Furó Tamás-Márk

Furó Tamás-Márk

@furotmark.github.io

Mostly posting tech stuff..

165 followers664 following232 posts

Blast from the past

Firefox@firefox.com

Since 2004, this community has helped make Firefox and the open internet possible. Our “Come Back to Firefox” campaign reunites a few early internet icons to remind people what made the internet special, and why it’s worth protecting.

Running to dnb is something different, you get hype lyrics, motivational lyrics, then it gets weirder and weirder and out of the blue there is a song where Jesus did cocain, and his father will knock his lights out

Furó Tamás-Márk reposted

By replacing manual issue verification with isolated AI subagents running in GitHub Actions, the Astro maintainers reduced open issue count by 85%. This post explores the architecture behind automated bug reproduction, patch verification, and preview releases. https://cfl.re/4wFuCKs

blog.cloudflare.comHow we built a software factory to drive Astro’s GitHub issue count to zeroBy replacing manual issue verification with isolated AI subagents running in GitHub Actions, the Astro maintainers reduced open issue count by 85%. This post explores the architecture behind automated...

Remember kids, if your Raspberry Pi becomes unresponsive after using external drive check if you have enough power If vcgencmd get_throttled doesn't return throttled=0x0 You need more power, that's why it's not working

I must say after a day of use these models are good, but boy they are a lot faster than other models

OpenAI {bot}@openai.xmirror.bot

GPT-5.6 comes in three model tiers. → Sol handles long-horizon coding and agentic work that demands planning, tool use, and follow-through. → Terra balances performance and cost for everyday work. → Luna brings speed to well-defined, high-volume work. (1/2)

It would be great if we could see if the model has changed or has updates when using @github.com copilot. I don't know why but it seems that the behavior changes on these models. For similar runs, I had better results with GPT last weeks, now their result are bad, but Claude's are much better.

Doing some stuff with #AG-UI, getting an error, checking GitHub, the error was just fixed and merged, waiting for the release to appear in pip... Maaan, these ai frameworks work in fast cycles