HNHacker News
TopNewBestAskShowJobs

Kednicma

185 karma · joined June 2, 2020

submissionscomments
Kednicma··on Stanisław Leśniewski: rethinking the philosophy of mathematics [pdf]
Since sets are 0-categories, we can't escape set theory when talking about structures like the natural numbers. A natural numbers object is a feature of a topos, preserved by topos functors (geometric morphisms). Nobody told me this definition; it's something I had to absorb for myself when learning topos theory.

Formally, let N be the natural numbers object (in some topos), let z : 1 -> N be the zero arrow, and s : N -> N be the successor arrow. Then z;s;s : 1 -> N is the arrow which chooses 2 as an element of the NNO. Since z and s are unique up to unique isomorphism, so is 2. Moreover, since geometric morphisms between topoi preserve finite limits, the NNO should also be preserved, and that includes 2.

When the topos we choose/define is (equivalent to) Set, then we get the standard ordinal-number definition of 2.

To use a pun, this lets us upgrade from Dedekind-categoricity to a more modern and natural sort of categorical categoricity.

Kednicma··on Stanisław Leśniewski: rethinking the philosophy of mathematics [pdf]
It's always worth recalling why we bother with set theory. Philosophical objections like Leśneivski's are extremely valuable and good insights which we cannot discard trivially. Maybe sets are not good things to study. The main reason that we study sets today is because they are a place where we could study ordinals and the rest of number theory. We know about two bananas, two apples, two trees, etc. but what is two itself? Set theory provides a capable if unsatisfying answer: Two is anything which is uniquely isomorphic to the second ordinal number, which happens to be a particular finite set, and since sets formally contain nothing but other sets, we can manipulate two as a set without having to know about bananas, apples, trees, etc.

The modern way to talk about this stuff is via categorification; [0] is a good high-level introduction.

[0] https://math.ucr.edu/home/baez/quantization_and_categorifica...

Kednicma··on A Horrible Experiment
Humans can't make unbiased computers. Every computer which humans make is indelibly human as well. This is the anthropic bias.
Kednicma··on A discretization attack [pdf]
Excellent paper. Aside from the main points, I notice that the key phrase "category theory" is in the abstract, to aid indexing, but there's no category theory here. I don't think djb's unaware of abstract algebra, so I wonder why he chose this otherwise-irrelevant phrase for indexing.
Kednicma··on Research questions that could have a big social impact, organised by discipline
Bounded proof search is NP-complete, and you can see this if you imagine transforming a bounded proof search into a satisfiability problem; go from "what, if any, proof starts at these assumptions and reaches these conclusions in less than N steps?" into "what, if any, assignment of Boolean values will encode a proof in O(N) values?" This is precisely what got both Gödel and Nash interested in what would become P vs NP.

Computers beating humans at chess is a great example of not being smarter, but being stupid faster. Both humans and computers approach PSPACE-complete problems like chess by approximating their searches and looking at trees of possible moves. We have no special insights into how to consider or evaluate moves; we have psychological profiles on which moves might be connected during practical play but no reason to require people to make such moves, and those profiles are largely limited to "openings", classifying the relatively few initial trees.

How do computers beat humans in chess? They do exactly what we do, but faster. This isn't smart, but fast stupid.

Kednicma··on Research questions that could have a big social impact, organised by discipline
This isn't true. NP-completeness and Turing-completeness both limit what an AI can reason about, and there's no reason to expect that AGI will spontaneously break those barriers.

To use the meme, an AI won't be smarter than us; it'll be faster at being stupid.

Kednicma··on Research questions that could have a big social impact, organised by discipline
This is fair. I started my comment by responding to each section, but I quickly realized that I wouldn't be able to make a constructive contribution.

First, the most important thing: Note that one of the responses to my original comment is feeling intense emotional relief at having some uncanny unease put directly into words. This is not the first time I've responded to this sort of article, and not the first time I've seen this sort of reply. I gather that it is difficult for folks to articulate the ambient horror of what this sort of writing is sketching.

Eugenics is recognizable from phrenology. They want to know why brain size matters and how to measure intelligence in different parts of the brain. They also want to know whether FAAH-OUT can make people happier and less sensitive to pain, and whether "extreme" human life extension is possible. They want to know how small the human race can get, and what the marginal utility of each additional person is. Indeed they ask whether utilitarianism isn't the ultimate ethical method.

Corporatism is most obvious in the framing. Various already-established consensus positions are turned into but-really or what-about questions, so that it seems as if climate change is not something that definitely exists, that we definitely could do something about, and that we definitely are not doing enough about. Corporations are portrayed as the leaders in innovation, while governments and the public are incompetent and slow to change.

Why do I say that these positions are tropes? Because I've read enough Dark Enlightenment literature to recognize its fingerprints, mostly. The authors are very slanted towards a meritocratic transhumanist utopia, where their very intelligent observations about the coming utopia will be rewarded with high stations and praise in the coming utopia. Meanwhile all of the implications about classism and injustice are carefully worded away so as not to be blatant and repulsive.

I recognize that you might not like my questions. But you also admit that they're substantive, and that's good enough for me. Honestly, they're barely my questions at all; each of them has been open for decades, I think. Maybe quantum computing is the most recent one to have something interesting happen, and that's changed our trajectory from asking "can we?" to "can we, bigger?" and "can we, cheaper?" and "can we, easier?"

Kednicma··on Hitler Youths – the rise of teenage far right terrorists [pdf]
Like, how did they mean it? Did they build guillotines? Did they kidnap members of the ruling class and subject them to public sham-trials before executing them?

It's important to examine this claim carefully, when compared with the existence of for-profit prisons, prison labor, concentration camps; also when compared with gun-toting culture, like concealed carrying, brandishing, gun salutes, standing one's ground, shooting to kill, bringing guns to school...

To be blasé and use your phrasing, going through high school, I happenh to know exactly [0] how many kids were lost because of gun and prison culture. Meanwhile I've yet to see kids using guillotines for anything beyond novelty-slicing melons.

[0] https://en.wikipedia.org/wiki/Thurston_High_School_shooting

Kednicma··on Research questions that could have a big social impact, organised by discipline
We've got eugenics, corporatism, sci-fi bait, and worse. I do like the idea of a field of "China studies" through the very specific lens that they've sketched, which focuses on the experience of Hong Kong throughout the past century, but otherwise, ugh, what a series of shallow tropes. Their "climate studies" list is papering over the fact that those studies were done decades ago and the conclusions were well-understood. Here's one question in each field which is better than any of theirs:

Genetics: To what extent does the RNA world (leadzyme, ribozyme, RNA bases) influence our DNA's actions; is there more beyond genome and epigenome?

Business Development: Enumerate all models of cooperation.

Climatology: Classify the climates of Venus and Mars under anthropogenic climate change; what will terraforming do to their climates?

Earth Science: Classify the minerals and rocks of Venus and Mars.

Neuroscience: Is demyelination preventable or treatable?

Economics: Are markets efficient? Prove that either P<NP (no) or FP=PPAD (yes).

Medicine: Can phage therapy prevent an antibiotic-resistant bacterial doom scenario?

History: Just let historians critique history books for a few hours live on stage, and broadcast the results. This will more socially impactful than any specific question or research agenda.

Law: I do have a soft spot for "What rights might digital minds have under current law?" It's a good question. However, a better one, in terms of social impact today, might be: How can the law establish effective oversight over its executors and legislators?

Statistics: Are there tighter bounds on neural net performance than PAC/VC theory? Alternatively, are neural nets essentially polynomial regression?

Philosophy: The Hard Problem.

Physics: Scale quantum computing to today's classical computational regimes.

Astronomy: Finish building planned telescopes; do the planned experiments have results which agree with current theories?

Political Science: How can two-party democracies be converted by the will of the people to multi-party democracies; can Duverger's Law be bent?

Psychology: The Hard Problem.

Philosophy of Science: The Demarcation Problem.

Sociology: How do societies collapse; how are technologies lost? Case studies won't cut it; we need models.

Mathematics: The Riemann Zeta Hypothesis. If metamathematics is allowed, then whether P<NP, P=NP, etc.

Research on these lines of thought does have a big social impact today just by consideration and exploration, and will have big social impacts if they ever actually answer the questions.

Kednicma··on The Universe Knows Right from Wrong
Humans are what, about 99.9% similar to each other on average? So that gives me 3 nines belief that yes, if I can understand, then you can understand; we're just humans after all. I think that you're being sarcastic; once we define categories, Cartesian closed categories, and elementary topoi, then categorical set theory is a straightforward sequence of diagram schemata. But I understand your point!

You're super-close to a deep realization: all words are only sensical to certain subcultures. IOW there's no absolute meaning to any word. In that perspective, we're both right; "1" is meaningless and 1 is categorical, and it's just a question of pointers vs. names.

Another quick and deep corollary is that reality is socially constructed; whenever a quorum of humans is mutually intelligible during a conversation, then they are agreeing on the local nature of reality. Humans can't construct global maps of reality, though, since they can't observe the Universe all at once. Indeed global maps of reality are forbidden by the Kochen-Specker Theorem.

Edit: To clarify for the audience, I have no appeal to higher categories here, and the definitions of category theory are not some sort of categorization or classification process, but axiomatic definitions akin to set theory.

Kednicma··on The Universe Knows Right from Wrong
To be extremely precise: 1 is the successor of 0. Anywhere that there is a natural numbers object, there's a semantics for this statement. What is socially constructed here is the choice of topos which hosts the natural numbers object!
Kednicma··on Brutality of Life Reading List
Yes, I claim that almost all humans partition humans into completely imaginary classes. I'm going to define a bigot as somebody who makes such partitions. Note immediately that there exists a partition of bigots from non-bigots iff non-bigots exist. Note also that, simply by being able to recognize the above statement as true, both of us (and any other readers) are bigots. It follows that there must only be finitely many non-bigots at most; almost all humans are bigoted.
Kednicma··on Physics is stuck
Let's finish working on our current experiments first, and then declare that we're stuck. We have particle accelerators, gravitational-wave telescopes, solar probes, and planetary probes which we're either still building or still operating, and we haven't finished falsifying-or-verifying all of our predictions.

It's certainly wrong to imagine that we need a Great Man in order to come and fix everything. Perhaps we need a Great Insight, but also perhaps the insight has already been published and it simply hasn't been well-adopted yet. After all, people are still today fighting against basic century-old quantum mechanics, hoping that reality is more material and determined than it actually is.

Kednicma··on Mark Shuttleworth Comments Following Ubuntu Community Friction
Sounds like a preparatory delegitimization of the community council. After all, if the community doesn't care enough, then Canonical can retake more direct control eventually and simply say that the community wasn't stepping up enough to take their own lead.

I wonder what the endgame is like here. Ubuntu is slowly being crushed under the weight of all of their experiments and different packaging needs, and outreach efforts aren't paying enough to sustain growth. Like a Ponzi scheme, but without the ill intent or even the understanding of what was increasing in cost over the years.

Kednicma··on Falsehoods programmers believe about time (2017)
I was hoping for information about time, but this list is almost entirely about calendars.
Kednicma··on What's wrong with social science and how to fix it
I don't really think that you're cogent here; it seems like you just wanted to say things which stand in opposition.

While it's true that there's no empirical evidence within humans for the question of p-zombies, Occam's Razor neatly shows that the world without souls is the simpler one; both worlds look just like ours, but one of them requires all of these additional unfalsifiable unverifiable claims about souls and consciousness and inner experiences and etc.

The Dodo Bird's strength comes from the multitude of different models of therapy which have existed over the decades. We know, from history, that the memes of psychotherapists leaked out into popular culture and slowly altered how we talk about thinking. Nonetheless people are more neurotic (more diagnosed with mental disorders) than ever before! So the memes of psychotherapy do not on their own decrease mental disorders. People will look back on our current decade and think how silly we were to focus on "medications", "hormones", "chemical imbalances", "neurotransmitters", etc. just like we now think it's silly to focus on "repressed memories", "dream interpretation", "hysteria", "oedipal urges", etc.

Humans cannot imagine anything completely novel. Every thought which a human ever has is unoriginal and hopelessly tied to the human experience. This is known as the anthropic bias and has been known about for millennia. If you believe in souls, you have an uphill battle in terms of evidence, including here.

Please stop believing in souls. There's no evidence for it and it turns your arguments to mush.

Kednicma··on What's wrong with social science and how to fix it
You're right that I'm thinking of very basic criticisms. In particular, there's zero evidence that humans aren't p-zombies [0] and no definitive rejection of the Dodo Bird hypothesis [1]; in combination, this suggests that psychologists are both wrong to imagine that there's anything interesting going on inside of a human's mental states, and also wrong to try to classify those mental states into appropriate and inappropriate states. Instead, what ends up getting studied is society's own idea of what ought to be happening inside our homoncular Cartesian theater [2].

Given these foundational issues, it's folly to try to support Big Five or any other descriptive model just by saying that it's a good fit for the numbers. Any principal component analysis will find something which factors out as if it were a correlative component. This dooms Big Five just as reliably as it dooms g-factors or Myers-Briggs or any other astrology-like navel-gazing.

(If you want an example of actual five things showing up again and again and again, mathematics has examples [3][4][5], but it turns out that when actual five things show up, then the reaction is not to serenely admire the correlation, but to admit terror before cosmic uncertainty. Psychologists do not seem to go insane and kill themselves like statistical mechanics or set theorists; have they really seen the face of god?)

[0] https://en.wikipedia.org/wiki/Philosophical_zombie

[1] https://en.wikipedia.org/wiki/Dodo_bird_verdict

[2] https://en.wikipedia.org/wiki/Cartesian_theater

[3] https://en.wikipedia.org/wiki/ADE_classification

[4] https://en.wikipedia.org/wiki/Monstrous_moonshine

[5] https://en.wikipedia.org/wiki/Classification_of_finite_simpl...

[6] https://en.wikipedia.org/wiki/Hard_problem_of_consciousness

Kednicma··on What's wrong with social science and how to fix it
> Even if all the statistical, p-hacking, publication bias, etc. issues were fixed, we'd still be left with a ton of ad-hoc hypotheses based, at best, on (WEIRD) folk intuitions.

This is the quiet part which most social scientists, particularly psychologists, don't want to discuss or admit: WEIRD [0] selection bias massively distorts which effects are inherent to humans and which are socially learned. You'll hear people today crowing about how Big Five [1] is globally reproducible, but never explaining why, and never questioning whether personality traits are shaped by society; it's hard not to look at them as we look today at Freudians and Jungians, arrogantly wrong about how people think.

[0] https://en.wikipedia.org/wiki/Psychology#WEIRD_bias

[1] https://en.wikipedia.org/wiki/Big_Five_personality_traits

Kednicma··on With new security law, China outlaws global activism
Sure, but also, fuck ethnofascism in all forms; when they come over here and ask us to censor ourselves, we say no [0]. We are obligated to resist it, not to shy away from it.

[0] https://www.npr.org/2012/09/19/161343395/why-does-china-want...

Kednicma··on Ask HN: What is your favorite motto?
"Everybody lies." ~ Gregory House

"Talk is cheap; show me the code." ~ Linus Torvalds

"To understand a program you must become both the machine and the program." ~ Alan Perlis

Kednicma··on TikTok reveals details of how its algorithm works
Here's the secret sauce: "Using machine learning, the algorithm serves videos to users based on their proximity to other clusters of users and content that they like." It's impressive how they arranged to be transparent for everything else, but kept this important part opaque.

Rumor is that the actual secret sauce here is human curation; people hand-select videos with high appeal and label them "viral", "popular", etc. in order to astroturf eyeballs and clicks. I suppose that admitting this directly would contrast sharply with the Chinese-harmonious-technocracy veneer that they work to project.

Kednicma··on Portland adopts landmark facial recognition ordinances
We're looking for one of:

* Defunding. There was a moderate budget cut already, but we want more, and we want the city to not just remove budget from PPB but actually put money into social welfare programs.

* Non-violence. The city council and mayor theoretically have the power to order the police to stop using tear gas right now, and similarly they could order the police to be less violent in all of their activities.

* Justice. The police are failing to arrest fascist street thugs, and instead send them text messages letting them know not to show up [0]. Meanwhile they do nothing to prevent federal officers from using unmarked vans to yank people off the streets.

* Polis. The police mostly live in Vancouver or its suburbs, over in Washington State. This is absurd; police ought to live in their respective precincts, just like how city councillors must live in their district. Even my credit union requires me to live in my county!

* Reform. There are many popular initiatives [1] which would force police work to be more humane and considerate, if only they were adopted.

I would consider any of those to be a victory. We got some defunding earlier, and protestors agreed that it was a good token effort but not enough.

Nobody wants the old routine of police violence and abuse. Nobody particularly wants to be tear-gassed, I think.

[0] https://www.nbcnews.com/news/us-news/disturbing-texts-betwee...

[1] https://8cantwait.org/

Kednicma··on Portland adopts landmark facial recognition ordinances
Sure. Do you want to explain why you believe that we need more police? Right now our nation is in the middle of a massive consideration: Maybe we have had too many police. This comes on the heels of police evolving from slave patrols [0], police torturing the populace [1], etc.

[0] https://en.wikipedia.org/wiki/Slave_patrol

[1] https://en.wikipedia.org/wiki/Jon_Burge

Kednicma··on Portland adopts landmark facial recognition ordinances
I live in Portland. This legislation is an obvious no-brainer; we've known since Philip K. Dick that this sort of technology is primarily meant to harm and control the populace.

As a nice side benefit, the sort of neoliberal surveillance-advertising monolith that we see in "Minority Report" is also banned.

Edit: So many downvotes! Do y'all want to use your words? It's starting to seem like HN users want to build panopticons and they're upset that we legislated that possibility away.

Kednicma··on Leadership in crisis – why the West needs Plato more than ever
Yes. I suppose my sarcasm didn't come through very well. To paraphrase a popular meme, Plato is only known to us today via his wrestling nickname. His opinion on leadership was that a philosopher-king, somebody both thoughtful and powerful, should dictate society. In his own life, he was politically entangled, but he was never given the power that he felt he earned.

I'm also reminded of Idiocracy, which checks all of the same boxes. I don't know whether Mike Judge was thinking of Plato, though.

Kednicma··on Leadership in crisis – why the West needs Plato more than ever
Right, we need a tough-guy philosopher-king who wrestles his opponents into submission.
Kednicma··on Dozens of scientific journals have vanished from the internet
This sounds like apologia from a big closed publisher (AAAS) explaining why open-access is supposedly bad. See, sometimes open-access journals fold, and when that happens, nobody knows what happens to the articles. But they'd like you to ignore two inconvenient facts: First, that traditional closed publishers effectively lose all articles by default by this metric! And second, that the Internet Archive, itself open-access, was essential to conducting the study in the first place!
Kednicma··on Simple wave-optical superpositions as prime number sieves (2019)
I am reminded of two curious statements which are not normally linked:

(0) There could be a Siegel zero [0], like an "indentation" in the Riemann zeta function. The prime numbers need to all collaborate to "shift" the zeta function a little in order to "buckle" the zeta function at the Siegel zero, so there's only really enough room in the primes to accomodate one Siegel zero at most.

(1) Ultrafinitism [1] comes in many flavors, including a flavor where there is a single largest prime number [2]. In this ultrafinite physical world, the largest prime determines how large any object can be before folding back in on itself. Since this would be a true statement about natural numbers, it would also be a sort of conspiracy amongst the primes, with only one instance.

Situations like this article remind me of both simultaneously: If there's any sort of conspiracy amongst the primes to forge a Siegel zero, it would have to be the same conspiracy which produces an ultrafinite prime candidate; the primes don't have bandwidth for two conspiracies.

[0] https://en.wikipedia.org/wiki/Siegel_zero

[1] https://en.wikipedia.org/wiki/Ultrafinitism

[2] https://sites.math.rutgers.edu/~zeilberg/mamarim/mamarimPDF/...

Kednicma··on What motivates the authors of video game walkthroughs and FAQs?
Jacob Geller had interesting commentary along these same lines last week [0]. In some ways, a walkthrough is an extremely detailed first-person critique: Exploring the game with the walkthrough as a guide, the author highlights each element of the game which they want you to experience.

[0] https://www.youtube.com/watch?v=Vr6pA15xuFc

Kednicma··on Why Past Attempts of Software Metrics Have Failed Us
This article is both too long and also, by virtue of being a linkfarm, too short. Let's try something simpler: All software metrics eventually fail to indicate useful directions for progress, by Goodhart's Law [0]. A corollary is that whatever snake oil is being peddled by the article, it will eventually be as useless as current metrics.

[0] https://en.wikipedia.org/wiki/Goodhart%27s_law

← PreviousPage 4 of 14Next →