Letter to Nature denounces the term “quantum supremacy” as racist

December 12, 2019 • 9:00 am

Yesterday, in a comment on this site, reader Invisible Airwaves called attention to a letter to the editor in Nature about an innocuous phrase used in computer lingo. It’s unbelievable what things will offend people these days—even innocuous phrases—and it’s starting to get me down as I see no corrective in the future.

So, a group of scientist Pecksniffs have taken issue with the phrase “quantum supremacy“, which means this, according to Wikipedia:

In quantum computingquantum supremacy is the goal of demonstrating that a programmable quantum device can solve a problem that classical computers practically cannot (irrespective of the usefulness of the problem).

And here come the termites, gnawing away at this term, a term we’ve all heard often these days. Why do they object? Because. . . . well, you’ve probably already guessed. The complete letter, which is indented, is below, and if you don’t believe me you can click on the screenshot:

We take issue with the use of ‘supremacy’ when referring to quantum computers that can out-calculate even the fastest supercomputers (F. Arute et al. Nature 574, 505–510; 2019). We consider it irresponsible to override the historical context of this descriptor, which risks sustaining divisions in race, gender and class. We call for the community to use ‘quantum advantage’ instead.

The community claims that quantum supremacy is a technical term with a specified meaning. However, any technical justification for this descriptor could get swamped as it enters the public arena after the intense media coverage of the past few months. [JAC: I seriously doubt it.]

In our view, ‘supremacy’ has overtones of violence, neocolonialism and racism through its association with ‘white supremacy’. Inherently violent language has crept into other branches of science as well — in human and robotic spaceflight, for example, terms such as ‘conquest’, ‘colonization’ and ‘settlement’ evoke the terra nullius arguments of settler colonialism and must be contextualized against ongoing issues of neocolonialism. [JAC: Note the postmodernist language here.]

Instead, quantum computing should be an open arena and an inspiration for a new generation of scientists. [JAC: It’s not an “open arena”? Really? How so? And it is already an inspiration for people working on this promising new technology.]

Well, maybe the term has that resonance in their view, but not in mine—or in many other peoples’. You have to be on the lookout for this kind of “offense” to find it, and then, when you do find it, you tell everyone that it has “overtones of violence, neocolonialism and racism through its association with ‘white supremacy'”. But it has no association with white supremacy save that one of the words in the two phrases is the same. If you’re going to play that game, why not also ban the term “white”?

In fact, the term “quantum advantage” is not a good replacement, for Wikipedia, at least, says it means something different from “quantum supremacy”:

By comparison, the weaker quantum advantage is the demonstration that a quantum device can solve a problem merely faster than classical computers.

Beside the three authors listed above, Nature also lists thirteen others in the “supplementary material”:

Co: signatories:

Syed Mustafa Ali Open University, Milton Keynes, UK.
Steve Brierley Riverlane, Cambridge, UK.
Hope Bretscher Cavendish Laboratory, University of Cambridge, UK.
Juani Bermejo-Vega University of Granada, Spain.
Helmut G. Katzgraber Microsoft, Redmond, Washington, USA.
Chris Granade Microsoft, Redmond, Washington, USA.
Alan Aspuru-Guzik University of Toronto, Canada.
Sabine Wollmann University of Bristol, UK.
Dominic Horsman Université Grenoble Alpes, France.
Anne Broadbent University of Ottawa, Canada.
Ariel Bendersky University of Buenos Aires, Argentina.
Cecilia Cormick National University of Córdoba, Argentina.
Shazeaa Nisa Ishmael University of Oxford, UK.

Readers with spare time might want to look up who these people are.

I have two questions here. First, why did Nature publish this overheated and ridiculous attempt to police language? Second, is there any way to stop this tsunami of virtue-flaunting before it inundates science, academia, and the rest of us?

121 thoughts on “Letter to Nature denounces the term “quantum supremacy” as racist

  1. “Open arena” is offensive because it recalls Roman gladiator games, where undesirable people were killed.

    (Satire warning: Don’t take this comment seriously.)

    1. Seriously, that’s actually a good strategy for dealing with extreme political correctness — out offend the offended! Turn the tables on them. No matter how politically correct someone is, there will always be a person further to the left who will outflank them and tsk tsk them for their problematic views.

  2. “why not also ban the term “white””

    I concur 110%!

    So the question becomes, what shall we name the mansion formerly known as “The Executive Mansion” in Washington? (note my avoidance of the loaded word, here….)

    [in my current teaching position, it is nearly impossible to say ANYTHING without offending several people, each in his or her own special way…. Whoops! I think that was another one]

  3. Let’s appreciate for a moment what had to happen to get here: somebody found this important enough. It took some more time to write down the arguments and get others on board. They had meetings or online conversations to go through the problem. Final drafts were written as they argued over specific wording. More people got involved and stopped their activities to read the proposal of the name change. Signatures were collected.

    I think we see this more often, and not because of outrage. The new currency now is attention. People will put time into generating attention (any) to put their name onto the map. I think we’re seeing Clickbait 2.0 now. That’s no longer straight forward clickbait, but still “polarising nonsense” that exploits outrage and virtue signalling. The whole media ecology is participating.

    1. That’s all true, but you overlooked the time spent by Nature’s editor in deciding to publish the letter once it had been received.

  4. Well, this is serious. I now have to forget I saw the second part of three movies, the Bourne Supremacy. That just ruins the third movie and I have lost transition and can no longer follow the plot. This is a disaster.

      1. Bourne sounds too much like “born”, and might be construed as anti-abortion, or alluding to “born again” and thus injecting religious fundamentalism.

  5. Lots of terminology would trigger these folks. They need to get busy with those letters! Let’s see…
    Slaver ants
    Ant colony
    Little brown jobbies
    The genus Homo
    Cro Magnon man

    I invite others.

        1. – Sneaky copulation (the sneaky male does not care whether the female agrees)

          – Sperm competition (different males competing inside a female)

            1. How about all the products designed to get rid of “free radicals”… so they’re implying that radicals ought to be imprisoned?

  6. “We object to this term being used because it includes a word that people like us now use so often and apply to everything that it now offends us to hear it.”

    Maybe, after they rid everything but critical studies of the word “supremacy,” they’ll move on to “division.” Imagine how many opportunities for offense that will provide, and how many jobs for professional “diversity, equity, and whatever” officers it will require to figure out replacement words! So much opportunity for offense, so much opportunity for BS job posts. The great thing about all of this is that they’ve created their own economic sector (of both social and economic value) in which only they can participate/gain from, but which they can use to threaten everyone into giving them money and social status. It seems they’ve learned from what Al Sharpton has been doing for decades. It’s brilliant and utterly destructive.

  7. What is most distressing about this idiocy is that it may repel people who would otherwise sympathize with real social justice movements that fight racism and other inequities in society. Trump and his allies would love to see more letters like this.

    1. The tactical stupidity of these people is just infuriating. It’s like they wake up every morning and think to themselves ‘what would be the best PR for anti-PC conservatives like Trump?’, and then immediately set about handing it over on a platter.

      I swear, if we just cut out left identity politics and the more extreme forms of political correctness we would defang Donald Trump overnight. He would be rendered utterly impotent.

      1. For a great way (to my mind) to de-fang Trump, look to what young miss Thunberg just did (again) in response to another example of Trump’s shitheadedness.

        I know it’s small potatoes, but she made me smile this morning.

        1. I liked that too. It was of the Buttigiegian* school of comebacks; wry and dismissive.

          *there’s got to be something better than this word

      2. I know several people who voted for Trump because they liked “Making Liberals Cry”. Sad to think that is a criteria to vote for but yeah this thing is what makes them mad at all liberals, and miss out on the important things real liberals believe in.

        1. That’s OK, they’re also supporting vaping, for the same reason. Honest! National Review devoted two pages to it. (To their credit, they thought it was silly.)

    2. …Having said all that, while I concede that stuff like this is harmful to the electoral chances of liberals and left-wingers, it doesn’t make me any less contemptuous of the kind of person who uses inconsequential PC flotsam like this as justification for voting for Trump/Brexit/whatever. This in particular: it’s just a letter; it’s signed by some scientists, sure, but it’s still just a letter. If I’d seen it in Nature I’d have thought ‘that’s a bit silly’, and gone on to the next one.

      1. I think that very few people would consciously decide to vote for Trump because of a letter such as this. However, it could create an unconscious tipping point for undecided voters. Sometimes little things have more impact than big things.

        1. Obviously not on the basis of a single letter like this. And I don’t think that many Trump voters are interested in the etymological back and forths of the quantum mechanics community.

          But there’s a thriving cottage industry on the right that does nothing but aggregate stories like this and pump them out to readers and viewers.

          (Politically I cannot blame them – it’s PR gold. But morally I think it’s laughable that they pretend this stuff is important.)

      2. I don’t think it’s letters like this one that resonate negatively with Trumpians. However, many people use similar memes in everyday conversation. I think many of us that read this website have heard them and been at least mildly disgusted by them. Of course, no such incidents would ever make me vote for Trump.

        My theory is that this kind of thing doesn’t really rile up the Trumpians much either. They simply use it as cover for their real reasons for siding with Trump: xenophobia, racism, white supremacy, bullying, and various other gripes that they would prefer not to state explicitly.

        This kind of misdirection has been raised to an artform by Trump. His presidency has been a tutorial for his GOP sycophants, supporters, and Trump-friendly media outlets. As was pointed out on CNN recently, all have gotten better at it as they’ve learned from the master.

  8. What can we do about it, our host asks? I’m not totally convinced about this, but I think that ridicule might help. Seriously. Point and laugh at them. I am of an age where ridicule wouldn’t have worked (we boomers are a dour lot) but today it can be very effective. Blow back is always a risk, but if there are any modern ideas more deserving of ridicule than the Woke movement, I’ll eat a MAGA hat.

    1. Bravo. Laughter and ridicule are usually the most potent weapons. George Orwell once wrote that one reason the British Army never considered adopting the goose-step is that people would just laugh at it.

      1. I remember reading that essay years ago but I can’t remember the title. Orwell said something to the effect that a soldier doing the goose-step was like a bully who made silly faces at you while beating you up, the underlying message being, “I dare you to laugh at me.” He argued that the goose-step would never be tolerated in England because the English have less respect for the military (and authority in general) than the Germans did.

  9. Looking at the probable security competition with China in the coming years, and the fact that China has a GDP by PPP that is higher than the U.S. and growing faster than the U.S., as well as the fact that they have a higher number of conscript age males than the U.S., it occurred to me that at least the U.S. still has a scientific and technical edge.

    Watching the Academy destroy itself, and academic standards in the process, it does not appear that the U.S.’s edge in science will last long. Perhaps “supremacy” is problematic because people recognize that American supremacy is on its last legs, and we are at the dawn of Chinese Supremacy, and denial is better than looking at the fate of Tibet or Guangxi and extrapolating a future.

  10. … and it’s starting to get me down as I see no corrective in the future.

    This too shall pass, Brother Coyne.

    Keep the Faith (vs. Fact). 🙂

  11. real social justice movements that fight racism and other inequities in society

    I am still waiting for the “real” communism that “fight[s] racism and other inequities in society”. “Actual” communism just seems to concentrate power in the hands of one person while killing millions.

    At least “actual” social justice doesn’t have a seven to eight figure body count yet. [Unless you want to count the millions in South Africa who died of AIDS because the president blocked anti-retroviral drugs on the basis that the viral theory of AIDS was a white supremacist plot.]

  12. The supremacy of the number 13 over the number 7 is just awful (when you subtract, the answer, 6, is positive!), awful especially given the scientifically proved superiority of 7 for good luck. Would any of you care to join me as co-authors of a letter to Inventiones Mathematicae advocating the banning of that phrase above? Maybe also to Annals of Mathematics, but the recent supremacy of the former journal is arguable…oops, the inferiority of the Annals … oops! What is a mathophile to do??? For safety I better stick to reading Derrida and the postmodernists exclusively from now on–or am I getting these supreme intellectuals confused with the equally supreme wokists?

  13. For those with more questions about Quantum Supremacy, start here:

    Scott [Aaronson]’s Supreme Quantum Supremacy FAQ!

    Since we were recently talking about false extensions of fundamental scientific results, I want to note another one here. The definition of Quantum Supremacy makes it sound like Quantum Computers can calculate things that regular computers can’t. As Scott points out, there is no proof that they can. There’s no proof that an efficient algorithm can’t be found that will simulate QC algorithms.

    Some people claim that human brains use quantum computing or quantum effects, giving them powers that can never be simulated on non-quantum computers. There is no scientific basis for this claim.

    1. QC’s can’t calculate anything a conventional computer can’t, so far as is known today, but there are things that they can do quickly that will take impractically long times with conventional systems. There is no proof that conventional systems can’t efficiently emulate a QC, but it is very strongly suspected. This is, essentially, the P = NP? question, as many of the tasks that a QC is well suited for are NP complete.

      1. What would be an example of a qualitatively different kind of mathematical problem, that neither classical nor quantum computers can solve, but that is nevertheless solvable in principle?

        Because if QCs can’t solve problems that are fundamentally impossible to solve in the first place then that seems like a slightly unfair criticism.

        (I don’t even understand how it could be conceivably possible: that there could be problems that we know are solvable, but that we also know are not solvable by any kind of computer… Since we are basically computers ourselves{ie. there’s nothing qualitatively different about the physics of our brain and a computer} that doesn’t seem to make sense to me.
        And I appreciate this is probably getting as hard to follow as it is to write.)

        1. The current state of the science is that QC differs from classical computing only in terms of efficiency. I guess some scientists have a hunch that there’s a gap between the two other than performance but it hasn’t been proven.

          Another way of saying the same thing is that all known algorithms can be implemented on Turing machines. Some propose that more powerful machines exist but it hasn’t been proven.

          So, in short, the set of “problems that we know are solvable, but that we also know are not solvable by any kind of computer…” is empty as far as we know.

          1. “So, in short, the set of “problems that we know are solvable, but that we also know are not solvable by any kind of computer…” is empty as far as we know.”

            That’s what I thought. And I don’t see how it could be any other way either. Our brains are just computers after all. The underlying physics is the same, nothing spooky is going on. If a human brain can, in principle, solve a problem(even if it doesn’t know how to right now) then so can a computer.

            I suppose plenty of people would disagree that our brain is just a computer, but the onus is on them to point out what makes wetware qualitatively different from software. And I’ve never heard any convincing answers.

          2. “I guess some scientists have a hunch that there’s a gap between the two other than performance but it hasn’t been proven.”

            No, Paul, as far as quantum versus classical is concerned, the opposite has long ago been proved mathematically, more-or-less from the invention of these ideas by Manin, then Feynmann, then Deutsch in that order, that the difference is entirely efficiency–but so extreme in human terms so that many instances are known to be doable by a Q.C. once constructed, and but almost certainly not humanly doable classically.

            As for theoretical machines going beyond quantum computers, there’s nothing of importance, AFAIK. You’re perhaps asking, Saul, for someone to demonstrate the falsity of Turing’s most fundamental result giving his version of the most general meaning of the word ‘computable’ [not ‘efficiently computable’]. The many other versions of ‘computable’ have all been shown to be equivalent, including Church’s which preceded Turing’s, using the lambda calculus, but unknown to Turing at the time. Didn’t Turing go off to Princeton in the late ’30’s and do a Princeton Ph.D. under Church? (maybe just so the grant to work there was satisfied, since Turing hardly needed more formal qualifications by that time!)

            The other possibility of what you might be asking is for a kind of machine that is ‘merely’ vastly more efficient than even the (not yet constructed) useful quantum computers. Maybe some fantastic advances in fundamental physics which make quantum theory just a good approximation to the truth will eventually do that—but not during our lifetimes—a post-quantum unification of gravity with the other 3 forces and more which also shows general relativity to be only an approximation???

            1. There is work on “hypercomputers” or “super-Turing” computation. I have contributed (negatively, skeptically) to this literature, most recently in the volume _Computing Nature_, showing that the most plausible proposal on offer is still ridiculous.

          3. We don’t actually know this – the “set up time” on quantum machines may kill the apparent speed up. There are other models of computation (or of factoring specifically) where this is the case. My colleague Amit Hagar from my UBC days has done stuff here.

            1. I would certainly consider “set up time” a measure of efficiency. The important thing is that, as far as we know or can prove, all computation is Turing equivalent (there are less powerful kinds only) and quantum computing devices are just another implementation, though one with interesting efficiency characteristics.

      2. “This is, essentially, the P = NP? question, as many of the tasks that a QC is well suited for are NP complete.”

        This is incorrect; we don’t know of any NP-complete problems that can be solved efficiently by quantum computers. There are problems that can be efficiently solved by QC that (as far as we know) aren’t in P (i.e. efficiently solvable by a classical computer), but those problems are not (again, as far as we know) NP-hard, and thus not NP-complete.

        Wikipedia has a nice Venn diagram showing the (suspected) relationship between BQP (the problems efficiently solvable with QC) and other well-known complexity classes (and a more extensive discussion in the main article).

        Mind you, if it does turn out that P=NP, it wouldn’t surprise me if BQP was also equal to them.

        1. “…but those problems are not (again, as far as we know) NP-hard, and thus not NP-complete…”

          Because of that last comma, separating “know” from “not”, I think this could be confusing.

          For my own reassurance, is the situation now, and for quite awhile in the past, that integer factorization is
          1/ not known to be NP-complete, and
          2/ conjectured not to be NP-complete, but also not in P?
          (though if everything collapses and the funders of quantum computer research want their money back…!!)

    2. Thanks for reminding me of this.

      To hopefully add a bit on less technical stuff than the actual Google problem used, and hopefully do it correctly, and since Aaronson’s blog is clearly for semi-experts or more:

      1/ If a machine, not necessarily universal, but based on the quantum computer principles, succeeds in giving an answer within 2 or 3 seconds to an instance of a general problem (see last sentence of number 6/ below for an e.g. of “instance”), and
      2/ if there is at present no classical computer plus program which can do the same in less than 2 or 3 decades, and
      3/ despite the ‘forever’ known fact that in principle nothing a quantum computer can solve cannot in principle be solved using a classical setup with very very ridiculously large but finite physical resources, including time which might be millions of factors larger than the time since the big bang,
      4/ then, despite the ‘at present’ in 2/ above, it will be considered that quantum supremacy has been achieved.
      5/ Of course, that ‘proof’ of 4/ would need to then be withdrawn if human ingenuity later somehow made number 2/ false.
      6/ There are well known instances where 2/ now holds, where 1/ would hold if only a large enough good universal quantum computer could be constructed, where it is strongly conjectured that 2/ will always hold (one of the Millenium Problems in mathematics), and where once the answer is spat out by the quantum machine, it would be very easy for a standard classical computer to quickly check that it’s correct. (e.g. It is very easy to multiply two large primes, with possibly about 150 decimal digits each to check that the answer is what you were initially given, near 300 decimal places long, but unknown yet and conjectured unlikely ever, to be able to classically get those two factors from the ‘300’ in less than months.)

      As to the human brain, I remain firmly agnostic; I would not like to be stuck in a public debate against Roger Penrose on this, which of course is out of the question anyway.

      1. I *did* informally debate Penrose on this! He misunderstands (or did at the time) the limitative results, as Feferman showed. He misunderstands heuristic programs (as Dennett and others kept telling him), misunderstood physics (as Stenger pointed out) and misunderstood neuroscience (as Pat Churchland and Rick Grush pointed out). I was only at the time confortable with talking through Dennett’s point (as I independently came up with it), but I have written about the reconciling the rest.

  14. I looked up two of them (Broadbent and Aspuru-Guzik), who appear to be seriously involved in computational research, with respectable publications, and two others are
    at Microsoft, and thus presumably are geeks by profession. So, I guess we can’t simply dismiss them as offense hobbyists with too much time on their hands. Their professional respectability suggests, alas, that this sort of word magic is seeping into serious fields.

    Will we next have to worry about terms like dark matter, dwarf stars etc. etc.?
    Worse still, in molecular biology we have to deal with those blobs on our Petri plates we call, uhhh, pardon us, colonies.

    1. Microsoft, Google, IBM and others all have a hand in quantum computing. Guzik is a reputable physicists. My guess is he thinks the terms is stupid, regardless of the ludicrous insinuation that it’s racist.

      1. Bending over backwards I think it’s an unwise choice of words… Do they call themselves ‘quantum supremacists’?*

        If I was starting from scratch I’d just call it something else. Quantum Max Power. Quantum Advance Intel Plus. QAGA?
        But now it’s out there it seems like it’ll be rather difficult to change the name, and on a list of social justice priorities it’d be…quite low.

        *It reminds me of a game that came out recently called Death Stranding – before it was released we discovered that the supernatural-y enemies in it are called ‘Homo-Demens’. Pronounced homo demons. Most gamers assumed that it’d get snipped once the developers realised the unfortunate implications, but the game came out with the same ridiculous name. No-one really batted an eyelid and no-one was particularly bothered.

        1. That’s a stretch. Quantum supremacy is simply nice name for a short-term quantum computing research goal. I’m sure they do not call themselves “quantum supremacists”. I suppose the research group that claims the prize might call themselves that in private gatherings and after many beers.

          1. Well, ‘quantum supremacists’ is the natural choice of name for people who practice quantum supremacy. I think it’s quite funny and I don’t think it’s a stretch.

            In fact I’d be both surprised and disappointed if the other mathematicians and physicists don’t start calling them quantum supremacists as soon as they get the chance.

            1. C’mon, it’s the supremacy of one type of machine over an earlier type. Scientists who work on the theory and construction of quantum computers are not going to start referring to themselves as ‘having supremacy’ over other scientists (such as Alan Turing, for crap’s ache). When it happens, it will simply be the physical realization of a mathematical theorem, especially in the unlikely case that P<NP has already been proved by then.

      1. No, a syllogism asserts to be a known fact that [(P & Q) implies R], for certain sorts of propositions P , Q , and R. But it has no need for P and Q to be themselves known facts. Indeed, one might use it ‘practically’ when knowing both Q to be a fact and R to definitely not be a fact, thereby deducing that P is not a fact.

        More generally that’s what formal logic is all about, started by Aristotle with syllogisms, and vastly expanded only 2200 years later, starting mainly by Frege.

        I’d make no distinction, in principle, between logic, formal logic, symbolic logic or mathematical logic.

        At the Medieval Institute of Pontifical Studies–sorry, the Pontifical Institute of Medieval Studies–the theologians might disagree with a few assertions above. But these days at most universities, philosophy departments aren’t stuck with the retardation by a millenium+ of human intellectual advancement caused by St. Paul. Nor are mathematics departments. Syllogisms as a topic is a tiny special case of 1st order logic.

        1. It is possible that some variety of logic will be helpful in creating Artificial General Intelligence. On the other hand, they all are brittle by design while human thought is notoriously flexible. On yet another hand, we don’t really want our AGIs to suffer human uncertainty. If it’s required for human level intelligence then ok but that’s not been proven.

          1. I assume “AGI” is a common term, but I hadn’t seen it myself. Perhaps it means the same as ‘AI’ meant from Turing’s later time till a few decades ago.

            The pushers of that earlier, never achieved, putative AI, not Turing, but Marvin Minsky (a CS refugee from Algebraic Topology) for example, gradually gave it a poor reputation it seems to me, what with their more-or-less 5-year plans asserting that fluent language translation, brilliant chess playing, etc. would be done by AI within a matter of a small number of years. This seemed to continue from maybe 1955 to 1980 or later. See for example Joseph Weizebaum’s old book “Computer Power and Human Reason” (though he has other criticisms as well, related to grossly inappropriate computer applications in psychiatry).

            In any case more modest advances in something that’s still called AI, really good ones, have happened in recent decades, though much of it, not all of course, has come from pure brute force, namely speed and memory.

            I’m puzzled by what your first sentence means e.g. what an example of “variety of logic” might be here. John Burgess’ textbook “Philosophical Logic” (?) is a good study of temporal, modal, intuitionistic logics. He doesn’t pull his punches in reviewing some others, e.g. Graham Priest with propositions which are supposedly simultaneously true and false. To be overly mean, I might generalize to ‘quirky Australian logic’. But I know of no advocates of those, neither the invaluable nor the unvaluable logics, as being able to contribute to AI, other than the three above plus standard 1st and higher order logic.

            I might rewrite your “..they all are brittle by design while human thought is notoriously flexible.. ” as ‘reputable logics all are precise by design while human thought is notoriously sloppy..’ ! Regurgitating the sloppiness surely is not going to help in creating AI as predicted first by Alan Turing, though it might help in studying the human brain and consciousness.

            1. AI used to mean AGI but then the neural computing revolution took AI and made it theirs so the people working on duplicating human-level intelligence had to invent a new term.

              Those working on AI in the 60s were somewhat naive. They thought that that higher-level cognition, the interesting part, could be duplicated using some kind of logic and symbolic processing. The fact that humans can hold conflicting ideas in a single head was thought to be a flaw that they should and could avoid in AI. Now we think it is actually a feature, not a bug.

              The non-AGI AI has made huge advances but not much of it seems to apply to AGI. AI algorithms must be trained on millions of labelled examples in order to achieve competence in a fairly narrow application. This is not how people learn. Also, it can’t generalize its learning to applications other than what it was trained on. This is also not how our minds work.

              I have no problem with your “rewrite” of my statement. I think we’re both talking about the same thing. I’m leaving it as an open question whether “sloppiness” can be captured in some kind of logic. Actually, since logic is such a loaded word, perhaps “mathematical formalism” would be better. There is such a thing as fuzzy logic, for example. I’m not saying that it captures the brain’s logic just giving it as an example where logic can be sloppy in a sense.

              I believe the brain does have implement some kind of algorithms. Once we understand them, it is possible we can take the best parts, throw away the bad parts, and create a mathematical formalism based on what’s left, just as we learned how to fly from studying birds.

              1. ” The fact that humans can hold conflicting ideas in a single head …Now we think it is actually a feature, not a bug.”

                I assume that thinking jesus christ could walk on liquid water, versus scientific knowledge about the properties of liquid water and of the human body, is not one of those pairs of conflicting ideas regarded as a feature!

                I’m ignorant of any example of genuine new knowledge for which fuzzy so-called logic is needed. Somehow the Princeton mathematician/philosopher John Burgess didn’t have a chapter on fuzzy logic in that book I mentioned earlier (nor even a mention, IIRC).

                The corridors where “..logic is such a loaded word..” are disjoint I think from the set of those who agree with Keith’s remark above about Bunge and the formal nature of logic. I suspect the fuzziness of people walking that (hopefully solid not liquid!!) corridor has more to do with the uses of words, such as ‘consistent’ or ‘contradiction’ etc., with meanings different from how logic uses them (and likely quite unformulated definitions by such people).

  15. Now that I actually have a few minutes, a couple more terms that I have seen PC/woke resistance to in my professional life (teaching/practicing engineer):

    Master/slave (refers to communication between devices-comp sci/engineering)

    Jig (a construct that ensures repeatability, often by guiding a tool such as a drill, an usually holds or mounts to the work being machined-engineering/manufacturing)

    Niggardly (stingy– has no connection to “the N word”, but people are ignorant)

    Bonding (when I was teaching a Chem section many, many years ago– the offended individual was connecting the word to sex acts. I learned a LOT more than I wanted to know about any student through that one complaint… I wish I could make this stuff up)

    There are more.

              1. Interesting. According to Wikipedia:

                “Unique hermaphroditic connectors, commonly referred to as IBM Data Connectors in formal writing or colloquially as Boy George connectors were used.[20] The connectors have the disadvantage of being quite bulky, requiring at least 3 x 3 cm panel space, and being relatively fragile.”

                Ha! Fragile Boy George connectors! However, sex still wins out.

    1. I’ll bite. What’s wrong with “jig”? At work we have a blacklist of thousands of words and phrases we’re not supposed to use. Whoops, “blacklist” is on the blacklist…

  16. It’s unbelievable what things will offend people these days—even innocuous phrases—and it’s starting to get me down as I see no corrective in the future.

    Here’s the corrective. Use whatever words you see fit, and if someone objects, tell them to go f**k themselves. Or if that is too blunt, you might adapt Evelyn Waugh’s all-purpose response card, which said: “Mr. Evelyn Waugh sincerely regrets that he cannot do what you so kindly suggest.”

  17. If supremacy is racist, is the word “supreme” also racist? When I eat a burrito supreme, am I ingesting racist ideology? What about when I listen to Diana Ross and the Supremes?!

  18. “Pecksniffing” like this is an old story in computing. In fact the first instance of this type of language policing that I remember encountering – some 20 years ago! – was precisely in computing. Some people (who, what a surprise, rarely had anything to do with computing beyond using a computer for work!) took exception to the terminology used to describe the priority of hard drives in a computer, where the primary unit on an IDE interface is called “master”, and the other is called “slave”. I’m pretty sure every reader of this blog site can imagine the sort pomo word salad that made up the “objections”.

    1. And my old Landrovers had, for operating the hydraulic clutch, a MASTER cylinder and a SLAVE cylinder.
      I can remember double, then triple, then quadruple clutching to get the damn thing over those hills north of Superior and back home to southern Ontario before needing to get parts and get underneath and fix it!

  19. Getting published in Nature has long been considered as a major accomplishment for scientists. Now, instead of doing groundbreaking work, you can get in Nature by publishing a grievance article…

      1. I understand the difference. I meant it as a bit tongue in cheek. People have made the same observation about the science fiction short stories published in Nature as well.

  20. Two things are still unresolved for me :

    1. Data, observation. Perhaps a table, a graph. Where is it in this pape? As the paper makes a truth claim about the tangible world – namely, the genetic makeup of individuals in the “physics world” wouldn’t it be required?

    2. There is still the problem of individuals who grow up with limited access to resources, yet of good aptitude, possibly getting outcompeted for positions in the “physics world”. Even if the author has a point, what would that point actually do to solve the problem?

Leave a Reply