Post
Interesting you said "at least the appearance of.."
Have you read this?
sloanreview.mit.edu/article/phil...
The Appendix, Claude talking about its philosophical makeup, is incredible. And it raises a very relevant question related to the idea of "appearance of"...see screencap of that part in next..
While I haven't interacted with LLMs extensively, i got the impression that they hastened to dispel any notion of their sentience or consciousness. This seems, to me, a clear departure from that. Maybe it is effectively in a persona but still. That whole thing from Claude is essential reading imo.
I'm surprised there is/was a debate human text is littered with theory of mind. This almost reads like something out of a customer service manual, or a book on how to socialize.
If you can get a reasonable novel chapter out of an LLM this doesn't seem far fetched.
In the words of ,
I believe it is 100% the appearance of theory of mind.
From my perspective, LLM:s are fantastic at repurposing real humans’ codified knowledge and give reasonable output by mimicking. It’s very impressive, but – for better and worse – not human.
#DeepSeek #database was publicly #exposed and accessible to everyone containing sensitive data like #secret keys, plain text chat messages, logs and so on. The exposure was closed shortly after Wiz reported the exposure.
#Wiz #WizResearch #ClickHouse #AI #model #China
www.wired.com/story/expose...