Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Google fires engineer who called its AI sentient

bigtechnology.substack.com

Post reply on HN
Print view View original

381–390 of 749 posts

Re: Google fires engineer who called its AI sentient

#381

Post by whittingtonaaon »

chmod775 wrote: post #323
whittingtonaaon wrote: post #202

Earlier quoted context omitted.

It is if you take “sentience” to mean “the ability to feel,” which is what my dictionary just told me. I think this category really is the most basic differentiating one. Higher level stuff like self awareness all depend on it. The most basic difference between a computer and a human (or even a dog…) is, in my opinion, the ability to feel.

>It is if you take “sentience” to mean “the ability to feel,” I don't like this definition much because "feel" is a fuzzy word. In this context it should be "feel" as in experience . I can build a machine that can sense heat and react to it, but I can't build one that can experience heat, or can I? You need to figure out what having the capability "to experience" means, and you'll be one step closer to defining senti…

I guess it is hard to define because it’s such a basic, essential thing. So does it matter that it’s hard to define? Even babies and puppy dogs experience pain and pleasure. They are feeling creatures. We don’t have any evidence that non-biological beings have pain, pleasure, fear, excitement… and so on.

Re: Google fires engineer who called its AI sentient

#382

Post by refulgentis »

SkyMarshal wrote: post #378
srvmshr wrote: post #4

LLM are definitely not sentient. As someone with a PhD in this domain, I attribute the 'magic' to large scale statistical knowledge assimilation by the models - and reproduction to prompts which closely match the inputs' sentence embedding. GPT-3 is known to fail in many circumstances which would otherwise be commonplace logic. (I remember seeing how addition of two small numbers yielded results - but larger numbers…

There was a recent post either here or on Twitter where someone took the questions Blake asked the AI about how it feels to be sentient, and replaced “sentient” with “cat” and had the same conversation with it. It’s clearly not self aware.

That's not possible because it's internal to Google: you're right about the post existing, though, I found it very misleading because it was replicating with GPT-3. Blake is certainly misguided, though.

Re: Google fires engineer who called its AI sentient

#383

Post by JoshCole »

mtlmtlmtlmtl wrote: post #205
robbrown451 wrote: post #166

Earlier quoted context omitted.

"Sentience I guess I would call awareness of self or something along those lines." Is a self driving car aware of itself? Surely within its logic it represents itself as a vehicle, while also being able to logically represent other vehicles on the road. Its logic accounts for there being similarities (they both obey the same laws of physics and such) and differences (its logic controls its own behavior but not that o…

I was a little sloppy in my formulation. A better way would be the ability to observe one's own thought processes and integrate them back into consciousness. Meta-cognition is a much better term, I think. It doesn't seem to me like cats for instance have this ability.

You should check out the YouTube of the cat that was trained to use buttons to speak. It sometimes says things that are surprising which it wasn't trained to say directly - like complaining that its owners music is bad. I realize that this isn't contradicting your thesis of language models not being conscious - but I just generally think people should check out things like that cat. So I'm not trying to disagree with you here. I just think the cat is cool.

I'm also a fan of ants. They have agriculture. They farm in their underground cities! They keep livestock too; insect livestock, but still! They give them medicines when they get sick, specially cultivated fungus for example. They are so cool.

Re: Google fires engineer who called its AI sentient

#384

Post by ge96 »

srvmshr wrote: post #4

LLM are definitely not sentient. As someone with a PhD in this domain, I attribute the 'magic' to large scale statistical knowledge assimilation by the models - and reproduction to prompts which closely match the inputs' sentence embedding. GPT-3 is known to fail in many circumstances which would otherwise be commonplace logic. (I remember seeing how addition of two small numbers yielded results - but larger numbers…

Corecursive listener by chance? (Eliza)

Re: Google fires engineer who called its AI sentient

#386

Post by Aeolun »

simonh wrote: post #265
Aeolun wrote: post #215

Earlier quoted context omitted.

Then is torture, rape and murder wrong because the victim is sentient, or because the perpetrator is?

There are arguments to be made for both. Some crimes, even if virtual, can stain or corrupt the perpetrator in ways inimical to society. There are plenty of examples of people who fantasised or role played abhorrent behaviour and went on to perpetrate it in real life, so there is a real danger. For example we tolerate computer games with virtual killing, but don’t tolerate virtual rape games. Even with virtual killin…

> Should we tolerate nazi death camp torturer simulation games?

This immediately brought the “Farming Simulator” imagery to mind. I can totally see how they’d make a nazi death camp simulator seem soul crushingly boring.

Re: Google fires engineer who called its AI sentient

#387

Post by Aeolun »

haswell wrote: post #300
simonh wrote: post #265

Earlier quoted context omitted.

There are arguments to be made for both. Some crimes, even if virtual, can stain or corrupt the perpetrator in ways inimical to society. There are plenty of examples of people who fantasised or role played abhorrent behaviour and went on to perpetrate it in real life, so there is a real danger. For example we tolerate computer games with virtual killing, but don’t tolerate virtual rape games. Even with virtual killin…

I'm trying to understand the connection between the original question and video games. This seems like an orthogonal set of considerations.

I think it has to do with the part where the perpetrator is a concious being. Clearly the enemies in the games aren’t concious, but does it still stain the human playing the game?

It was an angle I didn’t consider at all, so it was actually quite interesting.

Re: Google fires engineer who called its AI sentient

#389

Post by visarga »

dekhn wrote: post #13
bowmessage wrote: post #7

Earlier quoted context omitted.

What's the difference, really?

nobody can say with any epistemic certainty, but many of us who had worked in the field of biological ML for some time do not see language models like this as anything but statistical generators. There is no... agency... so far as we can tell (although I also can't give any truly scientific argument for the existence of true agency in humans).

If you want agency you have to put the AI into an environment and give it the means to perceive and act on the environment. The agent needs to have a goal and reward signals to learn from. Humans come with all these - the environment, the body and the rewards - based on evolution. But AIs can too - https://wenlong.page/language-planner/

Re: Google fires engineer who called its AI sentient

#390

Post by bee_rider »

kareemsabri wrote: post #24
klodolph wrote: post #17

Earlier quoted context omitted.

Smart CS students were fooled by ELIZA back in the day.

Smart CS students in the 1960s and professional AI engineers with Ph.D level education (as I believe he has) in 2022 are a world apart.

While we have become over-credentialed nowadays, I wouldn't say a smart CS student from the 1960s is worlds above a current day person with a PhD level education. We're at least in the same ballpark as the bright olden days bachelors degree holders.
Post reply on HN