So Tachi's avatar
So Tachi
so@nuts.cash
npub1dfed...lc7u
building nostr:nprofile1qqs8wfkyxlx00y0kmmvhmwkps3hxyqv7t77jfapwnke0vs8jx8pupxsx8yfv2
So Tachi's avatar
So Tachi 3 weeks ago
the WEF this year is actually entertaining
So Tachi's avatar
So Tachi 3 weeks ago
there is 2 ways to use AI. loop over your models, then loop over your agent, then loop over the loops to debug the error of the previous loop. or use your brain, cherry pick the code you want debugged or improved. both yields similar results, none is faster than the other, but the second is an order of magnitude cheaper.
So Tachi's avatar
So Tachi 3 weeks ago
opencode is awesome, and tbh, it might even become one of the most used open source software, both as a cli or as a dependency. but it's not quite as good as other closed source agent far from it!
So Tachi's avatar
So Tachi 1 month ago
Can you practice being grateful? That might be my new year resolution
So Tachi's avatar
So Tachi 1 month ago
the email: "Love where you're going with this! Just a few tiny tweaks to really make it sing ✨" the prompt: "translate 'this is dogshit, start over' into corporate speak that won't make them cry"
So Tachi's avatar
So Tachi 1 month ago
I get more bullish in Bitcoin every time I use AI
So Tachi's avatar
So Tachi 1 month ago
I haven't yet used multi-agent or the ralph plugin yet. I don't feel the need for it. I want my model on a leash, reporting to me every step of the way. how are you handling this multi-agent, infinite loop setup?