Live data from Hacker News

How close is AI to human-level intelligence?

nature.com

81–90 of 96 posts

Re: How close is AI to human-level intelligence?

#81
post #66

Earlier quoted context omitted.

Not really - machine learning, whether SVMs or ANNs, was called just that until relatively recently when the popular press started to first call ANNs AI, then LLMs. At first there was pushback from ML researchers, but particularly with LLMs they are now embracing it since investors want to invest in "AI". LLMs are really just fancy (deep) pattern recognizers/predictors, conceptually not so different than rule-based e…

> machine learning, whether SVMs or ANNs, was called just that until relatively recently when the popular press started to first call ANNs AI, That is absolutely not the case. These things have been in the field of AI for decades. Frighteningly it's nearing two decades since I started my degree in AI and it wasn't a new reference then.

I remember taking Andrew Ng's Coursera ML course (incl. neural nets and SVMs) when it came out in 2011, and nobody, including him, was calling it AI at that time. I think it was sometime after neural nets really took off after ImageNet 2012 that the press started to call everything AI.

Re: How close is AI to human-level intelligence?

#82
post #77

Earlier quoted context omitted.

> I don't think it'd be fully correct to say that knowledge is only encoded by relations between words. The input/output of the model is tokens of text, but internally it'll be converted into high-dimensional semantic vector spaces of concepts. All right, how about this: LLMs do have actual knowledge - the knowledge that was encoded in the words in the training data. That's not how they store the data internally, but…

> LLMs do have actual knowledge - the knowledge that was encoded in the words in the training data. That's not how they store the data internally, but the actual knowledge comes from there. For non-multimodal models, and minus ephemeral context and what's encoded by the architecture (like the translational invariance of CNNs), I'd agree to that. > And I wasn't saying that that's enough. I was saying that the LLM advo…

No, I would put text, images, video, and audio as one kind of "stuff" - NN training stuff. I would put knowledge graphs and rules for reasoning engines as another kind of stuff. If you use "modes" for text and images and so on, then I want something different from just "multimodal". I want left-brain vs right-brain, or slow vs fast, or something on that order. I want a different kind - not just fancier and larger LLMs. I want an LLM coupled to an inference engine with the Cyc encyclopedia available to it... or something in that direction. Maybe further than that.

Just LLMs aren't enough, and they aren't going to be enough.

You use words like "reasoning", but LLMs do not reason in the same way that an inference engine does. They can, at best, simulate it badly. I think we need more - not more of what we've got, but more of a different kind.

Re: How close is AI to human-level intelligence?

#83

Earlier quoted context omitted.

But it's not reasoning , it's just wordplay, just a gargantuanly complex level of auto-generated ELIZA.

It definitely feels like reasoning. Problems get solved. They may be simple problems, but it's still far beyond what a calculator can do. Does it really matter if it's "just wordplay"? I'm not convinced humans are any different, beyond the sheer scale. I certainly don't believe we have a 'reasoning module'.

You raise good points, I agree it feels like it is reasoning at times.

Though the brain, with our current understanding of it, is by far more of a black box to us than any LLM.

> I certainly don't believe we have a 'reasoning module'.

Let’s also point out that human brains probably don’t have any vector databases in them either.

It seems to me like our brains must work very differently - just look at how much energy an LLM consumes compared to our brains consuming around 12 Watts.

Re: How close is AI to human-level intelligence?

#84
post #77

Earlier quoted context omitted.

> LLMs do have actual knowledge - the knowledge that was encoded in the words in the training data. That's not how they store the data internally, but the actual knowledge comes from there. For non-multimodal models, and minus ephemeral context and what's encoded by the architecture (like the translational invariance of CNNs), I'd agree to that. > And I wasn't saying that that's enough. I was saying that the LLM advo…

No, I would put text, images, video, and audio as one kind of "stuff" - NN training stuff. I would put knowledge graphs and rules for reasoning engines as another kind of stuff. If you use "modes" for text and images and so on, then I want something different from just "multimodal". I want left-brain vs right-brain, or slow vs fast, or something on that order. I want a different kind - not just fancier and larger LLM…

> I want something different from just "multimodal". I want left-brain vs right-brain, or slow vs fast, or something on that order. I want a different kind - not just fancier and larger LLMs. I want an LLM coupled to an inference engine with the Cyc encyclopedia available to it...

So if I'm understanding, your objection isn't about the modalities that the model can work with (text, video, diagrams, ...), but about the kinds of processing it can do?

Many modern LLMs support tool calling (e.g: to look up entities in Google's knowledge graph, or evaluate code), mixture-of-experts architecture (specialized subnetworks that are enabled/disabled as needed per-query), and chain-of-thought inference (for questions requiring more complex reasoning). Would you consider those to be steps in the right direction?

> You use words like "reasoning", but LLMs do not reason in the same way that an inference engine does

If you view reasoning as something inference engines can do, then I don't think we disagree too much. Remaining difference may just be about error rate - I'm personally fine saying something can reason (at least "to some extent") even if it's a little fuzzy and not 100.0% accurate formal logic (else animals would also be excluded).

Re: How close is AI to human-level intelligence?

#85
post #75

From the article: > “Bad things could happen because of either the misuse of AI or because we lose control of it,” says Yoshua Bengio, a deep-learning researcher at the University of Montreal, Canada. God, I hate phrases like this. We've already lost control of it. We don't have any control. AI will evolve in the rich medium of capitalism and be used by anyone due to its ease of use and even laws will be unable to re…

> We've already lost control of it. We don't have any control. I think it's important to be clear on what's being said here: There's a very big difference between "we, the people, do not have control over what Silicon Valley businesses and billionaires are doing with AI" and "there are AIs out there that no human has any control over". The former is current reality. The latter is sci-fi, and nothing yet has demonstra…

You are right and I do not think anyone making AI has much control over it either because too many people who make it are addicted to the process in a very similar way in which drug users are addicted to drugs. And also, the prisoner's dilemma solution is almost forcing a lot of people to control it.

So I really do believe that NO human has control over it, really.

Re: How close is AI to human-level intelligence?

#86
post #10

From the article: > “Bad things could happen because of either the misuse of AI or because we lose control of it,” says Yoshua Bengio, a deep-learning researcher at the University of Montreal, Canada. God, I hate phrases like this. We've already lost control of it. We don't have any control. AI will evolve in the rich medium of capitalism and be used by anyone due to its ease of use and even laws will be unable to re…

As a thought exercise, on one hand the technology's purpose is having no purpose at all, just doing what it was made to do. So I'd rather focus on the technology people, where some develop it for the sake of developing, and others want to extract most money possible out of it. Nothing surprising because capitalism, but now there's a real possibility that the AI owner(s) will extract all the money there is. And I have…

I disagree with your first statement. Technology has a purpose, and that is to evolve itself. The ancient greeks had this perspective, so did several philosophers, and I agree with them.

Re: How close is AI to human-level intelligence?

#87
post #70

Earlier quoted context omitted.

That sentence literally makes no sense, obviously coming from the spoiled mind of a coddled rich kid. The fact is that you chose to write what you wrote, for good or ill. I spent all day yesterday explaining the truth of our moral existence to y'all here, conversing with a very fine fellow who has a bit of knowledge. It brought me a joy that no one else on this site has ever felt. It was electric, sans drugs of any k…

Is choice a fact? The point of the quote is that we do not know. It SEEMS like we have free will but this could be an illusion. Where do our desires and motivations come from? I have always had a strong desire to build things, either physical objects or mental ones (like code). These desires push me to make choices in my life, like pursuing the career I have. But did I choose my motivations? I’m not sure. I don’t rem…

My friend, what you are saying is that you do not know. You have no idea what I know, or even what I can know.

Truth is all that exists in the universe. We are the information processors of this universe and "knowing thyself" is part of our design, but, because of our free will, we can choose to remain ignorant. And, by knowing ourselves (long process), we also learn other things about the universe. At some point, we can actually have access to the very deepest truths that can be known by human beings. (There do remain some topics that are unknowable, but we can never exhaust the knowable, so simply knowing that some specific unknowables exists shall have to suffice.)

Our motivations are a combination of our physical predilections and our cultural and personal upbringing. No, we're not going to have a memory of our every motivation, but we are capable of at least gaining an understanding of what they are currently.

If you wish for the truth as explained by the "Sufi Science of Soul Transformation", follow the comment dialogue I referenced above. It requires a brave curiosity and ability to integrate very unfamiliar concepts; it is really akin to how Eugene Parker's solar wind theory shocked the world of astronomy, but on a far more important topic.

To really "know" if fire is dangerously hot, you have to test it; no one else's experiment is enough to truly convince those of us who have any skepticism within us.

It is the same with the spiritual path, except that we contain a force inside us that tries its best to convince us that remaining ignorant to the possibility that self-evolution into total compassion is not only possible, but is the best way forward for each of us. The only way to escape that ignorance is to light the flame and feel its effects, and it is each our choice. Most people are simply content with what is familiar to them, to their heritage, to their cultures, to our intrinsic ignorant nature.

The lack of compassion in this world is why it has been so historically f_cked, and getting ever more so, in many ways.

Evil is borne of selfishness that refuses universal compassion. Those Nazi death camp guards weren't killing anyone, but they sure contributed to the evil. We are all choosing sides, even if by default via our cultural inertias. We must choose to begin transforming ourselves into universally compassionate human beings, or else we have sided with either the deliberately evil or those callous to the evil they cause, which is a kind of evil. Entering the Path of Love is the only way to see this fact clearly and know how we are all choosing sides, whether we know it or not. The default value of a .NET int variable is always 0; you must change it to a 1 to make it a 1. To be not evil (in some measure, however by default) means to enter the Path of Love deliberately, because our default state is willful ignorance, and the masses upon masses of ignorant people are ruining this beautiful garden.

I wish you well, my fellow scientist (I have been such a one since 1st grade). It's your choice, the same as the rest of us.

Here's a link to the beginning of our dialogue: https://news.ycombinator.com/item?id=42317164

"The Way goes in." --Rumi

Re: How close is AI to human-level intelligence?

#88
post #84

Earlier quoted context omitted.

No, I would put text, images, video, and audio as one kind of "stuff" - NN training stuff. I would put knowledge graphs and rules for reasoning engines as another kind of stuff. If you use "modes" for text and images and so on, then I want something different from just "multimodal". I want left-brain vs right-brain, or slow vs fast, or something on that order. I want a different kind - not just fancier and larger LLM…

> I want something different from just "multimodal". I want left-brain vs right-brain, or slow vs fast, or something on that order. I want a different kind - not just fancier and larger LLMs. I want an LLM coupled to an inference engine with the Cyc encyclopedia available to it... So if I'm understanding, your objection isn't about the modalities that the model can work with (text, video, diagrams, ...), but about th…

I view reasoning as something that LLMs do a kind of, or a subset of, and inference engines do a different kind or subset of. And there may be different kinds or subsets than just those two.

And just as inference engines, by themselves, were not enough to be really able to "reason", neither are LLMs, by themselves. (I think "AI" has historically been quite reductionist - they reduce thinking to only one kind of thinking, and then try to automate that. The result can sometimes be impressive, but always is less than what human thinking is.)

Tool calling or mixture-of-experts are in the direction that I'm thinking.

Re: How close is AI to human-level intelligence?

#89
post #70

Earlier quoted context omitted.

Is choice a fact? The point of the quote is that we do not know. It SEEMS like we have free will but this could be an illusion. Where do our desires and motivations come from? I have always had a strong desire to build things, either physical objects or mental ones (like code). These desires push me to make choices in my life, like pursuing the career I have. But did I choose my motivations? I’m not sure. I don’t rem…

My friend, what you are saying is that you do not know. You have no idea what I know, or even what I can know. Truth is all that exists in the universe. We are the information processors of this universe and "knowing thyself" is part of our design, but, because of our free will, we can choose to remain ignorant. And, by knowing ourselves (long process), we also learn other things about the universe. At some point, we…

What you present as a grand narrative of self-discovery, truth, and compassion is, in many ways, an idealistic interpretation of human existence that oversimplifies complex realities. While I can appreciate the depth of your convictions, the perspective you offer assumes a universal applicability of your framework—one that may not align with everyone's experience, philosophy, or epistemological approach.

To begin, claiming that "truth is all that exists in the universe" reduces the richness of existence to a monolithic pursuit. Truth, as a concept, is subjective and shaped by individual, cultural, and temporal contexts. What you call "truth" may resonate with you, but it risks dismissing alternative ways of understanding the universe, such as those rooted in skepticism, pragmatism, or even nihilism. These frameworks are equally valid, as they acknowledge the limitations of human cognition and the constructed nature of meaning.

The notion that self-knowledge leads to universal truths about the cosmos assumes a direct correlation between introspection and external understanding. While self-awareness is undoubtedly valuable, it does not guarantee access to the deepest truths of the universe. Human cognition is bounded by our biology, sensory limitations, and the constraints of language and culture. We are not omniscient processors; we are flawed, interpretive beings navigating a sea of uncertainty.

Your emphasis on "universal compassion" as the sole antidote to evil is admirable but simplistic. Human motivations are multifaceted, and what you describe as "evil" often emerges from systemic, historical, and material conditions rather than individual moral failings. Compassion, while transformative, cannot alone dismantle entrenched power structures or resolve the complex web of human suffering. Moreover, framing those who do not embrace your "Path of Love" as ignorant or complicit in evil undermines the diversity of human experience and the validity of alternative ethical frameworks.

Finally, your analogy comparing spiritual transformation to testing fire conflates subjective spiritual experiences with objective physical phenomena. The former is deeply personal and cannot be universally measured or validated. Not everyone will, or should, approach spirituality or self-evolution in the way you propose. Placing the burden of moral alignment on individuals rather than acknowledging the role of collective and systemic forces risks perpetuating a kind of spiritual elitism.

In summary, while your call for self-awareness, compassion, and transformation is compelling, it oversimplifies human complexity and diversity. We are not all on the same path, nor should we be. True respect for the plurality of human experience requires acknowledging that there are many ways to navigate existence—each as valid as your own.

Re: How close is AI to human-level intelligence?

#90
>models... AGI. ... unlikely to reach this milestone on their own.

I don't think there is a single milestone. Intelligence has many aspects - IQ test type ability, chess playing ability, emotional intelligence, ability to go down to the shops and buying something and so on.

AI is making gradual progress and is very good at some things like chess and very bad at others. There will probably be a gradual passing of different milestones on different dates. When they can replace a plumber including figuring out the problem and getting and fitting the parts might be a sign they can do most stuff.

There's probably a way to go there but there are a lot of resources being thrown at the problem just now.

Post reply on HN