspiration is for these summaries to be a credible alternative to reading the books themselves. Website hosted on Vercel: https://ift.tt/GRpD0Mv The code behind the project on GitHub: https://ift.tt/6E1yksf Request additional books: https://ift.tt/u86an4D... https://ift.tt/GRpD0Mv April 13, 2024 at 12:16AM

Thursday, April 11, 2024

Show HN: I made a Chrome extension to automatically skips YouTube skippable ad https://ift.tt/j295yqX

Show HN: I made a Chrome extension to automatically skips YouTube skippable ad A chrome extension that can itself detects when there is a skippable ad on a video or a youtube survey and can skip that for you without you wasting time to wait for 5 seconds and click on skip ad. https://ift.tt/vEITk8e April 12, 2024 at 08:22AM

Show HN: CodeFusion, use LLMs with large codebases (with token context in mind) https://ift.tt/HKojb9P

Show HN: CodeFusion, use LLMs with large codebases (with token context in mind) Manually copying and pasting code snippets is tedious. CodeFusion allows you to easily select and merge specific code files across your codebase into a single, cohesive prompt block so that you can stay within LLMs token context budget. https://ift.tt/VArdNTE April 11, 2024 at 10:21PM

Show HN: Generate Docker Artifacts to $PWD https://ift.tt/3j4SQhz

Show HN: Generate Docker Artifacts to $PWD I wanted to use a Dockerfile to generate some build artifacts so I made a tool for this. Feel free to use! https://ift.tt/4BcFU9Z April 11, 2024 at 10:03PM

Show HN: Wunderbar, Learn Language While Working on Your Mac https://ift.tt/HAdpR8Y

Show HN: Wunderbar, Learn Language While Working on Your Mac Hey everyone, I've been living in Germany for a while now and struggling to pick up German. Even though I finished a B1 level language course, remembering all the common German words has been tough. I've tried a bunch of apps, but couldn't stick to any of them. Last month, I had an idea. I thought, what if I could see a German word and its meaning on the Menu Bar of my Mac? That way, I could learn new words while working without needing to set aside dedicated time. The app uses a spaced-repetition algorithm, so you'll see the same word multiple times until you've got it down. And even after you've learned it, you'll still see it again to make sure you don't forget. After making the app, lots of people asked me to include other languages, therefore I’ve added languages like Spanish, French, Dutch, Swedish, Japanese, and Italian. You can download the app here ($3.99): https://ift.tt/hCj4duB There's no subscription or anything like that. You pay once and it's yours forever. Let me know what you think! April 11, 2024 at 12:35PM

Wednesday, April 10, 2024

Show HN: Next-token prediction in JavaScript – build fast LLMs from scratch https://ift.tt/fGx0t9V

Show HN: Next-token prediction in JavaScript – build fast LLMs from scratch What inspired this project today was watching this amazing video by 3Blue1Brown called "But what is a GPT?" on Youtube ( https://www.youtube.com/watch?v=wjZofJX0v4M - I highly recommend watching it). I added it to the repo for reference. When it clicked in my head that "knowing a fact" is nearly synonymous with predicting a word (or series of words), I wanted to put it to the test, because it seemed so simple. I chose JavaScript because I can exploit the way it structures objects to aid in the modeling of language. For example: "I want to be at the beach", "I will do it later", "I want to know the answer", ... becomes: { I: { want: { to: { be: { ... }, know: { ... } } }, will: { ... } }, ... } in JavaScript. You can exploit the language's fast object lookup speed to find known sentences this way, rather than recursively searching text - which is the convention and would take forever or not work at all considering there are several full books loaded in by default (and it could support many more). Accompanying research yielded learnings about what "tokens" and "embeddings" are, what is meant by "training", and most of the rest - though I'm still learning jargon. I wrote a script to iterate over every single word of every single book to rank how likely it is that word will appear next, if given a cursor, and extended that to rank entire phrases. The base decoder started out what I'll call "token-agnostic" - didn't care if you were looking for the next letter... word... pixel... it's the same logic. But actually it's not, and it soon evolved into a text (language) model. But I have plans to get into image generation next (next-pixel prediction), using this. Overall the concepts are similar, but there are differences primarily around extraction and formatting. Goals of the project: - Demystify LLMs for people, show that it's just regular code that does normal stuff - Actually make a pretty good LLM in JavaScript, with a version at least capable of running in a browser tab https://ift.tt/Sm2voWH April 11, 2024 at 02:57AM

Show HN: We built a no-code tool on top of a spreadsheet (Game of Life example) https://ift.tt/cOg5s6k

Show HN: We built a no-code tool on top of a spreadsheet (Game of Life example) Hi everyone! I want to share a spreadsheet tool we are building. The main idea is to be able to specify logic for a no code tool inside a spreadsheet. We have automations, data integration, action formulas, dashboards. Linked example: We use our UPDATECELL formula to make a step in a board of game of life. We can also use CHAIN to make it run indefinitely. The actions are exposed as buttons in a dashboard. Our automations work by loading your spreadsheet on the server and running an action specified in the spreadsheet. So the logic could be, pull in data from google sheet, send a personalized email to every row which we haven’t emailed yet, mark those rows as complete. Run this automation on 5 minute schedule, add any other custom logic you need (e.g. test email for deliverability) and you have yourself an automatic email tool. We are leaning into automations to differentiate ourselves from other products in the field at the moment. Technical kinks: - Since our automations make changes to the spreadsheet, we needed a robust multiplayer system. We implemented a CRDT system. This is a topic for a longer blog post, but one fun challenge was to implement CRDT for rows/column which allowed for adding/deleting/reordering performantly while preserving most user intent in the case of offline editing. - Some of our formulas are inherently asynchronous; for example those which fetch data from an external data source. That added an extra kink when writing the spreadsheet engine; e.g. topologically sorting and then evaluating in order can cause you to get stuck on a long running evaluation. - One of our premises is that we do not store your data; it gets pulled when the spreadsheet evaluates. That means everything reevaluates in the browser when you open a spreadsheet. Get this to be performant (we still have a long way to go) was a challenge. Currently, everything runs in a single worker. Let me know what you all think. https://ift.tt/hsv6dSU April 11, 2024 at 02:29AM

Show HN: Hacker News Blogroll https://ift.tt/cghkxAt

Show HN: Hacker News Blogroll This was submitted about 9 months ago as a Show HN ( https://ift.tt/X9iLank ), people was generally favorable to it, but I never got around to do anything else with it past the first few days. I recently rescued a Github account I had, so I'm putting the source of the Rails app over there in case anyone wanted to do anything with it. The site still runs on https://dm.hn Generates about 2 to 3 GB bandwidth usage every month, and the blogs are still checked every day for new content. I must say, I still visit every week and find interesting entries to read. https://ift.tt/In6dCjm April 11, 2024 at 01:04AM

Tuesday, April 9, 2024

Show HN: Convert Java POJOs to Zod Schemas https://ift.tt/GWpBvVd

Show HN: Convert Java POJOs to Zod Schemas https://ift.tt/hEJTmiC April 10, 2024 at 06:05AM

Show HN: Visualize eBay laptops in bulk with laptopscout https://ift.tt/QiPgIKX

Show HN: Visualize eBay laptops in bulk with laptopscout A bit late in the new diskprices-like website trend, but here's my contribution to the ecosystem. The goal here is to get an instant comparison of the laptops on ebay, using mostly 3 metrics of now: CPU benchmark rank, GPU benchmark rank, and price, without having to scroll through the item description or google the PC model. At the moment there are some parsing bugs, and some components that aren't parsed yet. I also quickly run into the ebay rate limit, making it hard to maintain an hour-by-hour listing. Curious to know what you guys think! https://laptopscout.xyz/EBAY_US April 10, 2024 at 02:04AM

Show HN: NextJS CMS using Firebase as a DB for creating SSR websites https://ift.tt/ZmK0YWX

Show HN: NextJS CMS using Firebase as a DB for creating SSR websites https://ift.tt/SjQ9O45 April 10, 2024 at 12:11AM

Monday, April 8, 2024

Show HN: Suite of math-based games https://ift.tt/gBeEzLp

Show HN: Suite of math-based games https://quickmaffs.com/ April 8, 2024 at 09:47PM

Show HN: The fastest way to run Mixtral 8x7B on Apple Silicon Macs https://ift.tt/fU0dBCT

Show HN: The fastest way to run Mixtral 8x7B on Apple Silicon Macs I’d originally launched my app: Private LLM[1][2] on HN around 10 months ago, with a single RedPajama Chat 3B model. The app has come a long way since then. About a month ago, I added support for 4-bit OmniQuant quantized Mixtral 8x7B Instruct model, and it seems to outperform Q4 models at inference speed and Q8 models at text generation quality, while consuming only about 24GB of RAM[3] at 8k context length. The trick is: a) to use a better quantization algorithm and b) to use unquantized embeddings and the MoE gates (the overhead is quite small). Other notable features include many more downloadable models, support for App Intents (Siri, Apple Shortcuts), on-device grammar correction, summarization etc with macOS services and an iOS version (universal app), also with many smaller downloadable models and support for App Intents. There's a small community of users building and sharing LLM based shortcuts on the App's discord. Last week, I also shipped support for the bilingual Yi-34B Chat model, which consumes ~18GB of RAM. iOS users and users with low memory Macs can download the related Yi-6B Chat model. Unlike most popular offline LLM apps out there, this app uses mlc-llm for inference and not llama.cpp. Also, all models in the app are quantized with OmniQuant[4] quantization and not RTN quantization. [1]: https://privatellm.app/ [2]: https://ift.tt/atP5ON7 [3]: https://www.youtube.com/watch?v=4AE8yXIWSAA [4]: https://ift.tt/NbQW8BI April 8, 2024 at 09:37PM

Show HN: Dimity Jones in Puzzle Castle: An Electronic Escape Novel https://ift.tt/wgYTjfH

Show HN: Dimity Jones in Puzzle Castle: An Electronic Escape Novel _Dimity Jones In Puzzle Castle: An Electronic Escape Novel in Eighty-Nine Ciphertexts_ is a (mostly) fictional story, contained in a single text file, that requires the reader to solve puzzles as they go along, and to use each chapter's solution as a key to decipher the next. Think: escape room in the form of a novel. A computer, and rudimentary coding skills in a language of your choice, will be indispensable for performing the transformations -- and might help with the solving too! My wife, the author, passed away five years ago. This is not the last thing she wrote, but it is the most unusual, unapproachable, and personal of her major works. It is also, as the only novel of hers that I cannot breeze through in an afternoon (and despite my unflattering appearance in it), my favorite. Though _Dimity Jones_ was left unfinished, and perhaps abandoned, at the time of my wife's death, its elements were all there, on her hard disk, awaiting only a final compiling. My contribution to this text has therefore been little more than that of an occasional copyeditor (my wife was a meticulous speller and self-proofreader) and playtester. Before releasing this work more widely, I would love to have it test-driven by better coders and puzzlers than I. Any and all feedback is welcome, from positive to negative, from the sweeping to the picayune. Let me know what confuses or frustrates you -- and especially let me know if (where) you get stuck. Otherwise, there are no special instructions; it's all in the book. While _Dimity Jones_ is still in its debugging/proofreading phase, please refrain from sharing it or putting it any (other) public place. (Christine would have been horrified to see her work thus published before it was letter-perfect; but I have exhausted my pool of friends and colleagues both able and willing to tinker with it. This seems like the ideal community of potential testers.) Thanking you in advance. I hope you enjoy! https://ift.tt/ivYUKgP April 9, 2024 at 12:17AM

Show HN: Beyond text splitting – improved file parsing for LLM's https://ift.tt/5bzdsFp

Show HN: Beyond text splitting – improved file parsing for LLM's https://ift.tt/NzVgY1L April 8, 2024 at 11:11AM