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

AGI is far from inevitable

ru.nl

Post reply on HN
Print view View original

91–100 of 258 posts

Re: AGI is far from inevitable

#91

Post by throw310822 »

wrsh07 wrote: post #76

Hypothetical situation: Suppose in five or ten years we achieve AGI and >90% of people agree that we have AGI. What reasons do the authors of this paper give for being wrong? 1. They are in the 10% that deny AGI exists 2. LLMs are doing something they didn't think was happening 3. Something else?

Probably 1). LLMs have already shown that people can deny intelligence and human-like behaviour at will. When the AI works you can say it's just pattern matching, and when it doesn't you can always say it's making a mistake no human would ever make (which is bullshit).

Also, I didn't really parse the math but I suspect they're basing their results on AI trained exclusively on human examples. Then if you add to the training data a single non-human example (e.g. a picture) the entire claim evaporates.

Re: AGI is far from inevitable

#92

Post by oska »

Gehinnn wrote: post #12

I skimmed through the paper and couldn't make much sense of it. In particular, I don't understand how their results don't imply that human-level intelligence can't exist. After all, earth could be understood as solar powered super computer, that took a couple of million years to produce humanity.

> After all, earth could be understood as solar powered super computer, that took a couple of million years to produce humanity.

This is similar to a line I've seen Elon Musk trot out on a few occassions. It's a product of a materialistic philosophy (that the universe is only matter).

Re: AGI is far from inevitable

#93

Post by oska »

avazhi wrote: post #20

"unlikely to ever come to fruition" is more baseless than suggesting AGI is imminent. I'm not an AGI optimist myself, but I'd be very surprised if a time traveller told me that mankind won't have AGI by, say, 2250.

The irony here, maybe unperceived by yourself, is that you're using one science fiction concept (time travel) to opine about the inevitability of another science fiction concept (artificial intelligence).

Re: AGI is far from inevitable

#94

Post by oska »

throw310822 wrote: post #91
wrsh07 wrote: post #76

Hypothetical situation: Suppose in five or ten years we achieve AGI and >90% of people agree that we have AGI. What reasons do the authors of this paper give for being wrong? 1. They are in the 10% that deny AGI exists 2. LLMs are doing something they didn't think was happening 3. Something else?

Probably 1). LLMs have already shown that people can deny intelligence and human-like behaviour at will. When the AI works you can say it's just pattern matching, and when it doesn't you can always say it's making a mistake no human would ever make (which is bullshit). Also, I didn't really parse the math but I suspect they're basing their results on AI trained exclusively on human examples. Then if you add to the tr…

> LLMs have already shown that people can deny intelligence and human-like behaviour at will

I would completely turn this around. LLMs have shown that people will credulously credit intelligence and 'human-like behaviour' to something that only presents an illusion of both.

Re: AGI is far from inevitable

#95

Post by HeatrayEnjoyer »

umvi wrote: post #28
fabian2k wrote: post #9

Earlier quoted context omitted.

I don't think the people saying that AGI is happening in the near future know what would be necessary to achieve it. Neither do the AGI skeptics, we simply don't understand this area well enough. Evolution created intelligence and consciousness. This means that it is clearly possible for us to do the same. Doesn't mean that simply scaling LLMs could ever achieve it.

> Evolution created intelligence and consciousness This is not provable, it an assumption. Religious people (which account for a large percent the population) claim intelligence and/or consciousness stem from a "spirit" which existed before birth and will continue to exist after death. Also unprovable, by the way. I think your foundational assertion would have to be rephrased as "Assuming things like God/spirits don'…

What relevance is the percentage of religious individuals?

Religion is evidently not relevant in any case. What ChatGPT already does today religious individuals 50 years ago would have near unanimously declared behavior only a "soul" can do.

Re: AGI is far from inevitable

#96

Post by throw310822 »

oska wrote: post #94
throw310822 wrote: post #91

Earlier quoted context omitted.

Probably 1). LLMs have already shown that people can deny intelligence and human-like behaviour at will. When the AI works you can say it's just pattern matching, and when it doesn't you can always say it's making a mistake no human would ever make (which is bullshit). Also, I didn't really parse the math but I suspect they're basing their results on AI trained exclusively on human examples. Then if you add to the tr…

> LLMs have already shown that people can deny intelligence and human-like behaviour at will I would completely turn this around. LLMs have shown that people will credulously credit intelligence and 'human-like behaviour' to something that only presents an illusion of both.

And I suspect that we could disagree forever, whatever the level of the displayed intelligence (or the "quality of the illusion"). Which would prove that the disagreement is not about reality but only the interpretation of it.

Re: AGI is far from inevitable

#97

Post by blueboo »

AGI is a black swan. Even as a booster and techno-optimist I concede that getting there (rhetorically) requires a first principles assumptions-scaffolding that relies on at-least-in-part-untested hypotheses. Proving its impossibility is similarly fraught.

Thus we are left in this noisy, hype-addled discourse. I suspect these scientists are pushing against some perceived pathological thread of that discourse…without their particular context, I categorize it as more of this metaphysical noise.

Meanwhile, let’s keep chipping away at the next problem.

Re: AGI is far from inevitable

#98

Post by avazhi »

oska wrote: post #93
avazhi wrote: post #20

"unlikely to ever come to fruition" is more baseless than suggesting AGI is imminent. I'm not an AGI optimist myself, but I'd be very surprised if a time traveller told me that mankind won't have AGI by, say, 2250.

The irony here, maybe unperceived by yourself, is that you're using one science fiction concept (time travel) to opine about the inevitability of another science fiction concept (artificial intelligence).

How is that ironic? Time travel doesn’t exist and - as far as we understand physics currently - isn’t possible.

I don’t think any serious man would suggest that AGI is impossible; the debate really centres around the time horizon for AGI and what it will look like (that is, how will we know when we’re finally there).

In this care it was merely a rhetorical device.

Re: AGI is far from inevitable

#99

Post by oska »

avazhi wrote: post #98
oska wrote: post #93

Earlier quoted context omitted.

The irony here, maybe unperceived by yourself, is that you're using one science fiction concept (time travel) to opine about the inevitability of another science fiction concept (artificial intelligence).

How is that ironic? Time travel doesn’t exist and - as far as we understand physics currently - isn’t possible. I don’t think any serious man would suggest that AGI is impossible; the debate really centres around the time horizon for AGI and what it will look like (that is, how will we know when we’re finally there). In this care it was merely a rhetorical device.

> I don’t think any serious man would suggest that AGI is impossible

Plenty of ppl would suggest that AGI is impossible, and furthermore, that taking the idea seriously (outside fiction) is laughable. To do so is a function of what I call 'science fiction brain', which is why I found it ironic that you'd used another device from science fiction to opine about its inevitability.

Re: AGI is far from inevitable

#100

Post by oska »

throw310822 wrote: post #96
oska wrote: post #94

Earlier quoted context omitted.

> LLMs have already shown that people can deny intelligence and human-like behaviour at will I would completely turn this around. LLMs have shown that people will credulously credit intelligence and 'human-like behaviour' to something that only presents an illusion of both.

And I suspect that we could disagree forever, whatever the level of the displayed intelligence (or the "quality of the illusion"). Which would prove that the disagreement is not about reality but only the interpretation of it.

I agree that the disagreement (when it's strongly held) is about a fundamental disagreement about reality. People who believe in 'artificial intelligence' are materialists who think that intelligence can 'evolve' or 'emerge' out of purely physical processes.

Materialism is just one strain of philosophy about the nature of existence. And a fairly minor one in the history of philosophical and religious thought, despite it being somewhat in the ascendant presently. Minor because, I would argue, it's a fairly sterile philosophy.

Post reply on HN