If software and system architecture is solved by AI and people stop caring about prioritizing trainjng, why would we stay on CUDA and SRAM over Cerebras, or even silicon long term?
If people feel like the models are already so good they are always right, (1/?)
If software and system architecture is solved by AI and people stop caring about prioritizing trainjng, why would we stay on CUDA and SRAM over Cerebras, or even silicon long term?
If people feel like the models are already so good they are always right, (1/?)
Unless I am ignorant and there is a faster way to infer after software and infrastructure creation are no longer bottlenecks, (inb4 'just have the AGI surpass the speed of light' lol), I would think the average person would prioritize paying for that speed before training data
Why wouldnt we assume that we then quickly leverage that to create the "perfect" robotics programs and tools to then create the fastest theoretical inference hardware that would use almost entirely light and optics?
People would rather pay for the speed than more training, right?
The entire SWFL local government has always felt extremely corrupt.
LOOK UP CARMINE MARCENO THE SHERIFF OF SWFL, WHO THE CORRUPT DOJ JUST DROPPED HIS CORRUPTION CASE WHERE HE WAS FUNNELING TAX DOLLARS TO HIMSELF.
He is about to run as Donalds Congress replacement.
Go vote
I grew up around Fort Myers, where Byron Donald represents in the US House of Reps. Please do not fucking vote for him as Gov. He is as sleazy as they fucking come. If you are young GO VOTE!!!!!
AGAIN: GO VOTE THERE IS NO EXCUSE
DO NOT LET DYING BOOMERS KEEP OUTVOTING US
Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here and get back to R&D. I remain deeply passionate about education and plan to resume my work on it in time.
Claude Code 4.7 is insane.
i know literally NOTHING about coding. ZERO. and i just built 3 fully functioning web apps in 30 minutes.
http://localhost:3001/
http://localhost:8000/
http://localhost:5000/
check it out.