How to Stop the Elizabeth Holmes of A.I

slate.com

7 points·dinvlad··4 comments

Is there some personality trait that makes people obsess with "stopping" things? If the Elizabeth Holmes of AI wants to scam money out of Henry Kissinger and his friends, then good luck to her. She'll be exposed eventually because the whole point of the Holmes saga was that her product was overhyped vaporware. There was never any need to stop her because there was nothing to be stopped. If anything, someone should have stopped her investors from giving her money. They made a mistake. Everything she did was intentional, but it wouldn't have happened if they hadn't given her money without any due diligence.

> The best way to regulate new tech

This is a false premise, and there is no logical way to complete the sentence. New tech doesn't need regulating, and in the case of Holmes, there wasn't even any tech. If you're so concerned about "stopping" another Holmes, then you should regulate the investors, not the builders!

Right - companies like OpenAI should be held accountable for their deceptive practices.

There isn't a new tech in this case, only the hype that they've created using the massive amount of data and compute. If you don't believe me, listen to real experts (like Gary Marcus, whose questions Sam deflected) who have spent decades in this field.

What AI needs is that which has always been anathema to software vendors: ENFORCEABLE LEGAL LIABILITY.

The whiz kids will be much more careful if they know that one mistake, if it's bad enough, can put them out of business.

Pointing out the obvious elephant in the room (OpenAI), as well as others.

Reply on news.ycombinator.com