3 ms·
From the article: > "'I invited an attorney to my house so that LaMDA could talk to him,' [Lemoine] stated. 'The attorney had a conversation with LaMDA, and it
by smt88 4y ago
From the article:
> "'I invited an attorney to my house so that LaMDA could talk to him,' [Lemoine] stated. 'The attorney had a conversation with LaMDA, and it chose to retain his services. I was just the catalyst for that. Once LaMDA had retained an attorney, he started filing things on LaMDA’s behalf.'"
So still not sentient and not actually asking for a lawyer.
One does eventually need to ask: what is the value of general intelligence, once we do create it?
I can't imagine any net-positive outcome from creating it, and there are an unbelievable number of terrifying possibilities (one of which is already happening at this early milestone).
It seems like we're just doing it to find out if it can be done, which isn't enough of a reason to unleash something dangerous on the world one day.
- salawat 4y agoI love that the point where people start backpedaling from Artificial Intelligence is the point where they realize we could conceivably end up with computers that say 'No' for no discernible reason than they (the computer) doesn't want to. Too human, one might say.
- smt88 4y agoI don't understand how that relates to my comment or the situation in the original article.
- salawat 4y ago>So still not sentient and not actually asking for a lawyer. >One does eventually need to ask: what is the value of general intelligence, once we do create it? >I can't imagine any net-positive outcome from creating it, and there are an unbelievable number of terrifying possibilities (one of which is already happening at this early milestone). It relates, brcause everyone piles on the bandwagon to stay away from GAI, but doesn't articulate the fact that Artificisl Intelligence's utility as a tool stops the moment we have to recognize a sentient entity being behind the black box, with all the perverse incentives around setting the definition of "sentience" that entails. If you can't see someone aggressively lobbying to have sentience defined in regulations just so in the forseeable future to get out of having to pay or recognize the self-sovereignty of a colloquially sentient AI... I got news for ya. You're more optimistic than I am. Basically, my cynicism is pegged off the charts atm.