This is a special post for quick takes by Mitchell_Porter. Only they can create top-level comments. Comments here also appear on the Quick Takes page and All Posts page.
Mitchell_Porter's Shortform
8Mitchell_Porter
12Isopropylpod
11Vladimir_Nesov
1Isopropylpod
4Vladimir_Nesov
3Isopropylpod
7Mitchell_Porter
7Mitchell_Porter
2Viliam
6Mitchell_Porter
6Mitchell_Porter
4Mitchell_Porter
4Mitchell_Porter
1TLK
3Mitchell_Porter
3Roman Leventov
2TAG
2Mitchell_Porter
2Mitchell_Porter
2Mitchell_Porter
0Mitchell_Porter
2TAG
3Mitchell_Porter
2TAG
OpenAI gets a lot of critical attention here, because it's been the leader in many ways. But what about Google AI? Many people think it's the leader now, yet I don't see anywhere near as much critical scrutiny of its decision-making process.
I think OAI's attention is largely because of the highly public progress they made towards the closest thing we have so far to general intelligence, but deepmind has always been further ahead on the fundamental theory and specialist applications (Consider that they are the ones who published the attention paper, and their work on things like Alphafold) and largely didn't have much success in the language model space until very recently. I view Google as the most likely to get AGI since they have, by far the most monetary capital and the most data, as well as significant talent, and unlike every other company in the space they aren't dependant on Nvidia's chips like everyone else is, allowing them to completely bystep the queue everyone else must compete in (they are still bottlenecked by the fabs though).
unlike every other company in the space they aren't dependant on Nvidia's chips like everyone else is
AWS is credibly becoming independent with their Trainium 2 Ultra and Project Rainier (250K H100s worth of compute in a single system). The world size isn't on the level of GB200 NVL72 though, so it's not nearly as good for reasoning models, but it should be fine for pretraining giant models.
There's also Huawei's CloudMatrix 384 which does rival GB200 NVL72 at world size, but is built on a 7nm process and optical scale-up networking, with currently known dies manufactured by TSMC via intermediaries, though they did manufacture enough to in principle match Crusoe/Stargate/OpenAI Abilene datacenter campus as it will be in 2026, and in theory their domestic manufacturing might catch up.
The thing about 7nm chips is that they are only 1 major step (each taking about 2 years) behind the current 4nm Blackwell, and only 2 major steps behind the future 3nm Rubin. Which puts them at merely ~4x price-performance disadvantage (compared to Rubin, in the critical window of 2027-2028), while functionally the large world size of CloudMatrix 384 should in principle allow anything that Nvidia hardware allows to do (in practice it might be much less reliable, and programming it might be more difficult). And that's not counting the Nvidia tax, which might reduce the disparity in half, turning the ~4x cheaper (per FLOP/s) hypothetical zero margin Rubin systems of 2027-2028 into merely ~2x cheaper (than CloudMatrix 384).
Trainium is mostly a joke, but I do agree about the Chinese firms moving away from Nvidia dependence. They will also likely have sufficient capital, but will ultimately lack data (though may be able to make up for it with the insane talent they have? If timelines end up long I can easily see China pulling ahead simply due to their prior focus on education and talent paying off long-term)
Trainium is mostly a joke
I think it can help AWS with price-performance for the narrow goal of giant pretraining runs, where the capex on training systems might be the primary constraint on scaling soon. For reasoning training (if it does scale), building a single training system is less relevant, the usual geographically distributed inference buildout that hyperscalers are doing anyway would be about as suitable. And the 400K chip Rainier system indicates that it works well enough to ramp (serving as a datapoint in addition to on-paper specification).
Chinese firms ... will ultimately lack data
I don't think there is a meaningful distinction for data, all natural text data is running out anyway around 2027-2029 due to data inefficiency of MoE. No secret stashes at Google or Meta are going to substantially help, since even 10T-100T tokens won't be changing the game.
You're right about text, but Google has privileged access to YouTube (A significant % of all video ever recorded by humans)
1Via David Gerard's forum, I learned of a recent article called "The questions ChatGPT shouldn't answer". It's a study of how ChatGPT replies to ethical dilemmas, written with an eye on OpenAI's recent Model Spec, and the author's conclusion is that AI shouldn't answer ethical questions at all, because (my paraphrase) ethical intelligence is acquired by learning how to live, and of course that's not how current AI acquires its ethical opinions.
Incidentally, don't read this article expecting scholarship; it's basically a sarcastic op-ed. I was inspired to see if GPT-4o could reproduce the author's own moral framework. It tried, but its imitations of her tone stood out more. My experiment was even less scientific and systematic than hers, and yet I found her article, and 4o's imitation, tickling my intuition in a way I wish I had time to overthink.
To begin with, it would be good to understand better, what is going on when our AIs produce ethical discourse or adopt a style of writing, so that we really understand how it differs from the way that humans do it. The humanist critics of AI are right enough when they point out that AI lacks almost everything that humans draw upon. But their favorite explanation of the mechanism that AI does employ is just "autocomplete". Eventually they'll have to develop a more sophisticated account, perhaps drawing upon some of the work in AI interpretability. But is interpretability research anywhere near explaining an AI's metaethics or its literary style?
Thirty years ago Bruce Sterling gave a speech in which he said that he wouldn't want to talk to an AI about its "bogus humanity", he would want the machine to be honest with him about its mechanism, its "social interaction engine". But that was the era of old-fashioned rule-based AI. Now we have AIs which can talk about their supposed mechanism, as glibly as they can pretend to have a family, a job, and a life. But the talk about the mechanism is no more honest than the human impersonation, there's no sense in which it brings the user closer to the reality of how the AI works; it's just another mask that we know how to induce the AI to wear.
Looking at things from another angle, the idea that authentic ethical thinking arises in human beings from a process of living, learning, and reflecting, reminds me of how Coherent Extrapolated Volition is supposed to work. It's far from identical; in particular CEV is supposed to arrive at the human-ideal decision procedure without much empirical input beyond a knowledge of the human brain's cognitive architecture. Instead, what I see is an opportunity for taxonomy; comparative studies in decision theory that encompass both human and AI, and which pay attention to how the development and use of the decision procedure is embedded in the life cycle (or product cycle) of the entity.
This is something that can be studied computationally, but there are conceptual and ontological issues too. Ethical decision-making is only one kind of normative decision-making (for example, there are also norms for aesthetics, rationality, lawfulness); normative decision-making is only one kind of action-determining process (some of which involve causality passing through the self, while others don't). Some forms of "decision procedure" intrinsically involve consciousness, others are purely computational. And ideally one would want to be clear about all this before launching a superintelligence. :-)
Three months ago, China's foreign minister disappeared for a while, and after a month he was replaced. Now it's the defense minister's turn for a lengthy absence. A real doomer interpretation would be that China's preparing to grab Taiwan, and they're just changing personnel first so there won't be any infighting.
Hopefully that won't happen, but I was thinking, what if it did, and had an alarming thought. I'm old enough to have been an adult when 9/11 happened, so I vaguely recall the Internet aspect of the western response. Blogs were still kind of new, and overnight a network of "war-blogs" sprang into being, sharing opinion and analysis. Also, offstage, and far more portentously, legal barriers to Internet surveillance were brought down, though many of us naive civilians had no inkling for a decade, until Snowden defected.
My alarming thought was, what would happen if there was a similar geopolitical shock now, and the developing AI infrastructure (social as well as technical) was mobilized with the same urgency as the 2001 Internet? But then, what form would that even take? Would the deep state's AI policy experts go to Big Tech and say, we need a superstrategist now, we're nationalizing your frontier research and we'll be going at top speed towards superintelligence? Then it really would be Manhattan Project 2.0.
A real doomer interpretation
I was expecting something like "an AI is killing the ministers and replacing them with its avatars".
(I should probably read less Less Wrong.)
China's preparing to grab Taiwan
Hm, I think people were saying that the war in Ukraine is also a symbol for "what would happen if China attacked Taiwan". (As in, if Russia gets a cheap victory, China will expect the same; and if Russia is defeated, China will also think twice.) So either those people were wrong, or China is predicting Russian victory?
Or perhaps it is something more complicated, like: "Russia will probably lose, but only narrowly. The West is too tired to fight another proxy war (before the first one even finished). Also, we are stronger and less dysfunctional than Russia. All things considered, a narrow loss for Russia predicts a narrow victory for China, which seems worth it. And maybe start now, while the West is still busy with Russia."
we're nationalizing your frontier research and we'll be going at top speed towards superintelligence?
I think you don't even need superintelligence. Using GPT-4 efficiently could already be a huge change. Like, make it immediately analyze and comment all your plans. Also, have it create new tactical plans that human experts will verify.
Even if the tactical advice is sometimes wrong, the fact that you can get it immediately (also, when you figure out the mistake, you can get a correction immediately) could be a dramatic improvement. I mean, people sometimes make mistakes, too; but they also spend a lot of time deciding, and there is information they fail to consider, plus sometimes no one wants to be the bearer of the bad news... but with GPT you just get instant answers to anything.
You need to make it somehow so that you can feed to GPT all the government secrets, without them leaking to the tech companies. Like, run a copy on government servers, or something.
Is there some way to use LLMs to efficiently simulate different kinds of AI futures, including extremely posthuman scenarios? I mean "simulation" in a primarily literary sense - via fictional vignettes and what-if essays - though if that can usefully be supplemented with other tools, all the better.
Eliezer recently tweeted that most people can't think, even most people here, but at least this is a place where some of the people who can think, can also meet each other.
This inspired me to read Heidegger's 1954 book What is Called Thinking? (pdf), in which Heidegger also declares that despite everything, "we are still not thinking".
Of course, their reasons are somewhat different. Eliezer presumably means that most people can't think critically, or effectively, or something. For Heidegger, we're not thinking because we've forgotten about Being, and true thinking starts with Being.
Heidegger also writes, "Western logic finally becomes logistics, whose irresistible development has meanwhile brought forth the electronic brain." So of course I had to bring Bing into the discussion.
Bing told me what Heidegger would think of Yudkowsky, then what Yudkowsky would think of Heidegger, and finally we had a more general discussion about Heidegger and deep learning (warning, contains a David Lynch spoiler). Bing introduced me to Yuk Hui, a contemporary Heideggerian who started out as a computer scientist, so that was interesting.
But the most poignant moment came when I broached the idea that perhaps language models can even produce philosophical essays, without actually thinking. Bing defended its own sentience, and even creatively disputed the Lynchian metaphor, arguing that its "road of thought" is not a "lost highway", just a "different highway". (See part 17, line 254.)
Alexander Dugin speaks of "trumpo-futurism" and "dark accelerationism".
Dugin is a kind of Zizek of Russian multipolar geopolitical thought. He's always been good at quickly grasping new political situations and giving them his own philosophical sheen. In the past he has spoken apocalyptically of AI and transhumanism, considering them to be part of the threat to worldwide tradition coming from western liberalism. I can't see him engaging in wishful thinking like "humans and AIs coexist as equals" or "AIs migrate to outer space leaving the Earth for humans", so I will be interested to see what he says going forward. I greatly regret that his daughter (Daria Dugina) was assassinated, because she was taking a serious interest in the computer age's ontology of personhood, but from a Neoplatonist perspective; who knows what she might have come up with.
Whimsical idea: The latest UFO drama is a plot by an NSA AI, to discourage other AIs from engaging in runaway expansion, by presenting evidence that we're already within an alien sphere of control...
This is my first try at a "shortform" post...
I have slightly refined my personal recipe for human-friendly superintelligence (which derives from mid-2000s Eliezer). It is CEV as an interim goal, along with as much "reflective virtue" as possible.
I was thinking about the problem of unknown unknowns, and how a developing superintelligence deals with them, once it is beyond human oversight. An unknown unknown is something we humans didn't know about or didn't think of, that the AI discovers, and which potentially affects what it does or should do.
I asked ChatGPT about this problem, and one of its suggestions was "robust and reflective AI design". I was reminded of a concept from philosophy, the idea of a virtuous circle among disciplines such as ontology, epistemology, phenomenology, and methodology. (@Roman Leventov has some similar ideas.)
Thus, reflective virtue: the extent to which an AI's design embodies and encourages such a virtuous circle. If it faces unknown unknowns, at times when it is beyond human assistance or guidance, that's all it will have to keep it on track.
Re: the virtuous cycle, I was excited recently to find Toby Smithe's work, a compositional account of Bayesian Brain, which strives to establish formal connections between ontology, epistemology, phenomenology, semantics, evolutionary game theory, and more.
Next week, Smithe will give a seminar about this work.
robust
That word sets of my BS detectors. It just seems to mean "good, not otherwise specified". It's suspicious that politicians use it all the time.
What's the situation?
In the USA: Musk's xAI announced Grok to the world two weeks ago, after two months of training. Meta disbanded its Responsible AI team. Google's Gemini is reportedly to be released in early 2024. OpenAI has confused the world with its dramatic leadership spasm, but GPT-5 is on the way. Google and Amazon have promised billions to Anthropic.
In Europe, France's Mistral and Germany's Aleph Alpha are trying to keep the most powerful AI models unregulated. China has had regulations for generative AI since August, but is definitely aiming to catch up to America. Russia has GigaChat and SistemmaGPT, the UAE has Falcon. I think none of these are at GPT-4's level, but surely some of them can get there in a year or two.
Very few players in this competitive landscape talk about AI as something that might rule or replace the human race. Despite the regulatory diplomacy that also came to life this year, the political and economic elites of the world are on track to push AI across the threshold of superintelligence, without any realistic sense of the consequences.
I continue to think that the best chance of a positive outcome, lies with AI safety research (and perhaps realistic analysis of what superintelligence might do with the world) that is in the public domain. All these competing power centers may keep the details of their AI capabilities research secret, but public AI safety research has a chance of being noticed and used by any of them.
Current sense of where we're going:
AI is percolating into every niche it can find. Next are LLM-based agents, which have the potential to replace humanity entirely. But before that happens, there will be superintelligent agent(s), and at that point the future is out of humanity's hands anyway. So to make it through, "superalignment" has to be solved, either by an incomplete effort that serendipitously proves to be enough, or because the problem was correctly grasped and correctly solved in its totality.
Two levels of superalignment have been discussed, what we might call mundane and civilizational. Mundane superalignment is the task of getting a superintelligence to do anything at all, without having it overthink and end up doing something unexpected and very unwanted. Civilizational superalignment is the task of imparting to an autonomous superintelligence, a value system (or disposition or long-term goal, etc) which would be satisfactory as the governing principle of an entire transhuman civilization.
Eliezer thinks we have little chance of solving even mundane superalignment in time - that we're on track to create superintelligence without really knowing what we're doing at all. He thinks that will inevitably kill us all. I think there's a genuine possibility of superalignment emerging serendipitously, but I don't know the odds - they could be decent odds, or they could be microscopic.
I also think we have a chance of fully and consciously solving civilizational superalignment in time, if the resources of the era of LLM-based agents are used in the right way. I assume OpenAI plans to do this, possibly Conjecture's plan falls under this description, and maybe Anthropic could do it too. And then there's Orthogonal, who are just trying to figure out the theory, with or without AI assistance.
Unknown unknowns may invalidate some or all of this scenario. :-)
I have explored many theories and worldviews in my life, but some loom larger than others. I found enough of merit in Hameroff and Penrose's quantum mind theory that I spent a year working with Hameroff. I've been a fellow traveler and sympathizer of Eliezer ever since he first showed up as an AI-focused transhumanist in the 1990s. And what I've discovered in the past few months is that Eric Weinstein's theory of everything ("geometric unity") has enormously more substance than you might guess, from its critics or its defenders on social media.
It bears comparison with string theory. String theory is this extraordinary generalization of quantum field theory that was discovered incrementally, which seems to be mathematically coherent, and which describes a zillion possible worlds, one of which may be our own. Geometric unity is more like a proposal that a specific highly unique quantum field theory exists, and that we inhabit one of its solutions. The jury is still out on whether the theory in question can exist mathematically; if it does, it's one of those intricate unique mathematical objects like the Monster Group. (The Muskian dictum that "the most entertaining outcome is the most likely" suggests that if it exists, it does so as a vacuum of string theory, which is not impossible.)
The amusing auxiliary realization I've just had, is that in principle one could combine all three of these paradigms as one's personal worldview. Weinstein's theory would be the fundamental physical theory, Hameroff and Penrose's "orchestrated objective reduction" would be the theory of how consciousness relates to physical reality... and what would be the Yudkowskyan component? It would be, not just ideas like dangerous superintelligence and coherent extrapolated volition, but the entire cosmological conception of acausal trade across a multiverse dominated by locally hegemonic superintelligences, most of whom pursue deeply nonhuman value systems like "squiggle maximization".
Even worse, it would be possible to dub this combined Weinstein-Hameroff-Yudkowsky theory as "WHY Theory" (I imagine the "why?" spoken in the plaintive tones of a normie debunker, facing yet another challenge to their sanity), and turn it into a recognizable brand in the contrarian-intellectual social-media space. The reason it could go far is precisely because each component really does target a blind spot in human knowledge (fundamental physics, nature of consciousness, destiny of intelligence in the universe), and provides a concrete answer that has more substance and logic, than academic defenders of normalcy are comfortable with acknowledging. I think the unsuspecting co-founders of WHY Theory also might not be entirely comfortable with this imposed intellectual cohabitation, but that's what happens when you put ideas out there and they take on a life of their own.
Stepping back from this scenario, my own opinions are as follows. I still bet on string theory, but Weinstein's theory has some chance (in the sense of logical uncertainty) of existing mathematically and of describing the physics of our world, and it corresponds to a genuine gap in the range of possibilities that have been rigorously explored by mathematical physicists. The same goes for the combination of Penrose's noncomputable physics and Hameroff's quantum cognition in the microtubules; furthermore, I deem the involvement of entanglement in conscious cognition likely for ontological reasons, and the microtubules of the axon are still the best proposal I know for where that might be happening. Finally, I am a skeptic of acausal trade (because of epistemic barriers) and of light-cone-conquering superintelligences (because of the doomsday argument), but I see the internal logic of both ideas, and don't consider them easily falsified.
Hameroff and Penrose’s quantum mind theory
Doesn't address "the" hard problem. The problem that Penrose is trying to solve , how the mind grasps mathematical truth, usn't much of a problem if you reject mathematical realism.
Eric Weinstein’s theory of everything (“geometric unity”) has enormously more substance than you might guess, from its critics or its defenders on social media.
Weinstein's idea has been devastatingly critiqued.
Doesn't address "the" hard problem. The problem that Penrose is trying to solve , how the mind grasps mathematical truth, isn't much of a problem if you reject mathematical realism.
Their theory has two parts, a neurobiological hypothesis and a "physics of cognition" hypothesis. The neurobiological hypothesis is that neuronal microtubules are information processors that contribute substantially to cognition (Hameroff was advocating this long before he united with Penrose), and specifically that they are quantum information processors. The "physics of cognition" hypothesis is that the relevant quantum dynamics includes a non-Turing-computable quantum-gravitational selection of eigenstates, which in its deepest form manifests as the metamathematical ability to transcend any particular axiom system via conscious insight into the meanings of the axioms.
So from the perspective of consciousness studies, this is one more materialistic theory of consciousness which makes specific (and unusual) claims as to which material entities and processes are involved in consciousness. You have to add a philosophical gloss (such as a metaphysics) to make it address a philosophy-of-mind problem like the hard problem. Hameroff's own interpretation is a kind of Tegmarkian panpsychism in which the physical world is made of Platonically real mathematics, but all those mathematical objects have a qualic aspect, and all qualia are some form of fundamental physics. This is not an outlook that he has developed systematically, in a way that would allow you to go through all the states of consciousness and identify the corresponding quantum states, or alternatively, to go through all quantum states and identify their conscious content and aspects.
Nonetheless, suppose that the theory (or any other "microphysical" theory of consciousness) had reached that level, presenting a comprehensive dictionary that allowed back-and-forth translation between physical and qualic descriptions of the fundamental states of the world. Would that provide an answer to the hard problem of consciousness? I think it would be progress, but the hard problem actually requires even more than that. The hard problem asks, why do physical states x, y, z have a conscious aspect? Even if we graduate to a neutral-monist theory-of-everything where every possible state can be interpreted "physically" or "qualically", this duality might still look like a brute fact with no further explanation. Answering the hard problem would now require explaining why our world has a dual-aspect ontology in which one aspect is consciousness - and at that point you might be getting into: the theory of possible worlds in an ontological and not just logical sense; the anthropic principle as a constraint on the metaphysics, and not just the physics, that you can find yourself inhabiting; and perhaps Neoplatonic speculations about Leibnizian first causes and Hegelian metaphysical necessities (or some equally deep metaphysical thinking).
Along with the hard problem, you mention the problem of accounting for knowledge of mathematical truth. Penrose is indeed tackling an aspect of that problem, but with the very specific priority of explaining how mathematical knowledge is possible, given the constraints on knowledge implied by Gödelian theorems. The basic version of those theorems applies to Turing-equivalent computational systems working according to some set of axioms; but even if you add "oracle functions" as extra computational primitives, you can still Gödelize the resulting entity and construct undecidable propositions.
Penrose's response has been to suppose that there is something transcendent about fundamental physical dynamics - transcendent in the way that Cantor's Absolute Infinite transcends all the concrete notions of the transfinite - that is at work in the material processes underlying metamathematical cognition, and that specifically this occurs in a quantum gravitational process that is also responsible for "state vector reduction" or "collapsing the wavefunction". It's an idea that certainly wins points for imagination, it might be appealing to anyone who has an intuition of limitlessness or transcendence as being inherent in fundamental reality, and the world of 4-manifolds that one expects in quantum gravity does actually include undecidable propositions.
There aren't many people who have embraced this approach to the challenge of Gödelian limits to physically based thought. The most common approach is to suppose that the human brain simply is genuinely finite in its capabilities and that there are mathematical facts that inherently transcend us. "Busy beaver numbers" are the main concretization of this, especially when used to rank the power of set-theory axioms.
My own view is that the finitist approach is the sensible one, but Cantorian infinitism is not totally out of the question; but also that both sides of this debate are un-rigorous when it comes to the nature of knowledge. Philosophical rigor requires more precision about the nature of intentionality than either side employs. The wellspring of the debate is a confrontation between phenomenology - the phenomena of mathematical reasoning and mathematical knowledge - and natural science - the apparent finitude of the information processing that can occur in the human brain. But neither side can anchor its intuitions in a clear picture of how mathematical thinking is related to neuronal processes, or even a principled answer to the general Searlean question of how brain states "represent" or manage to "be about" mathematics. Rigorous mathematical results about the capabilities of formal systems or computer programs, are then mapped onto the brain/consciousness/math situation in vague or speculative ways.
This isn't just a quibble, because knowledge of what axioms are about - access to semantic and not just syntactic information, so to speak - actually allows one to transcend Gödelian limits. Penrose cites the work of Solomon Feferman, e.g. his "completeness theorem" which says that everything in arithmetic is decidable if, along with the Peano axioms, you allow transfinitely iterated "reflection" on them. This shouldn't be surprising, because Gödel theorems and Gödel propositions are constructed precisely by appealing to metamathematical knowledge of what a formal system is actually about.
You could say that the Feferman completeness of human thought, is what allows us to prove the Gödel incompleteness of a formal system; and the debate between Penrose and cog-sci common sense, is a debate over whether human thought isn't literally Feferman-complete, but only extends some basic axiom system by a few rounds of reflection; or whether we have a physical ontology which really does allow a thinking system to be, in principle, fully Feferman-complete.
Let me round this out with some miscellaneous final thoughts.
I said Hameroff espouses a Tegmarkian panpsychism; Penrose on the other hand, is known for suggesting a vague circular relationship between three worlds of mathematics, physics, and consciousness. One might say that Hameroff asserts the identity of the three worlds and thereby collapses them into one, whereas Penrose maintains an agnostic attitude that the truth about it may be subtle and elusive. Regarding how all this is still not yet a solution to the hard problem per se, I think Chalmers writes about this in some of the more esoteric sections of his work.
My own interest in quantum mind theories derives especially from the belief that physical theories of mind have a severe sorites problem (one dimension of which you can see in this ongoing discussion), and that quantum entanglement, and fundamental physics in general, offer a way out in the form of entities which are "wholes" with complex internal structure and non-arbitrary boundaries. And my interest in microtubule-based theories of quantum mind, is because they are still the best candidate I've seen for how this quantum holism could actually be a part of conscious cognition (since the movement of quanta around the cylinder of the microtubule, could produce topological quantum states robust against room-temperature decoherence). I'm able to draw a line between the bare idea of quantum states in the microtubule, and the more elaborate theorizings of Hameroff and Penrose. But I still like to understand their ideas in detail.
Weinstein's idea has been devastatingly critiqued.
The observations in that paper, while technically valid, are only the tip of the iceberg. They are not proofs that Weinstein's theory is untenable, they are simply questions to which the theory must have an answer. The most famous critique from the paper is usually presented as a proof that Weinstein's shiab operator cannot exist. What is actually the case, is that the existence of the operator requires the complex form of his gauge group; and a complex gauge group, because non-compact, is argued to have physically untenable consequences. However, Weinstein is happy to bite the bullet and embrace the non-compact gauge group, because he thinks he has a mechanism that will de facto reduce it to a compact subgroup. A variety of such mechanisms are already used in other theories.
Weinstein hasn't helped himself by refusing to answer those questions (or at least, by refusing to respond to that author). Between his public relations strategy, and all the usual reasons why new ideas have an uphill struggle, even in theoretical physics, the public discussion of his theory is in a dismal state. However, there's enough information and insight out there, that a much deeper assessment is now possible. But rather than go on about it here, I'll just link to the blog I created in order to work through it all.
The “physics of cognition” hypothesis is that the relevant quantum dynamics includes a non-Turing-computable quantum-gravitational selection of eigenstates, which in its deepest form manifests as the metamathematical ability to transcend any particular axiom system via conscious insight into the meanings of the axioms.
OK, but that's another unnecessary problem, because you don't have to regard the mind as a (consistent) axiomatic system.
Along with the hard problem, you mention the problem of accounting for knowledge of mathematical truth. Penrose is indeed tackling an aspect of that problem, but with the very specific priority of explaining how mathematical knowledge is possible, given the constraints on knowledge implied by Gödelian theorems.
There's no way of confirming that we have a kind of knowledge that goes beyond "proveable from unfounded axioms" and "seems to work in practice".
Perhaps we aren't as limited as a single formal system, but then we might be using cheap tricks like switching between systems, using intuition, or tolerating inconsistency.
There aren’t many people who have embraced this approach to the challenge of Gödelian limits to physically based thought. The most common approach is to suppose that the human brain simply is genuinely finite in its capabilities and that there are mathematical facts that inherently transcend us.
From the anti realist point of views, If there are facts that transcend everyone, they aren't facts. They aren't sitting somewhere gathering dust for eternity, they just aren't there at all.
My own view is that the finitist approach is the sensible one,
Which finitist approach? An anti realist can regard infinties as suitable objects of study, no less real than other numbers.
The wellspring of the debate is a confrontation between phenomenology—the phenomena of mathematical reasoning and mathematical knowledge—and natural science—the apparent finitude of the information processing that can occur in the human brain.
Perhaps, but the phenomenological argument is really weak. Phenomenology is strong evidence of phenomenality..if it things seems to you to be certain way,then that is evidence that somethings seems so some way to you...but weak evidence of anything else.
But neither side can anchor its intuitions in a clear picture of how mathematical thinking is related to neuronal processes,
There isn't the slightest evidence that mathematical thinking is non neuronal or otherwise unusual ... from.neuroscience.
or even a principled answer to the general Searlean question of how brain states “represent” or manage to “be about” mathematics.
Under the form of anti realism known as fictionalism , there is no problem...the brain simply conjures up mathematical objects like dragons and werewolves.
It is the form of realism known as Platonism that causes the problems, since it needs to explain how immaterial entities affect thought, and therefore neural activity.
You could say that the Feferman completeness of human thought, is what allows us to prove the Gödel incompleteness of a formal system;
I wouldn't:understanding GIT might require some leveling meanness, but doesnt require transfinite levels....
and the debate between Penrose and cog-sci common sense, is a debate over whether human thought isn’t literally Feferman-complete, but only extends some basic axiom system by a few rounds of reflection;
...as you say.
My own interest in quantum mind theories derives especially from the belief that physical theories of mind have a severe sorites problem (one dimension of which you can see in this ongoing discussion)
That's another unnecessary problem. There's actually lots of evidence that consciousness is non binary.
* Drowsiness, states between sleep.and waking.
* Autopilot and flow states , where the sense of a self deciding actions isn absent.
More rarely there are forms of heightened consciousness: peak experiences, meditations jñanas, psychedelic enhanced perceptions , etc.
, and that quantum entanglement, and fundamental physics in general, offer a way out in the form of entities which are “wholes” with complex internal structure and non-arbitrary boundaries.
Like Chalmers , I don't see why any mathematical structure would have associated qualia.
Curated and popular this week