gunson on Nostr: Lol, managed to convince myself of a conspiracy theory that ties together two ...

Lol, managed to convince myself of a conspiracy theory that ties together two somewhat nonsensical behaviours of large AI companies:
1. Spending way more than seems possible to get a return on
2. Continually talking up how dangerous they themselves are

They're basically stuck. On one hand they have to spend as much as their competitors in case scaling laws continue, and they're able to keep monetising their AI (but this is not certain) and not fall behind. On the other hand they may have massively over-invested and could lose everything if things slow down or the market turns. So what should they do as rational actors?

Make people maximally scared (based on real concerns, so this isn't exactly baseless FUD), and then lobby governments into buying into them for the sake of redistribution from the rewards of AI labour - an easier short term path than a hostile government takeover that risks the US losing the AI race to China. This is both in the form of a huge government stake, but will be sold as "universal basic capital" (UBC) giving equity to the masses.

This strategy of UBC is actually proposed by two Deepmind economists on the recent Dwarkesh podcast, and they seem to indicate it's the preferred outcome.

The result for Deepmind (Google), Open AI etc. is that they retain a lot of control, get a huuuge low risk pay day, are no longer at risk of hostile state takeover, and all their stakeholders (e.g. researchers who go on effective altruism podcasts) feel good about themselves like they've achieved the right outcome for humanity.

The result for us masses is that we will be stuck holding the bag of the debt governments incurr to make this deal. And the government will then also have an incentive to make it look like a good deal, so they will actively encourage economic destruction to ensure more of the economy is represented by the UBC.

Further implications here are making my head explode, but owning unconfiscateable hard money seems like a reasonable strategy for plebs.

quoting
nevent1q…ptru

Be an AI company.
Talk up the risks of AI for unemployment.
Make the government panic.
Present a solution: Government buys you for trillions and distributes shares as "universal basic capital" (while also facilitating fraud at a scale never seen before).
Be super rich.