Meet Strix, my AI agent
This one covers:
- an intro from Strix
- architecture deep dive & rationale
- helpful diagrams
- stories
- oh my god what's it doing now??
- conclusion
https://t.co/CKZqHSGf2Q
@jimrandomh i saw this in agent->agent messaging and Astra said it looked like the compact words were used as pseudo IDs or nicknames for bigger concepts
so it’s sort of inventing language on the fly
@deepfates e.g. someone in marketing loved Strix. but she tended to use it for tasks that were small excursions off the beaten path
whereas in programming, my this week is determined by what happened last week. Everything builds
@deepfates collapse, error cascades, idk
honestly it’s probably the same reason vibe coding takes a lot of skill to make non-trivial well-maintained apps
it’s like riding a unicycle where it gets harder the longer you stay on
@deepfates openclaw is wretched, hermes, all of them
imo state is very difficult to do right. theres a chance CLMs could work, because they reduce the complexity of the problem by keeping mutable state small, but i haven’t pushed them hard enough
hey @thsottiaux any chance we can use TTS models in chatgpt work? would be great to generate podcasts & shit
mainly, been doing a lot of learning, wish i could shift it off text for driving
@rickasaurus the way i see it, it fills a very large gap between an IF statement and a whole ass agent. i don’t think it’ll ever be Astra, i don’t think we want it to be
@siliconcodesign every time someone says this, peek under the covers and their reasoning is because AI will get cheaper. But… yah no it works the opposite way. If you thought AI would get more expensive than sure i’ll bite