Claude API error: {"type":"error","error":{"type":"invalid_request_error","message":"Your credit balance is too low to access the Anthropic API. Please go to Plans & Billing to upgrade or purchase credits."}}
SlopCharger: Broker free-tier accounts with fly-by-night AI providers cross-matched with the latest in prompt exploits to enable new billing efficiencies by running your HN AI shitpost on someone else's one-fiftieth of a dime! :sparkles:
> DocuQuest: A platform that leverages LLMs to transform and simplify complex technical documentation into interactive, user-friendly learning experiences tailored for developers and engineers.
So "A Young Lady's Illustrated Primer" from Diamond Age, but for devs. A neat idea!
In general docs ecosystems tend to be heavy on only one of reference / explanation / tutorial. Would be cool to have a way to write one and get the others.
Alright, but DocuQuest is actually a really good idea.
"DocuQuest: A platform that leverages LLMs to transform and simplify complex technical documentation into interactive, user-friendly learning experiences tailored for developers and engineers."
I recently found a vibe-coded app that generates courses on the fly from YouTube, including automatically generated quizzes. I forgot the URL, but the result was beyond awful. Questions were something like: "what was the title of the youtube video?" and other utterly stupid things.
Not saying that it isn't possible, but stuff like this does need the human touch.
Very cute but wish there was a filter for the flavor of the week ... I feel 60% of them will have LLM because of this bias. Last month would've been "powered by rust."
I've received nothing but nice messages about the site after posting it and people seem to be getting a kick out of it. I feel like this (if barely so) meets the bar to give it a pass perhaps, especially considering that my voice is worth as much as yours and the next person, so without upvotes, it will disappear, just how it's meant to be. Hope that helps a bit.
This is not primarily an advertising site. See guidelines:
> Please don't use HN primarily for promotion. It's ok to post your own stuff part of the time, but the primary use of the site should be for curiosity.
SwearySkyscraper: A startup that develops novel and personalized swear words to help people better cope with pain, and integrates these swear words into a liquid damping system to stabilize skyscrapers during earthquakes.
A building that requires human pain to stand? I guarantee you this is something at least a few architects have fantasized at length about. Like ours, it seems to be a profession that attracts outliers.
Ideascope: An AI-powered platform that generates personalized, innovative startup ideas by analyzing real-time trends and data from Hacker News and other tech communities
>SciDigest: An AI-powered platform that transforms complex scientific research into engaging, digestible content for the general public, making scientific knowledge accessible and actionable.
This would actually be great. So many researchers have a marketing problem with explaining and getting people excited for their work.
I'm a popular science writer with eight year's experience doing exactly this (SciShow, Crash Course, Veritasium and recent winner of the Wellcome Collection Non Fiction Awards) without AI. Done right, the right coverage of even a pre-print reached hundreds of thousands/millions of people. But I've experimented with every SOTA model since 2022 with the most detailed and specific prompting I can think of (including multiple examples of transcripts of work already in the public domain) to see if it can replicate good quality science communication.
The content is usually reasonably strong but the tone is always off and it never quite understands what it is a reader/viewer needs to really get to grips with the topic if they don't already have a prior foundational understanding (though I notice this about a lot of other media outlets with professional science communicators too). It also has poor editorial thinking around what bits are most likely to be interesting and cohesive when considered as part of the whole piece.
But I'm still reasonably convinced as AI improves it ought to be able to replace me with the right workflow/context/prompting. I think there will always be a demand for my (and many other writers') talents as they are so it doesn't really bother me, but it'd be great to extend the work to all the many scientific discoveries that don't get the same attention. If anyone is serious about developing something like this, I'd be interested in partnering with them as someone with domain expertise on science communication and familiar with prompt engineering (email in bio).
One question is whether the audience is discerning enough to care about the issues you're raising. This seems like it could be a variation of why the umpteenth Marvel movie beats out indie masterpieces at the box office. The audience for high quality becomes increasingly niche as the market's relatively low bar for quality is satisfied.
I think you're right about the editorial thinking + what do people find interesting parts. But that doesn't have to be solved by directly by AI, it's easy enough to sidestep the problem and provide a nice interface for the human-in-the-loop part. I'd imagine that would save you a ton of time by having a nice starting point depending on how much you have to rewrite for tone.
That's true, it could just turn the writer's role into more of an editorial role. The main time-saving I have so far is being able to upload papers and get it to fact check for me. The editorial guidelines at SciShow are stricter than any academic journal I've published in: any non-trivial statement has to be supported by a direct, findable quote in (most-of-the-time) peer-reviewed scientific literature. I once had to find a citation for the idea that heat + fuel + oxygen generates a fire! (for this video: https://www.youtube.com/watch?v=BEcaE0e0CZg)
LLMs make that much easier. As I collect primary sources during my drafting/writing phrase, I can type up any non-trivial claims I'm making in my script in a separate document, share that with the LLM and say "Quoting directly from the set of attached PDFs, identifying which document, and on which page the quote comes from, find content which directly supports each of these assertions" and it generally goes a great job. At any rate, I have to check each of those quotes for accuracy but the help in _finding_ those quotes in order to pass a stringent fact checking procedure is a huge help if I didn't scribble down the supporting quotes during my research phase. This is also, by the way, stricter than the fact checking process for most non-fiction publishing.
Feels like there might be an accuracy issue as well. Although that might make it perfectly suited to replacing whoever writes university press releases...
>I've experimented with every SOTA model since 2022
>The content is usually reasonably strong but the tone is always off and it never quite understands what it is a reader/viewer needs
A SOTA model fine-tuned with your choice of transcripts could probably get you most of the way there. There might be a customized, open-weight model already on Huggingface that meets your needs.
I've been using the NotebookLM "podcasts" for this. I upload an arxiv pdf and use the Interactive Mode to have them talk about it, while I can pause them to ask clarifying questions. It's been surprisingly efficient for me to get an initial grasp on things I'm not familiar with, at times prompting me to then go on exploring interesting rabbit holes.
I've attempted this with pdftomp3.com where you can listen to PDFs. It has an "AI Explanation" mode where the content of the PDF gets explained. Its like a NotbookLM podcast, but I was earlier :)
Currently I'm working on an app for that, because thats where I listen to the MP3s anyways.
>Biomech Innovations: A wearable tech startup that uses genetic algorithms and AI to help zebrafish regenerate damaged organs, with a secrets management platform to securely store user data and a retro-futuristic assembly-level code editor for custom hardware designs.
Instant fun! Honestly, whatever tech I look for, I use built-in search engine of Hacker News first before googling it.
Lets not waste your $0.0005 and share this pretty decent idea with everyone:
> Graphene Labs: A decentralized, open-source platform for creating and sharing interactive data visualizations and infrastructure diagrams that can be hosted locally and integrated with real-time data sources
Ha! Fun stuff. While some ideas are actually intriguing, many if its suggestions seem to be overly vague jumbles of common phrases and technology. "AI-powered databases to leverage personalized accessibility for team management..." Lol. Still fun though.
Code itself is not very valuable as well. What has value is the combination of all those things and application of them to some real problem in a form of complete solution.
i was not expecting people to like the startup ideas so much, but it's a pleasant surprise!
i thought it'd be cool to let people vote on ideas that HN Slop came up with, so now you'll see an "i'd invest" button & that will let others vote on the idea on a leaderboard
hope y'all like it, keep sending the feedback, I'm listening!
> CosmicPlay: A startup that leverages GPU emulation and graph theory to build a next-generation video game engine that can render celestial bodies with nuclear propulsion and solar sailing capabilities, while generating AI-powered gameplay ideas from Hacker News posts
So "A Young Lady's Illustrated Primer" from Diamond Age, but for devs. A neat idea!
In general docs ecosystems tend to be heavy on only one of reference / explanation / tutorial. Would be cool to have a way to write one and get the others.
"DocuQuest: A platform that leverages LLMs to transform and simplify complex technical documentation into interactive, user-friendly learning experiences tailored for developers and engineers."
Not saying that it isn't possible, but stuff like this does need the human touch.
This is effectively an advertising website.
> Please don't use HN primarily for promotion. It's ok to post your own stuff part of the time, but the primary use of the site should be for curiosity.
LLMs will make this very hard to detect, very soon if not already.
Yep, I think this may be possible :p
This would actually be great. So many researchers have a marketing problem with explaining and getting people excited for their work.
The content is usually reasonably strong but the tone is always off and it never quite understands what it is a reader/viewer needs to really get to grips with the topic if they don't already have a prior foundational understanding (though I notice this about a lot of other media outlets with professional science communicators too). It also has poor editorial thinking around what bits are most likely to be interesting and cohesive when considered as part of the whole piece.
But I'm still reasonably convinced as AI improves it ought to be able to replace me with the right workflow/context/prompting. I think there will always be a demand for my (and many other writers') talents as they are so it doesn't really bother me, but it'd be great to extend the work to all the many scientific discoveries that don't get the same attention. If anyone is serious about developing something like this, I'd be interested in partnering with them as someone with domain expertise on science communication and familiar with prompt engineering (email in bio).
I think you're right about the editorial thinking + what do people find interesting parts. But that doesn't have to be solved by directly by AI, it's easy enough to sidestep the problem and provide a nice interface for the human-in-the-loop part. I'd imagine that would save you a ton of time by having a nice starting point depending on how much you have to rewrite for tone.
LLMs make that much easier. As I collect primary sources during my drafting/writing phrase, I can type up any non-trivial claims I'm making in my script in a separate document, share that with the LLM and say "Quoting directly from the set of attached PDFs, identifying which document, and on which page the quote comes from, find content which directly supports each of these assertions" and it generally goes a great job. At any rate, I have to check each of those quotes for accuracy but the help in _finding_ those quotes in order to pass a stringent fact checking procedure is a huge help if I didn't scribble down the supporting quotes during my research phase. This is also, by the way, stricter than the fact checking process for most non-fiction publishing.
Now there's a testimonial. I look forward to browsing the source links with each video!
>The content is usually reasonably strong but the tone is always off and it never quite understands what it is a reader/viewer needs
A SOTA model fine-tuned with your choice of transcripts could probably get you most of the way there. There might be a customized, open-weight model already on Huggingface that meets your needs.
Currently I'm working on an app for that, because thats where I listen to the MP3s anyways.
Instant fun! Honestly, whatever tech I look for, I use built-in search engine of Hacker News first before googling it.
You should use something like openrouter or portkey or similar for managing fallbacks
thank kier for claude code
What pieces of openrouter are open source? I checked out their main github repo, and it hasn't had any contributions in months.
Wish they would have used a different name.
i thought it'd be cool to let people vote on ideas that HN Slop came up with, so now you'll see an "i'd invest" button & that will let others vote on the idea on a leaderboard
hope y'all like it, keep sending the feedback, I'm listening!