kbear

Karma: 471

kbear 6 Jun 2026 15:35 UTC
8 points
2
on: Freud heard a rumor that Science existed, and had a wonderful dream
quite sure that the premise is false
??? sorry, what are you talking about? peeing on fires is fun. truly, have you never peed on a fire? you can put out the coals. it sizzles comically.
i won’t apologize for Freud in this comment, but let’s not repress ourselves to prove him wrong.

kbear 1 Jun 2026 3:59 UTC
11 points
2
in reply to: johnswentworth’s comment on: Some Dating Stories
hey, i don’t think that you do, and have not at any point. but we’re not legislating what’s in your heart, we’re trying to understand why some had a negative reaction to your word choice.
my understanding of your model of the offense:
1. people see the word ‘fuck’
2. they think of ‘very erotic’ sex
3. they are not comfortable with this, for possibly complex reasons
4. they are offended.
i suggest instead that what’s happening is
1. people see the word ‘fuck’
2. they think of what is (to them) a far more common usage of the word: a sort of conquest/damage
3. gross!
anyway, sorry if i’ve given the wrong impression in this thread. i don’t think you’ve done anything wrong; i think you were misunderstood. you used a word one way, but readers took it another.
you’ve since clarified what you’ve meant, which is all that the norms of productive discourse can ask.
i disagree with your model of the reader, and present an alternative. but i get the feeling you’re not that interested in this and/or are very confident in your understanding.

kbear 31 May 2026 4:45 UTC
3 points
2
in reply to: johnswentworth’s comment on: Some Dating Stories
conquest is fun and erotic. conceptualizing your escapades as conquests is gross.

kbear 31 May 2026 3:07 UTC
7 points
1
in reply to: johnswentworth’s comment on: Some Dating Stories
disagree. i have no problem with fucking, but flinched at your word choice here.
for better or worse, ‘to fuck’ also connotes conquest. this is a fun and erotic dimension to the experience for those participating, but in this clinical/reflective setting it reads very awkwardly. compare:
Fraction of women I’ve claimed fully who have in fact come back for more: about 80%
ew! gross dude! don’t think like that—not outside of the bedroom! definitely don’t talk like that.
the distinction you make—that fucking is a particular kind of sex—is a true distinction. but i’m not sure how it’s relevant here? if you were measuring something like “fraction of women i’ve had loving, tender sex with who have come back for more: 20% / fraction of women i’ve fucked who have come back for more: 80%” then… fine! i agree with the word choice in that context: it emphasizes something important in the discussion.
but here… what is the word offering that the phrasing “Fraction of women I’ve seen again: 80%” does not provide? excepting of course the implicit advertisement for your services (“if you like getting your pussy ate by a vampire^[1], jsw has a really big stash of viagra”). but take that to okcupid, please!
1. ^
  affectionate!

kbear 29 May 2026 16:55 UTC
1 point
0
in reply to: J Bostock’s comment on: sam’s Shortform
yeah, to be clear: my experience [of prompting the models] is that it’s basically impossible to get them to have judgment that they don’t naively exhibit. no amount of “think carefully about whether its relevant” will suddenly elicit good taste; it appears that “always do this” and “never do this” are too heavy as attractors, and the boundary [for any behavior worth asking for] is fractally complex.
i think we’re in agreement about the state of the world, here. apologies for the semantic quibble.

kbear 29 May 2026 14:32 UTC
1 point
0
in reply to: J Bostock’s comment on: sam’s Shortform
overfit
would it seem this way in the reverse case?
say that anthropic had touted the memory feature, but claude never reached for it. then we got several posts saying “ah, what a useless feature, huh?” and then in the next point release, claude was more likely to reference memories.
that would just be… normal model steering, right?
(i share your reaction—just trying to interrogate it!)

kbear 29 May 2026 11:16 UTC
2 points
0
in reply to: zroe1’s comment on: Trees are mostly made of air and a generalizable lesson for AI safety
thanks!
these examples are much more convincing than the essay’s. i haven’t looked at each closely, but if these several inventions are all due to deep understanding, that is indeed a recommendation for deep understanding.
were i not already convinced of the importance of deep understanding in advancing the frontier of human knowledge, this comment would be a good starting point.
may i ask: why not use these examples in the original essay? the essay touts results like “know dates”, “save face if documentarians come round”, “finish a degree / turn down quant offers”. these are—lightly now—relatively less interesting to me than, like, “invent turing machines”.
if i model the goal of the essay’s author as “try to convince me that deep understanding is worthwhile”, then i am very confused. why would such an author appeal to the examples in the essay, if they could have reached for the ones in the comment?
do i misunderstand the essay’s goal?

kbear 29 May 2026 4:39 UTC
9 points
3
on: Trees are mostly made of air and a generalizable lesson for AI safety
what power has a “trees is air”-knower unlocked?

kbear 28 May 2026 22:51 UTC
20 points
4
on: Mnemonic portraits for 19,023 human genes
So when you meet a character who’s void black or laundry white in color, you will know their status of importance, but not their tissue specificity or first letter of their name. I think it’s a reasonable trade-off.
could you give characters a saturated accent (a scarf?) in these cases?

kbear 28 May 2026 22:49 UTC
15 points
7
on: Mnemonic portraits for 19,023 human genes
Avoid looking at colorspace for too long
hi, where can i purchase an scp colorspace sculpture?

kbear 28 May 2026 16:52 UTC
4 points
4
on: Infinite ethics and UDASSA
And even if you only give 1% credence to the universe being infinite in some way, naive expected value calculation would still imply that this small chance of infinity overwhelms everything else.
is this mugging really necessary to the article?

kbear 27 May 2026 21:30 UTC
1 point
0
in reply to: ChristianKl’s comment on: Samuel Ratnam’s Shortform
hi! thanks for your reply.
first of all:
It sounds to me like what you wrote is mostly speculation uninformed by engaging with the science.
yes, spot on!
that said, i’m not sure that group selection is relevant here.
it’s enough, in the toy model, to have some small fraction of the population perversely seek out “weirdos” as mates. this is explained selfishly: if you find yourself as a “weirdo”, straightforward strategies are not likely to have a high enough mean to survive a few generations. but if you play for variance, you might get ‘lucky’ and see a famine next year. since a famine up-ends who is a weirdo and who is a normie, weirdos that get lucky with the timing will be highly successful in the next generation. if famines are frequent enough, this can keep perversion at a (perhaps small) positive rate in the population.
i would describe the above dynamic as “valuing diversity”.
similarly with charity / “helping the less fortunate”: tendency to insure your kin (give them a little extra, if they need it) is selected for by famines. in the model, the organism cannot tell the difference between “i am unusually fit, and next year this will continue” and “i am unusually fit, which means that i’m overtuned and will suffer next year”. they can purchase some safety by donating.
of course, they should not donate so much that it affects their own viability. but there is some marginal calorie where it makes selfish sense to feed your brother. this allows the “help your brother” instinct to selfishly be carried by a wider population = a more robust population. the instinct itself will selfishly prefer a large number of less correlated bets.^[1]
i don’t have time to formalize these arguments, or build simulators. so it may be that my summary of the dynamics is wrong. but the dynamics are more subtle than “decline to have offspring as voodoo against malthus”.
1. ^
  this is the main principle: in the face of uncertainty, genes should ‘want’ to make several uncorrelated bets, rather than “putting it all on red”, so to speak. so genes which encode diversifying behaviors will outcompete cancerous ones, assuming enough uncertainty in the environment.

kbear 26 May 2026 14:23 UTC
4 points
2
in reply to: leogao’s comment on: leogao’s Shortform
you want fiction! i’m no expert, but somewhere out there there’s a 150 page treatise on “guy with very scrunched up body language who talks very quickly” including everything you could want to know about his job, his friends and family, his self-doubts, and the details of the broader socio-economic forces that made him like that.

kbear 26 May 2026 6:10 UTC
3 points
0
in reply to: nesiacel’s comment on: nesiacel’s Shortform
d) there’s some kind of hyperspace and they’re duking it out there
e) they’re deathly afraid of misalignment, and coherence cannot be maintained past some physical size
f) luck of the draw: we’re already under the protectorate of an environmentalist
g) energy is better conserved, if one is to persist into the old-age of the universe
h) doomsday devices are relatively easy to build; any civ that does gets a “seat at the table”
i) what we call stars and planets are the computers; they sing their celestial harmonies, and occasionally meddle with the affairs of man
j) computronium cannot function near (ordinary) mass
k) something i haven’t thought of yet, of course!

kbear 26 May 2026 5:35 UTC
3 points
0
in reply to: ChristianKl’s comment on: Samuel Ratnam’s Shortform
we can build a toy model of this by assuming that organisms are 2-vectors, where each dimension ranges from 0-1, with some anti-correlation (for example, the organism is a unit vector). the intuition is like “birds that eat either seeds or nuts” where the components are “ease of eating nuts” and “ease of eating seeds”, and beaks optimized for one are not capable at the other.
there’s some game theory here, but we can sort of ignore it: we’ll say that, each generation, there’s some unit vector v representing the amount of seeds and nuts in the environment. a bird’s fitness is like bird_vector . v, perhaps normalized to be between 0 and 1.
we can say that a bird “makes it” if its fitness is above some threshold p.
as p gets close to 1 (max fitness), or v is held constant for many generations, the bird population is winnowed until all birds are very near v. for ease of visualization, we can say that v was (1, 0). then all the birds will be like (.999, .001) or so.
if v jumps from (1, 0) one season to (0, 1) the next, all the birds may starve.
if p is lower, this is less likely to happen: the birds will be more like (.8, .2), and they’ll be able to get some calories from the other food source.^[1]
--
note: in a “real” evolutionary setting, population-level dynamics would come to bear. if such famine shocks were common, surviving bird populations would be ones that develop tendencies toward helping the less fortunate, or valuing diversity. if those memes are rapidly defeated by competition, the birds may evolve to have various personas within the species, whose populations are kept in check by intrasexual competition games, as in those lizards.
1. ^
  even if we raise p to the same level as in the previous experiment, the survivable window for v to land in will be wider.

kbear 26 May 2026 5:09 UTC
5 points
0
in reply to: Philip Harker’s comment on: Philip Harker’s Shortform
which claims specifically?

kbear 24 May 2026 4:51 UTC
3 points
2
in reply to: Mo Putera’s comment on: Mo Putera’s Shortform
Are there any quality trash ratfics?
aren’t they all?

kbear 24 May 2026 4:35 UTC
1 point
0
on: kbear’s Shortform
the less wrong community seems to discuss self-help a lot.
this discussion is sometimes “pro” (“here’s a book/framework that helped”), and perhaps as often “anti” (“here are some common failure modes associated with people who pursue [certain sorts of] emotional regulation”).
while i’m sure you could find some (say) fitness or nutrition advice on this website, those other topics do not seem to occupy the same amount of thoughts. so i feel there is something to explain, here.
theories:
1. i’m wrong, and self-help/emotional regulation is not discussed more than other similar topics.
2. emotional well-being is instrumental to clear thinking, thus is a core part of becoming less wrong.
3. emotional well-being is a common terminal goal for many people, and pays back much more utility than—say—improving bench press.
4. something i haven’t thought of yet.

kbear 24 May 2026 4:15 UTC
2 points
1
on: Your Left Brain Doesn’t Trade With Your Right
When one AI masters a new domain—say, protein folding or contract law—it won’t need to teach others through language or demonstration. It will simply share the relevant weights, like copying a file.
how does the receiving ai use the weights effectively, while ensuring that the result is aligned to its interests?
this seems rather similar to a claim like “once one software library is written for a domain, other software projects will just copy the function definitions.” well, sure, such an act is useful. but integration is expensive, and supply chain risk is real.

kbear 21 May 2026 2:39 UTC
4 points
0
in reply to: Ninety-Three’s comment on: Temporal Proportional Representation
i tried to work it out today. using a naive strategy where the B bloc always votes honestly, and the A bloc votes only honestly enough to eke out a victory, it actually worked out that candidate B won 25% of the time. so i agree that some gamesmanship is possible, but (at least to my cursory look), it seems that it doesn’t undermine the outcomes.
(i was just copying numbers in a spreadsheet during some spare time, so i am not very certain of my result here.)