đź’Ż
One of the most important political and consumer priorities is to get these companies to stop research on more-capable-than-GPT-4-stuff now. And I think it’s important that we don’t confuse model size or training cost with capability because GPT-4 I think is likely “bigger” than a truely super human thought-worker replacement.
Regulations can be slippery slopes, often ending up favoring the strongest and largest entities.
Definitely. And always favoriting the state. I just don’t know how else to weather this one. Is there a better way?
It's a complicated matter; personally, I believe that open sourcing it and democratizing its access to all people is the best way forward.
In simple words freeing the AI from corporate and state control.
Definitely is complicated. The problem that comes to mind with a regulation-free version of that is the classic ancap problem in of resource centralization would be much worse, as these bots will be 100x more productive at 1/10th the price. There’s no reason to pay a premium for some person to run an open source model at their desk or at home when the person never does anything. Cut them out and run it in-house autonomously.
If they get to the point that you don’t really need to hire real people, then companies who don’t will easily outcompete ones that do in providing all the world’s thought-services.
