A lot of them see themselves as Oppenheimer building the Bomb and that’s the only justification left for ramping up spending, so that’s the rationale they’ll use.
Comments
Log in with your Bluesky account to leave a comment
If American companies adopt similar architectures, can’t they potentially enhance their models’ efficiency, create superior AI models and while reducing projected capex?
I wonder how many of those decision makers are kind of pot-committed right now. If you advocated for your organization to spend huge sums of money on AI infrastructure, it’s kind of hard to be like, whoops, that was a mistake…
I wish these people could explain what the end use of their products will be. Bonus points if they tell me the practical use of these products or when they'll actually make money
After a few drinks they’ll likely say replacing service jobs, but what Id say to that is most service jobs are in the physical world, and the worlds of robotics and manufacturing have not found any use in LLMs. Reminding me a lot of Elon “FSD one year away” claims in the 2010s
Comments