Ask HN: Has anyone built anything useful using AI?
20 comments
Lots of "information extraction" tools - basically taking in a bunch of documents and other natural language material and pulling out a few key pieces of structured data. This isn't 100% reliable but is useful as a first pass (where a human can investigate further) or a backstop (against a human missing something important and obvious but buried in a long file or huge volume of files).
I also put together a half-baked web extension to try to steer the Youtube algorithm. It basically scrolls through shorts and tries to watch+like high quality content and skip+dislike junk (and worse) to scrub it out of the feed as much as possible. It's only looking at the transcript and a couple of thumbnails, so it's not super accurate, but in the short term it's been working pretty well. I figure that Google can probably tell that it's not a human though, and might disregard its inputs (or ban me) in the long term. Time will tell.
I also put together a half-baked web extension to try to steer the Youtube algorithm. It basically scrolls through shorts and tries to watch+like high quality content and skip+dislike junk (and worse) to scrub it out of the feed as much as possible. It's only looking at the transcript and a couple of thumbnails, so it's not super accurate, but in the short term it's been working pretty well. I figure that Google can probably tell that it's not a human though, and might disregard its inputs (or ban me) in the long term. Time will tell.
I generate unit tests, comments, and more. A critical aspect is well-defined system instructions. Defining system instructions and prompts with XML formatting provides significant benefits. Think of LLMs as a force multiplier for most use cases. Upload the codebase and generate comments, unit tests, functional tests, mc/dc tests, etc. What follows is a matter of review and edits to outputs. Never unquestioningly trust, and instead, slowly train and refine models for any given project, as this approach tends to reap rewards in my experience.
I've built a lot of useful things with AI!
- A pretty full features terminal UI application that works well with my company's quirky hosted jira which is nice and fast. Used to and tview.
- Simple game to learn c++, ecs, and game dev. Basically making a rougelike co-op scorched earth tank game.
- Lots of little bits of wezterm, neovim, and zsh stuff to help my workflows. Things like a wezterm project selector, a fzf function to connect to remote hosts, etc
- Converted my resume to latex. Made it way easier to apply to jobs.
- Heavy planning and refactoring for apps at work
- A pretty full features terminal UI application that works well with my company's quirky hosted jira which is nice and fast. Used to and tview.
- Simple game to learn c++, ecs, and game dev. Basically making a rougelike co-op scorched earth tank game.
- Lots of little bits of wezterm, neovim, and zsh stuff to help my workflows. Things like a wezterm project selector, a fzf function to connect to remote hosts, etc
- Converted my resume to latex. Made it way easier to apply to jobs.
- Heavy planning and refactoring for apps at work
Yes. I've build a couple of tools for myself as experiments. Nothing fancy, just CRUD apps in PHP & SQLite. Worked really well. I'm pretty weak on SQL and databases, and i was able to get them set up and working in one shot. Saved me days of farting around.
I have. The advent of "reasoning" llms means you can trust them with judgments.
Chatgpt basically invented it in september of 2024. We are only coming up on the concept's 1st birthday. Open source options have only been out for 6 months.
I highly recommend using these for your own purposes.
Chatgpt basically invented it in september of 2024. We are only coming up on the concept's 1st birthday. Open source options have only been out for 6 months.
I highly recommend using these for your own purposes.
some stuff i built using AI
- a personal trading manager: automatically reads signals from telegram, send the signal to an openai agent using gpt-4o with a custom MCP server I built for decision making and order placement
- a monitoring system on my home servers: claude code runs every5 minutes and update sql records regarding some monitoring metrics. it can also warn me if there is an issue vie email (using resend).
currently i'm working on a dashboard for remote jobs tracking and curation. im planning to use a scrapper and then a model that will curate and record the jobs with useful information for people who wants to apply
currently i'm working on a dashboard for remote jobs tracking and curation. im planning to use a scrapper and then a model that will curate and record the jobs with useful information for people who wants to apply
I have used coding agents with great success. So in some way I'm building useful things using AI, and the people who created those agents absolutely created something extremely useful.
I used Coding Agent to create an interactive screen for a small demo. I use it every time I do a demo, and it's very useful.
I think people in the content business have probably seen the most significant boost from AI, so far.
yes!
Use it everyday https://github.com/jharohit/team-timezone-wall
Also built a AI NDA tracker for all our NDAs in the company which is awesome! We will open source soon.
Use it everyday https://github.com/jharohit/team-timezone-wall
Also built a AI NDA tracker for all our NDAs in the company which is awesome! We will open source soon.
I'm having trouble seeing how AI is used/needed here, can you explain?
Yup, first project is a US Dollar currency converter tool online.
No. They are hype.
People are missing the point of the question. OP is asking if you've built a tool that uses AI without it just being a standard chat wrapper, not if you've built something with code generated from an AI model.
Yep.
A lot of the tools I see are chat wrappers or code generators.