Big speculative money just seeking regulatory capture so they can pull up the ladder and regulate to death any competition from open source. Simple as. Don't believe any hype or fear pr0n. LLM inference is a really rad tool, like Google Search was when it first emerged.
China will eat their lunch and the bubble will pop. Elon is smart to rent the shovels and fund space buildouts using frothy AI speculation bucks in the short term until the bubble pops, but ultimately inference will become dirt cheap when it is eventually etched onto silicon.
Consider all technology is intentionally released. LLM version of AI will never be a runaway autonomous super-intelligence. It can only be used to augment normal human intelligence. Isn't it interesting how safe that version of AI is? Useful, but safe. And it seemed to emerge as-if on a schedule, right around the start of the decade?
That's because it was released, on a schedule. Probably the military's timetable. For reasons.