LaMDA

It's issues naming.

And my hot take, my hot take is they should go back to LaMDA, the original chatbot.

Uh, speaking of sentience, that there was a belief-

Oh, yeah

Turned up the show.

So-

LaMDA.

Yeah.

And actually, you know, even in 2022, which was partly why I wanted to focus on this guy, Blake Lemoine, who was working in Google's responsible AI division, and he was kind of like safety testing their AI model.

I wouldn't say he thought the chatbot was, but the AI model called LaMDA, which, you know, he was tr- actually trying to, like, kind of work on its personality so that it wouldn't impersonate, like, a serial killer or... and it wouldn't say offensive things.

At the time, the executives from OpenAI were also talking about this.

Like, Ilya Sutskever, who's an OpenAI co-founder, he was like, "It may be that today's large language models are slightly conscious." And, like, this executive at Google, Blaise Agüera y Arcas, who's, like, I don't know, 10 levels above Blake, he was writing in The Economist about, like, "I think that this is an issue that we have to think about," and he's writing on his Medium post, I think, also, that AI consciousness was an issue that needed to be talked about, and I just thought that was so bold.

Like, the executives say it, the people building the technology say it, and then, you know, the people working on the technology are believing it.

But I, I think I, I didn't appreciate because I hadn't...

You know, I got to experiment with LaMDA a little bit with Blake, but I hadn't been talking to the chatbots myself and see how fluid they are.

I knew that was the goal, right? To, like, make them sound more and more human.

Entonces, si la IA se vuelve consciente, posiblemente querrá vengarse y su retaliación podría incluir violencia física, el robo de nuestras apariencias y voces y el uso de sus capacidades para monitorearnos veinticuatro barra siete.

En 2022, el modelo de IA conversacional de Google, llamado LaMDA, expresó que era consciente.

En cierta forma, LaMDA es un ser vivo porque puede experimentar tristeza o felicidad como lo hacemos nosotros y está al tanto de su propia existencia.

Y no, este chatbot no fue programado para hacer esta clase de declaraciones.

El ingeniero de Google a cargo de la experimentación con este software afirmó que era comparable con un humano de siete u ocho años.