mattv8
mattv8 on Hacker News
It's also the easiest to implement, imo. But there is a tradeoff: fewer waking hours during your best years cognitively for more years when you're older...
Any particular reason why not to open source it other than you plan on selling it as a PaaS?
Thanks for the response. I'll definitely give ctx a try.
Very nice. Does this support GitHub Copilot subscriptions (oauth/hmac) or do you have plans for it? That would make or break for me because of the API costs.
Similarly I built a self-host able replit-like server with RAG but it's more end-user focused than developer focused...
Had to dig to find this but back in 2009 I was bored so I made a font based of my handwriting. I had a Wacom tablet and used this font creator- I'm pretty sure it was called Fontographer. Anyways it's still floating around the Internet:
https://fontmeme.com/fonts/mattfont-font/
I'm glad you saw it as humor- thats how it was intended. I love my wife dearly but we spend our free time very differently. I'm an engineer, she's an epidemiologist. I'm more creative and like to build things, she's more interested in watching the news. We do have great conversations about cosmology and physics though.
Too late for me then I suppose, I'm married with two kids.
What I would give for a wife that would enjoy tinkering with me.
I used to run, but now that I have kids it's impossible to find the time... I even built myself a cool run tracker dashboard (
https://www.visnovsky.us/?page=running)
Oh yeah- worth mentioning: carry pepper spray. I got attacked by a German Shephard one morning.
Posted by
mattv8 on Tuesday February 10, 2026 @01:42PM
from the show-hn dept.
https://github.com/mattv8/ragtime
RAGtime is a self-hosted MCP server and FAISS/PGVector manager that lets AI assistants run real operations on your infrastructure: SSH commands, SQL queries through SSH tunnels, git repo indexing, filesystem searches. It connects Claude, OpenAI, or Ollama to your environment via both MCP protocol and OpenAI-compatible chat completions API.I originally built this over Christmas break as a...
4 hours is an awfully big investment... Especially for those of us with multiple young kids and who no longer own their own free time. Care to give the gist?
Cheyenne Mountain Complex and Stargate are inseparably connected in my mind.
Late reply but I have a local restaurant that's willing to beta test it. I'm trying to get all the edge cases ironed out before I go live. There's a whole lot of branding logic I'm trying to account for so nervous it won't be reliable and they'll flake on me.
What do you guys think of this?
https://www.textaurant.app. It's an AI "agentic" SMS ordering system that's hopefully better than Taco Bell's attempt... I got sick of navigating every restaurants nuanced online order placing and figured I'd try to standardize it myself with an SMS based assistant (yes I'm aware of the XKCD). The idea is every restaurant would have their own number, or down the road I could have a single number for all restaurants but I'm somewhat token/context limited right now. It uses GPT 4o and I've been working on it for the past 4 months. Closed source for now but who knows I might open it up but I'm deciding if it's worth trying to patent.
This is a great article, and I largely agree but I feel like we're giving ourselves an excuse to be lazy. Because I've absolutely seen this principle swing in the opposite direction, where someone writes slop code, without ever having had a real conversation with the end user(s) and consequently the software goes out the door without having considered top 5 most common edge cases that would have been so obvious if a little more effort had been put in.