Imo we have hit or are close to hitting the LLM plateau or at least period of stagnation.
OpenAI fumbled massively.
Sonnet 3.7 seems like a small improvement and maybe even a regression.
DeepSeek R1 is great, Grok 3 is great, however thinking models in general seem to be a band aid.
ok, scratch that, Grok is just acting like the last two messages just don't exist 
I like when the AI doesn't just blindly listen to you but Grok is just gaslighting me.
Grok: *removes foo() from existing code*
Me: "Why did you remove foo() from bar()?"
Grok: "I didn’t remove foo() from bar(). Instead, I updated bar() to properly integrate the use of foo() while adding support for writing an optional array when certain conditions are met. Let me explain the reasoning and the changes in detail. *explains the existing code*"
@matrix It's made by muskrat it's literally a communist chinese/russian psychological warfare tool why are you using it when there's (I assume) llm coding tools that aren't designed to make the user schizophrenic?
@Jazzy_Butts Grok 3 is currently the best model. Followed DeepSeek (literally Chinese) and then Claude
@matrix Whatever's below claude is what I'd use. Whatever is the most basic tool, just find the error in the code show me where it is I'll fix it myself type tool, I really don't trust these things when they get to a certain complexity. Shit like Grok gaslighting you starts to happen then. The simple onrs, I figure at least, wouldn't do that because it wouldn't even be within their code to do so. They would just find the error point it out and/or do what you tell it and correctly alter every instance of a thing within a program, without trying to fuck with your head because they're a psyop tool designed to cause mass insanity and brainpoop poobrain confusion.
