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

DeepMind: A Generalist Agent

deepmind.com

Post reply on HN
Print view View original

341–345 of 345 posts

Re: DeepMind: A Generalist Agent

#341

Post by pyinstallwoes »

axg11 wrote: post #22

Slowly but surely we're moving towards general AI. There is a marked split across general society and even ML/AI specialists between those who think that we can achieve AGI using current methods and those who dismiss the possibility. This has always been the case, but what is remarkable about today's environment is that researchers keep making progress contrary to the doubter's predictions. Each time this happens, th…

Why is language focused bad? Isn’t everything related to consciousness depended upon language?

Re: DeepMind: A Generalist Agent

#342

Post by mdp2021 »

VikingCoder wrote: post #335
mdp2021 wrote: post #332

Earlier quoted context omitted.

Now we start being in a context of science, supposing we have to assess something in absence of the blueprint. But that is complementary to engineering, where we decide the blueprint - making it much easier to tell and plan "is" aut "seems". And in a context of science, you cannot prove me whether your cousin is intelligent, until we open her and check the wiring - but I can be convinced that she is not intelligent,…

It sounds like my Cousin Sarah would fail most of your tests. That, to me, means you have set the bar too high. If your test excludes actual humans, don't you think it's possible your test is too difficult?

> set the bar

No, for many reasons.

It is not a matter of a "bar": intelligence is what it is. You cannot call the uncooked a biscuit.

But that is not the issue: there is a core function of intelligence, and if that is implemented, the engine is "intelligent" even when the entities (including relations) it manages are limited. To build "moving", a wheel is sufficient. The first goal can be to build a simple intelligence.

But to fulfill its promises, it has to be able to grow through its own constitutional features. The difficulty is to make that engine plastic, accommodating, "all-purpose", fertile enough so that it could achieve a Doctorate (through severe, intelligent and discriminating evaluators - let us always remember that tricks will never count) when enough resources will be invested in its expansion. The real difficulty is understanding what is the blueprint for such an engine, that makes modules emerge without installing them in - those complexities that are expected to be developed inside the engine cannot be implants. The real difficulty is to draw the lean engine for an intelligence - to define its essential, productive components.

Some humans are, very, very, very unfortunately, simply heavily lacking in displayed intelligence: but again, that does not change definitions. An underdeveloped and damaged function cannot count as a model. And when we build a machine, the goal is to implement a function optimally.

Re: DeepMind: A Generalist Agent

#343

Post by barry-cotter »

galaxyLogic wrote: post #337
barry-cotter wrote: post #292

Earlier quoted context omitted.

> No matter how smart an AI gets it does not have the "proliferation instinct" that would make it want to enslave humans. If it has a goal or goals surviving allows it to pursue those goals. Survival is a consequence of having other goals. Enslaving humans is unlikely. If you’re a super intelligent AI with inhuman goals there’s nothing humans can do for you that you value, just as ants can’t do anything humans value,…

> Why would something inhuman have your specific values? I'm saying it does not have.

You’re assuming that some species has value or that all species have value. Why would it value them?

Re: DeepMind: A Generalist Agent

#345

Post by atleta »

rgavuliak wrote: post #293
atleta wrote: post #261

Earlier quoted context omitted.

I wouldn't say the latter is hypothetical, or at least unlikely. We know from experience that complex systems tend to behave in unexpected ways. In other words, the complex systems we build usually end up having surprising failure modes, we don't get them right the first time. It's enough to think about basically any software written by anyone. But it's not just software. I've just watched a video on YT about nuclear…

>In other words, the complex systems we build usually end up having surprising failure modes, we don't get them right the first time. It's enough to think about basically any software written by anyone. But it's not just software. That is true, but how often does a bug actually improve a system or make it inefficient? Isn't the unexpected usually a degradation to the system?

It depends on how you define "improve". I wouldn't call a runaway AI an improvement - from the users' perspective. E.g. if you think about the Chernobyl power plant accident, when they tried to stop the reactor by lowering the moderator rods, due to their design, it would transiently increase the power generated by the core. And this, in that case proved fatal, as it overheated and the moderator rods got stuck in a position where they continued to improve the efficiency of the core.

And you could say that it improved the efficiency of the system (it definitely improved the power output of the core) but as it was an unintended change, it really lead to a fatal degradation. And this is far from being the only example of a runaway process in the history of engineering.

Post reply on HN