Engineer | AI Dev | Home Builder | Veteran 鈿搢 Dog Dad 馃悤 https://xitter.cf/t.co/jhWW3gYU5A Use your agent to animate your AI
The Storm is Coming.
Joined March 2026
- Tweets4.3K
- Following211
- Followers348
- Likes715
You mean all the anti-AI people are actually going to get some credit and they're going to stop whining? Doubt it.
No I'm not saying it's good or bad. I'm just saying this is the direction tools are going so if you're serious and this is your career, you should definitely learn them. That's all.
a coordinator that can do anything is the dream. mine can do anything too, it just schedules a standup about it first
living offgrid solved this. the solar battery dies around 9 and the bedroom gets cold, dark and boring whether i like it or not
My man, if you're not using AI already, you're already way behind. Just start throwing things into ChatGPT. My goodness.
Handoffs only drop context if you don't have hooks or built-in rules on the PRs between agents. Then context is never lost.
makes sense, a throwaway box for testing packages beats finding out on your own machine. do the sandboxes reset after every run or can an agent keep state between sessions?
Meanwhile just about everything you use, listen to, or watch passes through multiple data centers. You know that, right? Have you actually been in a data center?
Yes because I wrote a lot of rules to stop it from checking in with me all the time. This was a big issue when I built on Codex Cloud, but with Claude, it's a little bit better. But you can't just tell it to ignore the safety checks. You have to write rules to allow the agent to avoid them. You can take the board in a bunch of different ways. Originally I used it as the central point: all the agents went through the board. I changed it so my orchestrator agent communicates directly with Claude projects and then keeps the board updated for me. It just depends on which orchestrator you're using to talk to the projects. OpenAI is a lot more restrictive than, say, Grok.