@zenarinn They are smart enough to know that everyone thinks Indians are gross. Anything they can do to distance themselves increases their odds of positive outcomes.
The most deranged thing about zei_squirrel’s meltdown is that they achieved AI psychosis from GOOGLE GEMINI. It’s like developing a gambling addiction from Dave and Buster’s.
@norpadon They aren't written in Python. Python just calls the underlying C via FFI. And the Python bit (and the FFI call) is absurdly slow. Literally none of this is true.
@MatthewParrott@ArtemisConsort Lmao it shows literally the exact opposite. ML 101 is the bitter lesson. To anyone who actually knows what they're talking about, it's incredibly obvious that you don't.
@MatthewParrott@ArtemisConsort This is anthropomorphism. It's a formula. If you don't provide good inputs to the formula, you can't expect good outputs.
Structured logging is absolutely not a prereq or even a good way to do context management. I'm skeptical you know what that actually is.
@atomkirk@lincwilb You are using free (dogshit, old, largely useless) models and your entire world view is colored by it.
If they only make a mistake 5% of the time in the work you're doing, you should be very alarmed about whether your work will continue to exist.
@MatthewParrott@ArtemisConsort The only people who could earnestly make the argument that harnesses and tooling are the bottleneck are people that have absolutely no idea wtf they are talking about.
@UnmarredReality Advice for people that don’t want to sound retarded on Twitter:
Understand basic logical fallacies and falsifiability before posting stupid takes