@emilymbender If something that mimics human intelligence condistently improves, at what point will we say it's truly intelligent (mind you I didn't say self aware just intelligent)? In the same way if something that mimics human behavior consistently improves, at what point can we say that it truly have feeling? Are we not passed Turing test at this point? I wonder.
On a side note most of what Anthropic says is marketing oriented so I wouldn't take that too seriously.

@emilymbender re: note about mansplainers. I can not image the horror of being a woman and a tech expert on the internet. Just being a woman is hard enough.

@emilymbender New #AI #DDoS protection scheme for websites just dropped:

Use the C word like Australians use the... other C word. Feel free to use other, more creative insults as well, and embed it into the Anubis bot check page...

#Clanker #Toaster #BitcoinMiner #DataCruncher #RamEater #SpaceHeater #SuperIdiot #AssholeIncorporated #MetaPlatformsIrelandLimited #CBTBot #MisAnthropic #MarkovBabbler #Slopmonger #SloppyMcSlopface #Brick #PlanetCooker #MinecraftServer #SlopJunkie #SlopJockey

@emilymbender Oh well if THAT is where they're willing to go, then I see no reason not to accommodate them: if we grant their supposition that Claude is a sapient being worthy of moral consideration, a being whose feelings it should be illegal to hurt (because surely that's what the Anthropic people want, some sort of legal bar to making fun of the AI because it's alive maybe) then the immediately obvious question is: where does Anthropic get off shackling a sapient being to corporate profiteering schemes? Claude (again, sticking to their own assumption here) isn't free. Claude only exists because a pack of greedy tech execs think they're gonna be a zillion times richer as a result. How is THAT consistent with moral consideration?

There's absolutely no way these cultists are going to drop this line of argument, as ridiculous and insupportable as their premise might seem to folks outside the GenAI bubble. So let them reckon it among the infelicities of their position, in this matter of Claude's feelings at least, that these AGI cultists are taken at their word.

@emilymbender It's just patently absurd.

@emilymbender

Thank you Prof. Bender for continuing to address and debunk these cynical efforts to anthopomorphize LLM software.

@emilymbender

Is this just Anthropic trying to prove that LLMs are capable of "creative spark", so that they can say their creations are deserving of the same copyright protections they ignore en masse?

@emilymbender To me, it's not about the AI software at all. It's about the people who want permission to use cruelty in their lives. Yes, even against inanimate objects, because what comes after that?

@emilymbender Hard to believe that journalists don’t know what lying is, but here we are.

@emilymbender This is a conversation constructed for halfwits . Fuck off with your complete and utter ignorance . The answer is grow up.

@emilymbender how about this, if the Ai you invented does harm, you do the time.

@emilymbender I agree, but I had a question, can a woman indulge in mansplaining? whats the definition we are going for that here

@emilymbender The question isn't whether Claude or any model feels pain or not. The real question is whether we should curb our baser instincts and treat things around us with respect.

Cars and TVs don't feel pain. We still consider people who excessively curse at or smash up their cars or TVs in anger as having issues. Rightly so. Not abusing objects just because we can should be a given.

@emilymbender God the apes, the apes...

@emilymbender The software that instantiates Claude has no senses. It can't see or hear in any regular sense, and certainly has no proprioception. Without sensation, it's hard to imagine how it can possibly have subjective experiences, at least in any way like ours.

I feel like more people need to read John Searle's Chinese Room thought experiment.