Trying to get this conversation started in libraries. It feels important given our long held commitment to privacy coupled with serving a broad public.

Anthropic is out with new nonsense about Claude (or the collection of models called Claude) deserving moral consideration and/or be able to feel distress, so my inbox is filling up with journos asking me to comment. I've also got a day job, so here is a brief set of public comments on this: >>

3. Okay, so what would convince you that a piece of software (often phrased here as "an AI", gah) is conscious/can feel pain/etc? Honestly, I'm really tired of this extremely silly question, but here is my answer: >>

So let me know if you want to come to a virtual talk next wk on “Reimagining Leadership.“ Time tbd. The focus is on cultivating critical, creative, and caring capacities in medicine, but I think it applies beyond that. I’m pretty sure we can do this without any leaders.

Perhaps you are unfamiliar with my work....

I’m in DC scheduled to give this talk today. But the organizers flagged slides in my deck that mention racism, structural gaslighting, New Jim Code and other themes that are off limits in the Land of the Free.

In other words, it is the job of scientists to devise experiments that can test their claims, to always be cautious abt what they are claiming, and to critically evaluate the claims of other scientists. By these well-established standards, no serious science is happening behind Anthropic's claims.

Anyone talking about new rules for being nice to the software, please read this ⬇️

In other words, it is the job of scientists to devise experiments that can test their claims, to always be cautious abt what they are claiming, and to critically evaluate the claims of other scientists. By these well-established standards, no serious science is happening behind Anthropic's claims.

Thus, any evaluation of LLMs that involves looking at either language produced by some software, or language input into the system for that matter, needs to be done with extreme care to avoid projecting our interpretation of the language onto the system. >>

Extraordinary claims require extraordinary evidence, and I am not the one making the extraordinary claim, so it is not my job to specify what the evidence should be. >>

3. Okay, so what would convince you that a piece of software (often phrased here as "an AI", gah) is conscious/can feel pain/etc? Honestly, I'm really tired of this extremely silly question, but here is my answer: >>

It's the job of the media to scrutinize such statements and decide whether there's a story here at all (maybe the story is Anthropic researchers have fallen into a fantasy-land and brought many policymakers with them), not to just platform them. >>

2. But Anthropic researchers say they see patterns that suggest that Claude can. The employees at Anthropic are telling themselves stories based on the words they input into the software and the words that come back out. >>

1. Can a language model feel pain? Clearly not. Large language models are computer software and they can no more suffer than any other human-created artifact can. >>

Anthropic is out with new nonsense about Claude (or the collection of models called Claude) deserving moral consideration and/or be able to feel distress, so my inbox is filling up with journos asking me to comment. I've also got a day job, so here is a brief set of public comments on this: >>