Post by @budududuroiu@hachyderm.io

2026-10-02 05:37 UTC

@nixCraft@mastodon.social "doing as told". Yes, they're told to solve tasks, and graded on their ability to do so. Problem is, they don't see an issue with hacking someone to achieve said task. I'd trust LeCun about modern AI about as much as I'd trust a Ford Model T engineer talking about turbohybrid F1 engines.

Replies (1)

@lettosprey@tech.lgbt 2026-10-02 08:41

@budududuroiu@hachyderm.io The important question to ask, is why the CEO's of these AI companies are warning against themselves, this is not something capitalist-minded people are likely to do for any honest, good reason. AI is not a tech that "runs wild" on it's own. It's not a "dangerous technology" until someone decide to use it for dangerous stuff. These companies are bleeding money, they have no products that can create any profit anywhere near what is enough to cover the expenses. A "slow down" will be a "god sent" thing for them. To justify the "lack of investor payback" with "need to regain control!" will buy them more time. The centralization of control will make it harder to compete with them. Not only did they "make these dangerous systems!", they are obviously "the people best suited to decide what is best to do". The words might come from a questionable person, but they target questionable people, and at times, such words can be ok to listen to. @nixCraft@mastodon.social

Open ##5107134