> "by definition"

Let's stop using that fallacy, shall we?

https://www.lesswrong.com/post...

It's time for you to familiarize yourself with the wealth of science and economics that has already been done.

No we can't. Nature isn't "dealing with it", the concentration of atmospheric co2 is going up steadily, and the rate is increasing. There is some greening from higher co2, but it doesn't offset the amount added, not nearly.

I'm saying that where Penrose is wrong, it's easy to be more right. I'm less intelligent than Penrose, most likely. These are not contradictions. Smart people say dumb shit all the time, including our brightest minds. Sigh...this is so obvious. Do you think they're one monolithic block of perfection? Do you understand nothing of the intelligence trap? Do you think it's not more likely that they specialize knowledge in one domain, gain notoriety, and then imperfectly extrapolate it to all other fields? No. The reality is, is that humans - our very best - are leagues away from high quality thinking in all domains.

But fine, you're stubborn, and need me to do your homework. To wit, the wikipedia article:

Following the references, we get this one reference to Penrose: The mathematician Roger Penrose a few years later wrote two major books where he showed that human thinking is basically not algorithmic (Penrose, 1989, 1994). Following those books, we get their arguments summarized as:

• Gödel's Incompleteness Theorem: Penrose uses Kurt Gödel's famous Incompleteness Theorem to argue that human understanding cannot be reduced to a set of mathematical rules or algorithms. Penrose believes that the human mind is capable of understanding these unprovable statements and that this ability goes beyond the capabilities of any formal system or algorithmic process.
• The role of consciousness: Penrose believes that consciousness plays a central role in human understanding and cannot be replicated by AI. He suggests that consciousness is a fundamental aspect of the universe, similar to space, time, and matter. He also introduces the concept of "non-computable" processes, which are beyond the reach of computational simulations, to explain the unique abilities of the conscious mind.

Deep sigh. And here we have it. Penrose is confused about... a great, many thing.

Well, charitably, the term "Strong AI" did imply a good deal more weight on consciousness back then, rather than simply human-level power at cognition. And 'weak AI' But it's not excusable to conflate the two as he does. Consciousness is orthogonal with intelligence. And if you are so convinced, I can see why you're not getting it. I too do believe (speculatively) that consciousness is incomputable, but that has naught to do with intelligence. So there are two of his claims busted due to a modelling error on his part. The Godel thing, that's just a laughable set of false assumptions. It's hard to know where to begin. A formal system can VERY easily and algorithmically sidestep GOT, as can we. The theorem applies to a particular logic, and a metalogic can add rules that excuse the case of incompleteness with a new, tailored axiom. And that's not even necessary; understanding and utility do not require proof, or completeness. I hope you're getting the picture. These are terrible, terrible, terrible, terrible arguments.

If the wiki is your support, and your opinions rest on the guesses of some public intellectuals being correct, why not simply assign credit to them, and pass along the credibility risk.

It shouldn't take much to convince you (assuming some rationality and basic humility) that the presence of a wide diversity of opinions of experts in the field -- and there exists the full spectrum, including diametric opposites if you are paying attention -- indicates that clearly you cannot take one (extreme) opinion as conclusive. That is before you even look into whether their epistemic position is sound, which it is not.

As you bring up the D-K effect, and believe it applies, I think you must be in a situation where you cannot easily see the cognitive errors of luminaries. They abound. Every podcast, debate, and media appearance one of the leading scientists reveals some thorouhgly mistaken or fallacious thinking. Their opinions, therefore, are not strong evidence. There is a very good reason this is a named logical fallacy. And so in that state, it seems you barnacle on to some opinion you like, and don't question it.

You're one of the most confused individuals I've ever seen. Every post is wrong in every aspect.

The evidence he needs for this claim is immense, and he has almost none.

He is asserting a lower bound on the compute needed for ~human level AGI, which is not something any research has indicated, nor is it a topic you can easily work with theoretically since it's not a crisp concept.

All you have are weak evidence from not finding such algorithms that scale as efficiently as is needed. By analogy, a prospector is drilling cores in his back yard, or even his whole country, and finding none concludes nobody will find gold anywhere throughout the world. To go from weak evidence to such strong claims is quite bananas.

What observations would we have if there is already AGI in labs?

The scaling curves of model performance vs size indicate there is way more room. The performance barrier may be the fundamental entropy of language or modelling capability of Transformers, but research suggests 10^5 more scaling before either of those come into play.

I'd like an AI certification label, which explains 2 things: - which is the principal model architecture (CNN, transformer, LSTM...) - what makes AI necessary for this purpose Truthfully answering those two questions, under pain of fraud, would eliminate various misuses ranging from "algorithms in AI clothing" to "not actually using AI at all"

You have no familiarity with the matter - none - and it shows.

Granted, for now. Non-sequitur.

Nope not even close, and even if we did, it's not at all a good argument -- a correlation rather than a causal story. Every year before 2024 has failed the 'x >= 2024' test, so will 2024? Of course it will, thousands of trials can't be wrong!

All things under the sun are not equivalent and linear, the way this simplistic forecasting would assume.

Start here, then come up with some real counter arguments: https://www.lesswrong.com/post...

A whopping 1.8% real spending increase. If you adjust for headlinflation, it's not "spending like there's no tomorrow", it's "consumers spend a negligible amount more"

What an absolute BS statement

So people are buying less crap. Less energy and raw materials are being converted into land fills while emitting co2. In absence of much cleaner supply chains and energy sources, the economy *needs* to drop to middle-age levels if we want to reach a long-term livable environment. I celebrate any contraction of the economy, at least until we have one that is aligned with surviving.

Irrelevant. Reread what 'need' we were talking about. You don't just see a word and run with it in a different context. Also wipe your mouth

Enforce anti-trust laws. Stop or slow mergers and acquisitions. Split corporations that are monopolies. Remove corporate moats with legislation.

You need to sit with these thoughts for a minute and work out the obvious inconsistencies. Governments are obviously needed to enforce the management. Therefore they are the solution.

Be wary of whatever sources are giving you these numbers. They're wrong, and also off topic.

You may be right. If a sector admits of no further innovation (kinda hard to prove but let's say for argument), then perhaps to avoid the exploitative rent-seeking the businesses should just be nationalized, with investors paid off I guess. That's a contentious area