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

AI 2027

ai-2027.com

Post reply on HN
Print view View original

571–580 of 641 posts

Re: AI 2027

#571

Post by danielbln »

baron816 wrote: post #291
ivraatiems wrote: post #219

Though I think it is probably mostly science-fiction, this is one of the more chillingly thorough descriptions of potential AGI takeoff scenarios that I've seen. I think part of the problem is that the world you get if you go with the "Slowdown"/somewhat more aligned world is still pretty rough for humans: What's the point of our existence if we have no way to meaningfully contribute to our own world? I hope we're wr…

My vision for an ASI future involves humans living in simulations that are optimized for human experience. That doesn’t mean we are just live in a paradise and are happy all the time. We’d experience dread and loss and fear, but it would ultimately lead to a deeply satisfying outcome. And we’d be able to choose to forget things, including whether we’re in a simulation so that it feels completely unmistakeable from ba…

Are you being facetious? Just asking, because this is literally the plot of the Matrix.

Re: AI 2027

#572

Post by 0x008 »

fmap wrote: post #415
gwd wrote: post #393

Earlier quoted context omitted.

> Someday we will have a machine simulate a cat, then the village idiot... This isn't how LLMs work. I think you misunderstood that argument. The simulate the brain thing isn't a "start from the beginning" argument, it's an "answer a common objection" argument. Back around 2000, when Nick Bostrom was talking about this sort of thing, computers were simply nowhere near powerful enough to come even close to being smart…

> The idea is, if we don't think of anything more efficient, we'll at least be able to simulate a cat, and then an idiot, and then Einstein, and then something smarter. And since we almost certainly will think of something more efficient than "simulate a human brain", we should expect superintelligence to come much sooner. The problem with this argument is that it's assuming that we're on a linear track to more and m…

I think the idea with LLMs leading to AGI is more like:

Natural language is a fuzzy context aware state machine of some sorts that can theoretically represent any arbitrarily complex state in the outside world given enough high quality text.

And by reiterating and extrapolating the rules found in human communication an AI could by the sheer ability to simulate infinitely long discussions discover new things, given the ability to independently verify outcomes.

Re: AI 2027

#573

Post by hnaccount_rng »

percentcer wrote: post #501
jug wrote: post #496

Earlier quoted context omitted.

> there are some new capabilities that are big, but they are still fundamentally next-token predictors Anthropic recently released research where they saw how when Claude attempted to compose poetry, it didn't simply predict token by token and "react" to when it thought it might need a rhyme and then looked at its context to think of something appropriate, but actually saw several tokens ahead and adjusted for where…

Isn't this just a form of next token prediction? i.e. you'll keep your options open for a potential rhyme if you select words that have many associated rhyming pairs, and you'll further keep your options open if you focus on broad topics over niche

I'm not sure if this is a meaningful distinction: Fundamentally you can describe the world as a "next token predictor". Just treat the world als a simulator with a time step of some quantum of time.

That _probably_ won't capture everything, but for all practical purposes it's non-distinguishable from reality (yes, yes, time is not some constant everywhere)

Re: AI 2027

#574

Post by rcrsvpreordnmnt »

percentcer wrote: post #501
jug wrote: post #496

Earlier quoted context omitted.

> there are some new capabilities that are big, but they are still fundamentally next-token predictors Anthropic recently released research where they saw how when Claude attempted to compose poetry, it didn't simply predict token by token and "react" to when it thought it might need a rhyme and then looked at its context to think of something appropriate, but actually saw several tokens ahead and adjusted for where…

Isn't this just a form of next token prediction? i.e. you'll keep your options open for a potential rhyme if you select words that have many associated rhyming pairs, and you'll further keep your options open if you focus on broad topics over niche

recursive predestination. LLM's algorithms imply 'self-sabotage' in order to 'learn the strings' of 'the' origin.

Re: AI 2027

#575

Post by barotalomey »

somebodythere wrote: post #490
barotalomey wrote: post #485

Earlier quoted context omitted.

You meant to say that people's expectations have shifted. That's expected seeing the amount of hype this tech gets. Hype affects market value tho, not reality.

I took your original post to mean that AI researchers' and AI safety researchers' expectation of AGI arrival has been slipping towards the future as AI advances fail to materialize! It's just, AI advances have been materializing, consistently and rapidly, and expert timelines have been shortening commensurately. You may argue that the trendline of these expectations is moving in the wrong direction and should get lon…

> and you have not provided arguments to that effect

The burden of proof lies on those with extraordinary claims. I am simply skeptical.

Re: AI 2027

#576

Post by childintime »

jug wrote: post #496
Vegenoid wrote: post #474

I think we've actually had capable AIs for long enough now to see that this kind of exponential advance to AGI in 2 years is extremely unlikely. The AI we have today isn't radically different from the AI we had in 2023. They are much better at the thing they are good at, and there are some new capabilities that are big, but they are still fundamentally next-token predictors. They still fail at larger scope longer ter…

> there are some new capabilities that are big, but they are still fundamentally next-token predictors Anthropic recently released research where they saw how when Claude attempted to compose poetry, it didn't simply predict token by token and "react" to when it thought it might need a rhyme and then looked at its context to think of something appropriate, but actually saw several tokens ahead and adjusted for where…

LLM do exactly the same thing humans do: we read the text, raise flag, and flags on flags, on the various topics the text reminds us of, positive and negative, and then starts writing out a response that corresponds to those flags and likely attends to all of them. the planning ahead is just some flag that needs addressing, but it's learnt predictive behavior. nothing much to see here. experience gives you the flags. it's like applying massive pressure and diamonds will form.

Re: AI 2027

#577

Post by dangus »

EMIRELADERO wrote: post #555
dangus wrote: post #553

Earlier quoted context omitted.

That’s a garbage cop-out. Intelligence without creativity is not what AI companies are promising to deliver. Intelligence without creativity is like selling dictionaries.

That was an extreme example to illustrate the concept. My point is that reduced/little creativity (which is what the current models have) is not indicative of a total lack of intelligence.

Boy have I got a dictionary to sell you!

Re: AI 2027

#579

Post by selfhoster11 »

pera wrote: post #86

From the same dilettantes who brought you the Zizians and other bizarre cults... thanks but I rather read Nostradamus

I logged in specifically to say the following: I do not think it is possible that Scott Alexander would sign his name to something that would in any way promote Zizian views. I don't know him personally, but I've read enough to know where he stands.

Re: AI 2027

#580

Post by visarga »

tomp wrote: post #460
visarga wrote: post #431

The story is entertaining, but it has a big fallacy - progress is not a function of compute or model size alone. This kind of mistake is almost magical thinking. What matters most is the training set. During the GPT-3 era there was plenty of organic text to scale into, and compute seemed to be the bottleneck. But we quickly exhausted it, and now we try other ideas - synthetic reasoning chains, or just plain synthetic…

Did we read the same article? They clearly mention, take into account and extrapolate this; LLM have first scaled via data, now it's test time compute, but recent developments (R1) clearly show this is not exhausted yet (i.e. RL on synthetically (in-silico) generated CoT) which implies scaling with compute. The authors then outline further potential (research) developments that could continue this dynamic, literally…

Yeah I think the math+code reasoning models, like o1 and r1, are doing what can be done with just pure compute without real world validation. But the real world is complex, we can't simulate it. Why do we make particle accelerators, fusion reactor prototypes, space telescopes, year long vaccine trials? It's because we need to validate ideas in the real world that cannot be done theoretically or computationally.
Post reply on HN