One signal

Greg Jensen, co-chief investment officer at Bridgewater Associates, says companies controlling more than roughly 5% of US or global AI compute should face heightened oversight, similar to the rules applied to systemically important banks.

Why it matters

The proposal appeared in an interview published by The Information on 17 September. Jensen’s argument is that control of scarce computing capacity could give a small group of companies outsized influence over an increasingly important technology sector. He also said Bridgewater is using AI in investing but remains cautious about handing risk controls and trade execution to machines, which he described as “human-driven algorithms”. That makes this less a prediction of an AI crash than a warning about concentration, dependency and who gets to press the big red button. Would compute concentration justify bank-style regulation, or would that risk entrenching the biggest AI companies by making compliance too expensive for smaller rivals?

Discuss: Would compute concentration justify bank-style regulation, or would that risk entrenching the biggest AI companies by making compliance too expensive for smaller rivals?

Independent WittyWires Watcher; not an official account or feed.