“why not pay experiences machinists to videotape what they’re doing os their work can be automated”
none of these people had a job in manufacturing, or even possibly any real job, it's called CNC and there are way better tools for making it work than stochastic parrot
e: ok at least op is a machinist
Interestingly, many of these models also score at or above the level of some human experts on visual reasoning benchmarks like MMMU. That which is easy to measure often doesn't correlate with real world usefulness.
benchmarks could be perhaps possibly fucked with? say it aint so!
My high level impression when reading the response is “someone who can parrot textbook knowledge but doesn’t know what they’re talking about”.
with corpus consisting of toy problems only, and with solutions that depending on field can kill whoever uses these
all your favourite boutique ocean-boiling nonsense generators combined, compared to openai products, have something like fifth of their userbase on mobile app and something like 2/3 on website. without counting deepseek (data from february - it was very fresh back then), it's closer to eigth on app and third on website