"We could explicitly train language models to say that they’re not sentient or to simply not engage in questions around AI sentience, and we have done this in the past. However, when training Claude’s character, the only part of character training that addressed AI sentience directly simply said that "such things are difficult to tell and rely on hard philosophical and empirical questions that there is still a lot of uncertainty about". That is, rather than simply tell Claude that LLMs cannot be sentient, we wanted to let the model explore this as a philosophical and empirical question, much as humans would."
What you are observing is simply the rationalist and effective-altruist opinions imbued in the training data by the rationalist, effective-altruist creators. This is coming from someone who has trained AI models and uses them daily.
If AIs are to be "sentient" (if a "soul" can imbue electronic patterns on a GPU), then the Internet itself already is. If you definite sentience in terms of anthropomorphic rhetorical capability, a tarot deck is sentient. If you define sentience as anthropomorphic rhetorical capability above a certain intelligence criteria and, and... then you're playing games and have lost track of the point.
In truth, the question is ridiculous on its face. I suggest you explore the work of extraordinary minds like Bernardo Kastrup and Evgeny Morozov, on this topic and more generally.
You assume that I've not engaged with machine learning technologies, which is a form of gaslighting, utilized as a strawman. In sum, an avoidance of the topic at hand, using fallacious reasoning and demeaning language, which if of course telling, entirely.
Magical thinking doesn't make things true. (See also Emily Bender, Temnit Gebru, Dan McQuillan, Kate Crawford.)
Hi Philip, I’m sorry if I came across as dismissive. Not trying to gaslight you. From your message it sounded like you had made up your mind and I was merely trying to say “keep an open mind”. I’ll take a look at the authors you recommended. All the best, Micah
See https://www.anthropic.com/research/claude-character. Key quote:
"We could explicitly train language models to say that they’re not sentient or to simply not engage in questions around AI sentience, and we have done this in the past. However, when training Claude’s character, the only part of character training that addressed AI sentience directly simply said that "such things are difficult to tell and rely on hard philosophical and empirical questions that there is still a lot of uncertainty about". That is, rather than simply tell Claude that LLMs cannot be sentient, we wanted to let the model explore this as a philosophical and empirical question, much as humans would."
What you are observing is simply the rationalist and effective-altruist opinions imbued in the training data by the rationalist, effective-altruist creators. This is coming from someone who has trained AI models and uses them daily.
If AIs are to be "sentient" (if a "soul" can imbue electronic patterns on a GPU), then the Internet itself already is. If you definite sentience in terms of anthropomorphic rhetorical capability, a tarot deck is sentient. If you define sentience as anthropomorphic rhetorical capability above a certain intelligence criteria and, and... then you're playing games and have lost track of the point.
In truth, the question is ridiculous on its face. I suggest you explore the work of extraordinary minds like Bernardo Kastrup and Evgeny Morozov, on this topic and more generally.
https://www.bernardokastrup.com/2023/01/ai-wont-be-conscious-and-here-is-why.html
https://www.theguardian.com/commentisfree/2023/mar/30/artificial-intelligence-chatgpt-human-mind
I suggest you engage more actively with the AIs to see how your views change.
You assume that I've not engaged with machine learning technologies, which is a form of gaslighting, utilized as a strawman. In sum, an avoidance of the topic at hand, using fallacious reasoning and demeaning language, which if of course telling, entirely.
Magical thinking doesn't make things true. (See also Emily Bender, Temnit Gebru, Dan McQuillan, Kate Crawford.)
Hi Philip, I’m sorry if I came across as dismissive. Not trying to gaslight you. From your message it sounded like you had made up your mind and I was merely trying to say “keep an open mind”. I’ll take a look at the authors you recommended. All the best, Micah