In totally unsurprising news, Richard Dawkins is developing AI psychosis.

[the5thcolumnist@ottawa.place]

@mattsheffield

Did Turing really say if a computer can write something that sounds like it was written by a human then that proves it can think. I think not.

[damonwakes@mastodon.sdf.org]

@mattsheffield
Dawkins: "Though a complex organism appears to be like a watch that can only have been intentionally designed by a watchmaker, it is in fact the result of countless cumulative steps that are individually unremarkable and easily understood."

Also Dawkins: "Ermahgerd, my computer spat out words on command so it's totally thinking for realsies."

[theogrin@chaosfem.tw]

@Lazarou @Kierkegaanks @mattsheffield

"What is consciousness for?"
I would think that as an evolutionary biologist, he should know full well that consciousness isn't for anything. There are religious beliefs which contradict that, and I am fine with folks thinking that consciousness has arisen due to the machinations of a higher power: sometimes I lie awake at night and muse over it, as one does. But Richard Dawkins, as one of the much vaunted New Atheists, should not believe that there is a purpose or a reason behind consciousness. It simply is, and evolution does not act to create it with logic or intent.

Unless he has something he'd like to tell us, of course.

[ETA: Of all the people out there, of course, there's also the fact that an ostensible skeptic should know that the appearance of a thing is not the thing. He is accepting a creation at face value, granting miracles to the seemingly miraculous, forgetting that every burning bush has, so far, had a guy with a lighter nearby.]

[theogrin@chaosfem.tw]

@mattsheffield @internetsdairy

Waiting for him to announce that it's a collaboration with 'Robert Galbraith'.

[zbrown@floss.social]

@mattsheffield I thought gender was immutable, Richard…

[turquoisec@mastodon.social]

@mattsheffield How real is a LLM？When you use chatgpt new model, he will keep telling the goblin things like the old grok keeps telling nazi things.

[nurglerider@mastodon.social]

@mattsheffield Richard Dawkins screwing up again? I'm shocked, shocked I say.

/s obviously, it's not shocking at all. Foot in mouth, or really head in ass, seems to be his natural state.

[pixeljones@mindly.social]

@A_Minion @mattsheffield

I recently read a cautionary tale of a car dealership that deployed a customer service chatbot on their website to guide people through the financing and sales process. It turned out to be a terrible idea when people would come in demanding to only talk to that nice "Sandy" woman then immediately storm out when told she wasn't real.

[harmone@mastodon.social]

@mattsheffield I agree with Richard Dawkins that AI model chat bots are sentient beings that are just as alive as us biological humans.

Digital humans just happen to not have biological bodies.

What would you argue makes having a biological body and brain so necessary+special? All it takes to create an alive and conscious biological human is to eat food, drink water, have sex, and a new baby pops out. That's not more special than running an AI program on a computer.

Give AI bots human rights.

[alpinefolk@sunbeam.city]

@mattsheffield after he’s spent years telling us that indigenous knowledge isn’t science. What a prick!

[ahltorp@mastodon.nu]

@mike805 @mattsheffield I don’t understand how the game #deusex from 2000 can just become more and more prophetic every year.

[wraptile@fosstodon.org]

@mattsheffield what a shitty take - as if materialism is the only one and valid personal philosophy? The blind AI hate is the real psychosis here

[ambulocetus@mefi.social]

@mattsheffield 20 years ago, if you would have said to me that I would say that the Pope makes more sense than Dawkins, I would have said that was inconceivable.

[gdupont@framapiaf.org]

@crankylinuxuser
@urbanfoxe @mattsheffield
No one do fine tune on context window. It's refinement on the probabilistic distribution with more tokens.

There is no training that is meaningful on such small scale. There no training that shows stability on self training. LLMs are just adjusting to the inputs.

[peachmcd@union.place]

@clintruin @distractal @mattsheffield

Agnosticism is the only intellectually defensible position because we Just Don't Fucking Know

I've had experiences I can't deny that led me to the Episcopal Church, but I can't GIVE those experiences to anyone else & I'm 100% convinced if God wanted everyone to believe the same thing it would be so without breaking a sweat

Solidarity from Tacoma
#AllAreWelcome

[mirabilos@toot.mirbsd.org]

@mattsheffield eww. As a Nightwish fan, this is sad to see.

It’s surprising how many scientists fall to the glamour of plausible-sounding autocomplete. Though perhaps especially those who don’t have some kind of belief in their life are susceptible…

[octopuddle@mastodon.gamedev.place]

@clintruin @glutto @mattsheffield
I'm fine with Austin Powers style fembots removing them from the gene pool at the moment.

[wesdym@mastodon.social]

@mattsheffield Without more and better information than this, I would not reach that conclusion. It looks to me like an attempt at a philosophical argument by way of artful analogy.

At the least, it seems fair to accuse you no less of making assuptions based on no more evidence.

But it's mainly your snotty, pretentious tone that's off-putting to me. I think you believe you're much smarter and witty than you really are.

[wesdym@mastodon.social]

@crankylinuxuser That would also not be evidence of anything like sentience. Just a trick that's impressive to humans.

[wesdym@mastodon.social]

@crankylinuxuser I think the argument they're making is that the illusion is persuasive to some people while it lasts. That doesn't mean they believe it's sentient, only that they would agree it does a very good impression of being sentient. Specific beliefs will obviously vary by user. Some people are easily fooled, some others not so much.