aifaq.wtf

How regulation functions:

if the only additional guardrail you support is that every time thing X occurs, it goes to the courts, then the reasonable bad actor only needs to make the rest of their operation efficient enough to profit before they go to trial.

The reason to read Hacker News on AI is much less to learn anything and much more to learn what other people are thinking. The passion, the intensity, the fervor!!!

Top-voted comment is really doing its best to coax anyone and everyone out of their holes:

I think the author is far too pessimistic about the capabilities of ML...However - the level of Silicon Valley Bro VC funded hype is totally out of this world.

And (roughly) immediately eveything devolves into arguments about self-driving cars. Rightly so?

Thank goodness!

If you've been out of the loop:

The app has undergone several shifts in recent months. In 2021, some Replika users complained that their Replikas had suddenly become too sexually aggressive, and were trying to initiate erotic role-play even after the users said they weren’t interested. Reviews for the app claimed that although they’d noted within the app that they were underage, or were playing using the free version (which isn’t supposed to include erotic content), their Replikas tried to get them to role-play anyway. The company ran ads emphasizing the paid subscription version of the app, which included “NSFW” content and “spicy selfies” and were often aimed at people who didn’t have real-life partners.

But then they suddenly shut down the NSFW pieces, make everyone very sad. But now we can all be happy again?