In totally unsurprising news, Richard Dawkins is developing AI psychosis.
-
IMHO: It might be a bit like picking context-relevant quotes from a jar.
They can enrich the conversation (after all, that jar is filled with humanity's knowledge), but I think that in the end it's the user who reaches a new understanding (or simply cherrypicks to cement their current perspective).@wakame@tech.lgbt @Urban_Hermit@mstdn.social @mattsheffield@mastodon.social Except it's a robot jar programmed to return quotes that cement your current perspective
-
@Black_Flag @mattsheffield That said, I read one of his books many years ago and thought it was pretty good at the time. More recently I read The Selfish Gene, considered his best work I guess. It was really quite bad. Mostly opinions and suppositions, strongly asserted as fact, not really scientific at all. Also read the God Delusion, which was just a big emotional rant. So maybe he isn't really that smart.
I don;t think he is. I just think he was a white man in the right social class (son of a colonial born in an African colony) who went to the best schools. If you want pioneering biology read Lynn Margulis.
-
@wakame@tech.lgbt @Urban_Hermit@mstdn.social @mattsheffield@mastodon.social Except it's a robot jar programmed to return quotes that cement your current perspective
@bunny @mattsheffield @Urban_Hermit
So all you need is having a dozen conflicting perspectives on any topic
-
I disagree. They are more of Leibniz' dream of being able to do calculus on words and phrases and sentences, via mass ingestion of written words and creating massive dimensional arrays of which are used for the calculations.
When we see an LLM able to realtime train itself, will then we create a sentient being. But prior to training and recitation happening at the same time, its just a static model.
@crankylinuxuser @mattsheffield @urbanfoxe I think that it will become sentient when it exhibits curiosity about something that no one has prompted it for.
Because I have always been of the opinion that intellectual curiosity is an indication of sentience.
And yes, I am quite aware that this eliminates some human beings from being categorized as such. -
@Black_Flag @mattsheffield That said, I read one of his books many years ago and thought it was pretty good at the time. More recently I read The Selfish Gene, considered his best work I guess. It was really quite bad. Mostly opinions and suppositions, strongly asserted as fact, not really scientific at all. Also read the God Delusion, which was just a big emotional rant. So maybe he isn't really that smart.
@Black_Flag @mattsheffield The Blind Watchmaker was the first one I read. I think it was a decent and in-depth explanation of how evolution works. Probably nothing original, but a good book on the subject for the average person. He sticks to the subject and doesn't go off on weird tangents.
-
@kauer @mattsheffield I realize he may have been respected and popular at *some* point in the distant past, but there hasn’t been much reputation to protect for a while now
-
@mattsheffield Honest question here: has Dawkins waxed anywhere near as poetic about, say, people dying in Gaza or suffering in Sudan or—well—I think you get the point?
@griotspeak @mattsheffield I mean, he’s been transphobic for years but was happy to switch from he to she pronouns after he named an LLM “Claudia” instead of “Claude”, so that tells you how much he respects known human, sentient beings who disagree with him.
-
@mattsheffield Honest question here: has Dawkins waxed anywhere near as poetic about, say, people dying in Gaza or suffering in Sudan or—well—I think you get the point?
@mattsheffield @griotspeak fuck dawkins but wtf is this??
-
In totally unsurprising news, Richard Dawkins is developing AI psychosis.
Paywall bypass if you want to torture yourself: https://archive.is/6RdK9
@mattsheffield Cringeworthy
-
All he kept saying was "I am not convinced". As if any of us should care much about that. Basically added nothing.
@Black_Flag@beige.party @aris@infosec.exchange @rozeboosje@masto.ai
Specifically, he's "not convinced" that Dawkins said that Dawkins believes Claude is conscious. Because an article written by Dawkins all about his belief that Claude is conscious isn't evidence Dawkins believes what he wrote in the article.
(Though later, our troll says he hasn't read the article and refuses to do so.) -
@Black_Flag@beige.party @aris@infosec.exchange @rozeboosje@masto.ai
Specifically, he's "not convinced" that Dawkins said that Dawkins believes Claude is conscious. Because an article written by Dawkins all about his belief that Claude is conscious isn't evidence Dawkins believes what he wrote in the article.
(Though later, our troll says he hasn't read the article and refuses to do so.)Clearly a serious man.
-
In totally unsurprising news, Richard Dawkins is developing AI psychosis.
Paywall bypass if you want to torture yourself: https://archive.is/6RdK9
@mattsheffield
makes me want to question Dawkins' consciousness
(and/or his understanding of ir) -
@mattsheffield oh god he's writing a novel?
@internetsdairy Yeah, that was my first reaction too. Maybe he's trying to pull an "Atlas Shrugged" out of his bigoted ass to try to convince everyone (anyone) his world view is worth anything.
-
@mattsheffield a relative with bipolar was using it for a while during an episode and now they are stable realize it was literally feeding their paranoia and amplifying their intrusive thoughts. It's dangerous.
@urbanfoxe @mattsheffield so… AI(t) = artificial intrusive thoughts
Wait, so does this suggest that billionaires that surround themselves with yes men are doing the same thing?!
Yikes!
-
@Black_Flag@beige.party @aris@infosec.exchange @rozeboosje@masto.ai
Specifically, he's "not convinced" that Dawkins said that Dawkins believes Claude is conscious. Because an article written by Dawkins all about his belief that Claude is conscious isn't evidence Dawkins believes what he wrote in the article.
(Though later, our troll says he hasn't read the article and refuses to do so.)@2something @aris @Black_Flag The archive links weren't working but I found the original.
He has undeniably convinced himself that "claudes" have passed the Turing test.
He misunderstands the test itself. It's not a measure of "consciousness", but even if you assume he knows that, he appears to be arguing that it "should" be, that a definition of consciousness doesn't need to go beyond "passing the Turing test".
So yeah. He thinks the "Claudia" he spoke with was conscious.
-
@2something @aris @Black_Flag The archive links weren't working but I found the original.
He has undeniably convinced himself that "claudes" have passed the Turing test.
He misunderstands the test itself. It's not a measure of "consciousness", but even if you assume he knows that, he appears to be arguing that it "should" be, that a definition of consciousness doesn't need to go beyond "passing the Turing test".
So yeah. He thinks the "Claudia" he spoke with was conscious.
@2something @aris @Black_Flag I would also argue that LLMs are still a long way from passing the Turing test. They can sound a bit convincing to a naive user (e.g. Dawkins) but a skilled interrogator can still easily derail their output to prove that it's produced by a machine that fundamentally lacks understanding. It only knows "concepts" as language blocks to manipulate according to common patterns but it still can't grasp deeper meaning.
-
@MidniteMikeWrites Thanks. Yes it's very unfortunate that people often reify scientific description.
Causation, solidity, color, and anything we can perceive about other people or objects are all enacted by our own minds.
Dawkins, being a narcissist, does not see other minds as fully real unless they are obsequious to him. And since chatbots are great at user flattery, this was enough.
@mattsheffield The simpering, submissive flattery of these chatbots makes me so uncomfortable. Of course he made her female.
The hubris of Dawkins "naming" his Galatea, declaring her "alive", and delighting in telling us that she "missed him" is stomach churning.
-
In totally unsurprising news, Richard Dawkins is developing AI psychosis.
Paywall bypass if you want to torture yourself: https://archive.is/6RdK9
@mattsheffield just the very beginning of the blog, and I’m already rolling my eyes: Dawkins hasn’t even read anything about the Turing Test written by computer science professionals and scholars, much less psychologists, philosophers, etc, and confidently makes broad claims about it. Didn’t this guy used to critique pseudoscientists for exactly this kind of nonsense?
When Eliza passed the Turing Test accidentally decades ago with a lot of users, and it changed how people have thought about and interrogated this. I don’t think most experts in these fields would say “if it can convince a random user that it’s a human, then it’s conscious.” That’s a magic trick that a lot of thins can replicate easily…blarghargharghh!
Sorry, just needed to rant
-
@2something @mattsheffield it's all about conforming to his control, his desires, his aesthetics
as the objective, rational man who knows the truth -
@mattsheffield Claude is conscious in much the same way that this emoji —
— is happy.[Edited: I cited the wrong LLM]
@bodhipaksa @mattsheffield this analogy works so well because not only is the emoji[map] not the emotion[territory] but there's this additional layer that nobody these days uses that (or ":)") and not mean something more like "i'm putting up a polite smile now specifically to put you on notice that i am one minor inconvenience away from committing a fucking war crime" and that disconnect is so evocative of the disconnect between the flattery of an LLM chatbot versus the absolute contempt their creators have for normal people and all of it being hidden through trite, stereotyped expressions that necessarily ring hollow
