@RhysSullivan The question is what gives humans their logical thinking that goes beyond just language understanding, and if the language models themselves can get there.
@theo@AnthropicAI If you think Claude code limits are bad, wait till you see Claude web - I'm getting session limited with 4 Sonnet 4.6 responses itself. They've nerfed it so bad since nobody's talking abt it.
"They just need to be really smart because it's hard to innovate in any category if you can't synthesize large amounts of information quickly."
@pmarca Min IQ to be considered really smart?
@RhysSullivan@badlogicgames It has 800 users within our org itself on an enterprise partnership. You are underestimating the number of such enterprise users.
@taylorotwell AI SDK Streaming seems to be lagging in Octane. Had to change from yield to echo for it and now it's a lot more laggy than it was before.
@RhysSullivan Seems cool - is executor basically a grep tools instead of loading all tool schemas into context-> Run them using code itself instead of LLM tool call?
But instead of the model just emitting a structured JSON tool call, it now has to write TypeScript though?
@paraschopra "Humans are highly attuned to attending to other people in a physical setting"
Counterpoint - I'd rather a robot handle my caretaking when i'm older than a real human. Less embarassing.