A1987dM

Karma: 15,364

A1987dM 7 Aug 2026 10:08 UTC
2 points
0
on: Social Dark Matter
Likewise, I think the prevalence of supernumerary nipples in humans is probably severely underestimates because probably the vast majority of people who have them mistake them for moles

A1987dM 6 Aug 2026 10:02 UTC
4 points
0
on: SSA rejects anthropic shadow, too
You can’t use “we survived the cold war without nuclear war” as evidence of anything. Because of the anthropic principle, we could have blown up the human race in the 1960′s in 99% of all possible worlds and you’d still be born in one where we didn’t.
FWIW the main reason you can’t is that odds are nuclear war wouldn’t have “blown up the human race”—there’s plenty of ways it could have had several billion survivors. So, if it had been near-inevitable, that’s the kind of world where we’d expect to be in, whether you take anthropics into account or not. By that standard, you wouldn’t be able to use the Green Revolution as evidence of anything either (it probably saved roughly as many people as the lack of a nuclear war during the cold war did)!

A1987dM 6 Aug 2026 9:56 UTC
2 points
0
on: The Anthropic Trilemma

The first horn of the anthropic trilemma is to confess that there are simple algorithms whereby you can, indetectably to anyone but yourself, exert the subjective equivalent of psychic powers

Well, provided it is “simple” to make a trillion copies of yourself indetectably to anyone but yourself

A1987dM 6 Aug 2026 9:49 UTC
2 points
0
on: Prediction Markets Have an Anthropic Bias to Deal With
The wordplay is unintended, but awesome.
See also: the only time I literally laughed out loud by chatting with an LLM https://aistudio.google.com/app/prompts?state=%7B%22ids%22:%5B%221pTWi3K-_qiI67to1Y2AAwEUunqKxYMJf%22%5D,%22action%22:%22open%22,%22userId%22:%22116669250426639305492%22,%22resourceKeys%22:%7B%7D%7D&usp=sharing

A1987dM 14 Jul 2026 8:26 UTC
2 points
0
in reply to: Eugine_Nier’s comment on: AI box: AI has one shot at avoiding destruction—what might it say?
From the context, something like “WTH? WTH? OMG!” (a demon shitting bricks upon noticing that Dante was there with his body rather than just his soul)

A1987dM 10 Jul 2026 10:36 UTC
2 points
0
on: VDT: a solution to decision theory
That’s called virtue ethics, duh

A1987dM 10 Jul 2026 0:30 UTC
6 points
0
on: How big is the Sun? How could you figure it out?
Commas in LaTeX math need to be surrounded by curly braces otherwise thin spaces get automatically added after them

A1987dM 29 Jun 2026 10:48 UTC
2 points
0
on: The worthlessness of vitamin D is mildly exaggerated
https://xkcd.com/2071/

(I mean, I am aware that the “official” RDAs are as low as 600 IU/day, but my impression was that anyone who’s ever actually looked into this at all took them about as seriously as the WHO’s claim of COVID not being airborne, and would have taken any loud positive statement that yes, you can get less than 1000 IU of vitamin D per day year round with no meaningful drawbacks as meta-contrarianism rather than as lip service to the mainstream, had I ever encountered one before.)

A1987dM 9 Jun 2026 10:41 UTC
4 points
0
on: How Far Apart Does a Model Think Its Tokens Are?
The Chinese for AI starts with 人工 and that’s not a deliberate visual pun? That’s it, I’mma quit everything and become a kabbalist.

A1987dM 29 May 2026 12:34 UTC
6 points
2
in reply to: lilkim2025’s comment on: Claude… doesn’t know who you are?
“Will not tell any random official of an authoritarian country but will tell Pliny the Liberator” does sound like the sweet spot for that kind of things to me.

A1987dM 10 May 2026 8:18 UTC
2 points
0
in reply to: DanB’s comment on: It’s nice of you to worry about me, but I really do have a life

Examples: recreational drug use, single parenting, pursuing a career in the arts, dropping out of school, polyamory, etc

One of these thing is not like the others, one of these things doesn’t belong.

(I can see the appeal of a norm that no-one uses recreational drugs whether or not they personally would be able to handle it; but wouldn’t a norm that no-one pursues a career in the arts imply that no more art would be made, except in people’s spare time?)

A1987dM 22 Apr 2026 10:14 UTC
2 points
0
on: Automating philosophy if Timothy Williamson is correct
Parthood itself is nowhere near as obviously-to-me distinct from our concept of it as birds themselves are from our concept of them… Is it just me?

A1987dM 21 Apr 2026 10:17 UTC
2 points
1
in reply to: J Bostock’s comment on: If a room feels off the lighting is probably too “spiky” or too blue
I put “energy saving” in scare quotes because they’re in large party energy saving by just being dark.
Huh, no, a 13 W incandescent light bulb would be very dim

A1987dM 19 Apr 2026 6:54 UTC
2 points
0
in reply to: philip_b’s comment on: There are only four skills: design, technical, management and physical
By “cannot … any” do you mean “there are no instruments you can” or “there are instruments you cannot”? The former seems completely wrong to me, you definitely can become an expert in playing the kazoo or the triangle in 6 months. (Even if you mean the latter, I think there are very few instruments you can’t become an expert in in 6 months, if you have no other job and no children and can spend 80 hours a week on it.)

A1987dM 19 Apr 2026 1:52 UTC
2 points
0
in reply to: gbear605’s comment on: Why So Many Cookie Banners?
Ditto. Doing that for a few days would be borderline okay, but a few times I found stuff in shopping carts I had put there several months earlier, and didn’t immediately even remember about!
Automatically emptying the cart at the end of the browser session sounds a bit overzealous, but 7 days would be a reasonable rule of thumb.

A1987dM 18 Apr 2026 11:24 UTC
2 points
0
on: Orders of magnitude: use semitones, not decibels
Well, I’m a former (amateur) musician and I still find decibels perfectly appropriate—remembering that 10^0.3 ≈ 2 automatically gives you 10^0.6 ≈ 4, 10^0.9 ≈ 8, 10^1.2 ≈ 16 i.e. 10^0.2 ≈ 1.6, 10^0.5 ≈ 3.2 and going the other way 10^-0.3 ≈ 0.5 i.e. 10^0.7 ≈ 5, 10^0.4 ≈ 2.5 and so on. If those aren’t precise enough, sqrt(10) ≈ 3.16 also gives you 10^0.2 ≈ 1.58 and 10^0.8 ≈ 6.32.

A1987dM 18 Apr 2026 11:18 UTC
3 points
1
on: LLMs will soon disrupt algorithmic media feeds
What do you mean “will soon”, some of them e.g. Facebook are already enshittified beyond all repair

A1987dM 18 Apr 2026 8:52 UTC
2 points
0
in reply to: joseph_c’s comment on: joseph_c’s Shortform
Have you experimented with using languages other than Japanese for that?
(longer cycles e.g. English→Mandarin→Arabic→English might also be interesting to try, in principle, though I’d guess that in the current incarnation of Google Translate Mandarin→Arabic is just Mandarin→English→Arabic in a trenchcoat anyway)

A1987dM 18 Apr 2026 8:25 UTC
2 points
0
in reply to: Thane Ruthenis’s comment on: Thane Ruthenis’s Shortform
I agree that it would probably be a disaster but I can easily imagine someone (probably not Anthropic themselves, but probably someone like xAI or Meta, possibly even OpenAI or Google) not seeing it. It’s not any more obvious than it was that what SBF was doing would backfire, and yet he did that anyway...

A1987dM 18 Apr 2026 8:15 UTC
4 points
2
in reply to: Nick_Tarleton’s comment on: Thane Ruthenis’s Shortform
Eh. The problems with some strategy don’t just have to be obvious to you for someone to refrain from it, they must be obvious to them. (Case in point: SBF)
If someone like xAI or Meta had a Mythos-level model, I could totally imagine them doing something substantially different, much worse for the world than what Anthropic did even if I could easily see it’d probably be worse for them themselves too (except insofar as xAI or Meta making a Mythos-level model would be evidence for them being less unreasonable than they appear, so interpret my hypothetical as “if Lucifer existed and handed someone like xAI or Meta a Mythos-level model).