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

Google DeepMind CEO says some form of AGI possible in a few years

wsj.com

Post reply on HN
Print view View original

271–280 of 388 posts

Re: Google DeepMind CEO says some form of AGI possible in a few years

#271

Post by nologic01 »

dS0rrow wrote: post #243
nologic01 wrote: post #200

Earlier quoted context omitted.

Unfortunately this argument, while imho entirely valid and also frequently seen in discussion, is unable to stop the massive train that has been set in motion. There are various types of retorts: i) the brain is also doing gradient descent; ii) what the brain does does not matter (if you can fake intelligence you have intellegence) iii) the pace is now so fast that has not happened in decades will happen in the next…

If not convincing what argument for and against do you think is the strongest?

I don't think there is any strong argument for AGI (i.e., that it is somehow coming any time soon). In my explanatory framework all developments make sense as purely statistical algorithmic advances. The surprising and interesting applications involving images, language are not really changing that fundamental reality.

There is a case to be made that with sufficient ingenuity at some point people will expand the algorithmic toolkit into more flexible and powerful dimensions. It may integrate formal logic in some shape or form, or yet to be conceived mathematical constructs.

But the type of mental leap that I think would be required just to breakout of the statistical fitting straight-jacket cannot be invented to satisfy the timing of some market craze.

If you look at the broad outline of the development of mathematics there are many areas where we have hit a complexity wall and generations of talent have not advanced an iota.

Even if we condition on some future breakthrough, the next level of mathematical / algorithmic dexterity we might reach will follow its own intrinsic logic, which will probably be very interesting but may or may not have anything to do with human intelligence.

Re: Google DeepMind CEO says some form of AGI possible in a few years

#272

Post by agentultra »

jabradoodle wrote: post #131
agentultra wrote: post #67

It seems to me that software developers are all too eager to attribute properties such as “intelligence,” to LLM’s which lead to strange, reductive conclusions that this is all humans are: token matching algorithms. There is much to intelligence that those who’ve been studying it in areas such as biology and ecology still do not understand. The role of emotions and how they work have a strong influence on our cogniti…

I don't think it is necessarily important how the human brain works. An interesting metaphor here is that aeroplanes don't work in a similar way to birds, the end goal is flight, not having wings that flap and are covered in feathers.

Precisely. LLM’s are an engineered system. Artificial.

What I keep hearing from the AGI folks is that we’re on the verge of replacing humans. That these systems, “think,” on their own and will be superior to us in every way: dangerous even!

I highly doubt they will be a danger on their own. Bing isn’t going to decide one day that it thinks you’ve been a bit distant lately and doesn’t want to answer your query until you apologize. It will answer the query because it’s an algorithm run on a computer that is designed to answer queries.

The danger of AGI still comes from people and corporations that wield them.

Although it would be very convenient if future cases against Google could absolve them of responsibility because of a “rogue AI.”

Re: Google DeepMind CEO says some form of AGI possible in a few years

#273

Post by HybridCurve »

ActorNightly wrote: post #2

I wish I could bet real money on this cause I would bet hard against this statement. The issue is that nobody seems to agree what AGI even is.

I agree on this.

I think we might also assume that we will believe we have created it before we actually do, but then simultaneously deny that we have created it after we already have.

Re: Google DeepMind CEO says some form of AGI possible in a few years

#274

Post by hammyhavoc »

jabradoodle wrote: post #240
ChatGTP wrote: post #163

Earlier quoted context omitted.

You forgot to quote Max Tegmark? :) It was a good talk…

Thanks, forgot where the analogy came from and should have taken the time to source it.

First you call it a metaphor, now you call it an analogy. Which are you using it as?

An analogy highlights similarities, a metaphor is saying a bird is a plane.

Is a nuclear power station a coal-fired power station? Both produce electricity.

Is an EV an ICE? Both can get you from point A to point B.

A statistical model is not the same as a complex neural network, and its capabilities are not comparable.

I don't think pointing out similarities is very useful. Why not treat it like what it is? An LLM.

Re: Google DeepMind CEO says some form of AGI possible in a few years

#275

Post by hammyhavoc »

boplicity wrote: post #18
ActorNightly wrote: post #2

I wish I could bet real money on this cause I would bet hard against this statement. The issue is that nobody seems to agree what AGI even is.

It's boils down to a question of "personhood." It turns out that's a political question and not a scientific one. Another way to say this: when will someone allow an AGI to have free rein and independent choice and power? How much power will we allow it?

Personhood has nothing to do with a hypothetical AGI happening. It actually boils down to hypothetical architecture even leading to AGI in the first place. It's a technical question, not a political question.

First it would need to actually have independent thought to be granted anything.

Re: Google DeepMind CEO says some form of AGI possible in a few years

#276

Post by RugnirViking »

_acco wrote: post #266
PheonixPharts wrote: post #41

As someone who has worked in the field of AI/ML for quite awhile now, the problem with current AGI predictions is ML hasn't done anything new since the 80s (or arguably earlier). At the end of the day all ML is using gradient descent to do some sort of non-linear projection of the data on to a latent space, then doing some relatively simple math in this latent space to perform some task. Personally I think the limits…

Yes, that’s right. AGI ≠ a lot of AI. They are fundamentally different things. The first computer was designed in 1837, long before a computer was ever built. We know how fusion reactions work, now we’re tweaking the engineering to harness it in a reactor. We don’t know how human intelligence works. We don’t have designs or even a philosophy for AGI. Yet, the prevailing view is that our greatest invention will just s…

your hypothesis is that there is a qualitatively different property to agi than whatever we have now. Most people here are just saying "if chatgpt is smarter that would be agi".

Going by the differences between gpt3.5 and gpt4 is really interesting. It is better able to reason in basically any problem I throw at it. Personally I think that a hypothetical system that is able to generate a sufficiently good response for ANY text input is AGI.

There aren't really any "gotcha" cases with this technology that I'm aware of where it just can't ever respond appropriately. Most clear failings of existing systems involve ever more contrived logic puzzles, which each successive generation is able to solve, and eventually at some point the required logic puzzle will be so dense few humans can solve it.

This isn't a case a case of "studying for the test" of popular internet examples either. I encourage you to try and invent your own gotchas for earlier versions then try them on newer models. Change the wording and order of logic puzzles, or encase them within scenarios to ensure its not responding to the format of the prompt

There are absolubtely cases of people overhypting it, or it overfitting to training data (see the debacle about it passing whatever bar exam, university test etc.). But despite the hype there is an underlying level of intelligence that is building and I use it to solve problems pretty much every day. I think of it atm as like a 4 year old that has inexplicably read every book ever written

Re: Google DeepMind CEO says some form of AGI possible in a few years

#277

Post by hammyhavoc »

blagie wrote: post #17
dougmwne wrote: post #10

Earlier quoted context omitted.

Seems like a no lose bet no matter the odds. If you win, life goes on and you can collect a bit of money. If you lose, money becomes worthless and the surface of the earth gets transformed into a computing substrate.

I don't know. Having interacted with LLMs at different levels, they resemble a very sophisticated, alien intelligence trying to pretend to be human. It's like me pretending to be a dog; even if I were to emulate a dog perfectly, I wouldn't have the same emotions; I'd be pretending. We have no idea what emotions, motivations, behaviors, or goals AIs have, will have, or if they'll have something as of yet unconvinced t…

> We have no idea what emotions, motivations, behaviors, or goals AIs have, will have, or if they'll have something as of yet unconvinced that's not emotions or motivations, but just alien.

AIs don't have emotions, motivations or goals.

They don't pretend, because pretending implies intent, they don't have intent. They do what they're created to do.

Humans are already brainwashed consumers. Welcome to marketing/advertising and late-stage capitalism. The ability of human beings to do what you're describing is much more effective than that of AI at present, ergo the "danger" has been here for decades. Smoking? Junk food? Radium water? Fast fashion? Equestrian ivermectin? Shall I continue?

Re: Google DeepMind CEO says some form of AGI possible in a few years

#278

Post by dmd »

meindnoch wrote: post #170
somewhereoutth wrote: post #164

Earlier quoted context omitted.

You don't need non-linearities, an infinite series of sine functions is enough to model any function. For extraordinary claims ('intelligence'), the burden of proof is on those making the claim, not on others to prove the negative.

>You don't need non-linearities, an infinite series of sine functions is enough The sine function is nonlinear.

So is IEEE floating point https://www.youtube.com/watch?v=Ae9EKCyI1xU

Re: Google DeepMind CEO says some form of AGI possible in a few years

#279

Post by fancyfredbot »

somewhereoutth wrote: post #164
earthnail wrote: post #123

Earlier quoted context omitted.

I disagree with you. > Essentially all we've done is pushed the basic model proposed by linear regression to it's absolutely limits No, we haven't pushed linear regression to its limits. If it was only linear regression, it wouldn't work. Neural networks need a non-linearity to model complex things. The beauty is that given an infinite series of nonlinearities, one can model any mathematical function. In practice we…

You don't need non-linearities, an infinite series of sine functions is enough to model any function. For extraordinary claims ('intelligence'), the burden of proof is on those making the claim, not on others to prove the negative.

Sine is non-linear, but you are right about not needing non-linear functions, as recently proven by Tom VII in a groundbreaking SIGBOVIK paper:

http://tom7.org/grad/

(Edit: this whole thread feels like a setup for this punchline)

Re: Google DeepMind CEO says some form of AGI possible in a few years

#280

Post by vidarh »

kumarvvr wrote: post #230
vidarh wrote: post #202

Earlier quoted context omitted.

How do we know those two are different?

Understanding and assimilation can lead to generating relations between disjoint sets of tokens. For example, "squeeze a tube to cut water flow" and "put pressure on a deep wound to stop blood loss" can only be related, if not already in the training data, if there is understanding and intelligence. The ability to do that is intelligence. Otherwise, it's just a search and optimization problem.

Firstly, transformer architectures are not "just a search and optimisation problem". They do generalise. Whether that generalisation is sufficient to be structurally equivalent to what we consider intelligence is an open question, but getting them to demonstrate that they generalise is easy (eg. ChatGPT can do math, albeit badly, with numbers large enough that it is infeasible for it to just have occurred in its training set)

Secondly, this poses the problem of 1) finding examples like the one you gave that it can't understand (regarding your specific example, see [1]), 2) ruling out that there was something "too close" drawing the equivalence in the training data, 3) ruling out that the failure to draw the equivalence is something structural (it can't have real understanding) rather than qualitative (it has real understanding, but it just isn't smart enough to understand the specific given problem)

So I'm back to my original question of how we would know if these are structurally different things in the first place.

[1] vidarh: How does squeezing a tube to cut water flow give you a hint as to what to do about a deep wound?

ChatGPT (GPT4): Squeezing a tube to cut off water flow demonstrates the basic principle of applying pressure to restrict or stop the flow of a fluid. Similarly, applying pressure to a deep wound can help control bleeding, which is a critical first aid measure when dealing with serious injuries.

[followed by a long list of what to do if encountering a deep wound]

Post reply on HN