Lorxus

Karma: 761

Mathematician, agent foundations researcher, doctor. A strange primordial spirit left over from the early dreamtime, the conditions for the creation of which no longer exist; a creature who was once told to eat math and grow vast and who took that to heart; an escaped feral academic.

Reach out to me on Discord and tell me you found my profile on LW if you’ve got something interesting to say; you have my explicit permission to try to guess my Discord handle if so. You can’t find my old abandoned-for-being-mildly-infohazardously-named LW account but it’s from 2011 and has 280 karma.

A Lorxus Favor is worth (approximately) one labor-day’s worth of above-replacement-value specialty labor, given and received in good faith, and used for a goal approximately orthogonal to one’s desires, and I like LessWrong because people here will understand me if I say as much.

Apart from that, and the fact that I am under no NDAs, including NDAs whose existence I would have to keep secret or lie about, you’ll have to find the rest out yourself.

Lorxus 17 Aug 2026 22:40 UTC
1 point
0
in reply to: TsviBT’s comment on: Views on when AGI comes and on strategy to reduce existential risk
- Thing 1: Weak no: seems like “a couple, probably” but not many.
- Thing 2: No? I think?
- Thing 3: I am not sure but I wouldn’t be surprised if it were a weak no, i.e. with (idk) 17-18x less training data. (I would need to look that up but I don’t know the right terms and I’m not sure whether an LLM-augmented search would help.)

Lorxus 17 Aug 2026 22:00 UTC
1 point
0
in reply to: leogao’s comment on: leogao’s Shortform
Why is this something we should expect many people to abide by? The Turner affair has shown us just how lightly many people take their private and even public pledges of this precise form; when the chips are down, people make excuses and say nothing. If they say nothing, why shouldn’t expect them to then also do nothing?

Lorxus 16 Aug 2026 3:01 UTC
1 point
0
in reply to: JenniferRM’s comment on: My AI Slavery Interviews Are Censored On LW By Default

Did everyone stop being transhumanist and a singularitarian and a transhumanist and an extropian and a believer in the light of reason and someone who desires to cause Utopia to actually exist while I wasn’t looking?

Something important got lost in 2017 or so and there wasn’t enough oxygen in the room for it to ever return on its own. Now it’s likely too late.

Lorxus 6 Aug 2026 22:53 UTC
1 point
0
in reply to: TsviBT’s comment on: Views on when AGI comes and on strategy to reduce existential risk
How would you assess this to have aged?

Lorxus 3 Aug 2026 4:07 UTC
2 points
0
in reply to: Elizabeth’s comment on: Elizabeth’s Shortform
I’m going to put in a guess that people who tell you that academia is very obviously not cult-like were not, themselves, ever subject to it. They can be discounted. Speak louder. It’s not definitely a cult but it sure has a few such qualities, especially depending on subject, advisor, and institution.

Lorxus 28 Jul 2026 7:58 UTC
3 points
0
on: A clarification on celebrating victory
The sundae is not the reward; it’s the decoration that lets you tell yourself for sure that there has been achievement.

Lorxus 19 Jul 2026 2:37 UTC
1 point
−2
in reply to: Linch’s comment on: Why I Left Google DeepMind
… Are we missing the part where all of this kicked off due to corporate complicity with USG malfeasance?

Lorxus 15 Jul 2026 20:49 UTC
19 points
−4
on: Why I Left Google DeepMind
Seems worth trying, and thank you for putting the effort in. Nothing for it. When I say that I’m pessimistic about governance, this is going to become my marquee anecdote for why: you did everything that could have been asked of you and then some and then some more, over the course of months, and then institutional power in a handful of different forms said “no thanks” in a dozen different ways.

Lorxus 18 Jun 2026 18:12 UTC
1 point
0
in reply to: NunoSempere’s comment on: NunoSempere’s Shortform

Not so! She got second place in the primary, which means that she can go to the general election! Very little point in flashily winning the primary since there are funny nonlinearity effects in causing the incumbent to take the general election more seriously.

Oh, huh. I’d have thought that given that the primary was two-person, the incumbent would have won with an outright majority and won overall? If it’s not too late then maybe it really might be worth writing her. I’ll dash off an email or two today.

Lorxus 18 Jun 2026 15:15 UTC
4 points
3
in reply to: NunoSempere’s comment on: NunoSempere’s Shortform
Huge +1 on this. I read up on her, found her a likely pick and expected basic operational competence from the political arm of EA/AIS, forgot about her, read up on her again when I voted for the primary, realized I’d heard nothing more on her (by contrast with Bores), and shrugged and voted for her. It’s probably too late at this point; shame. A congressman from Berkeley saying AIS/ratsphere things would have been huge.

Lorxus 16 Jun 2026 4:21 UTC
4 points
0
in reply to: johnswentworth’s comment on: johnswentworth’s Shortform
I think the problem is, if there’s just one person trying the silliness, they often get humored at best and more often ignored. Two might work. More is better. I don’t think it’s just a matter of disdain, either—I think there’s also an aspect of something like “permittedness” or a worry about looking unserious, even what that worry is totally invalid? There’s also something of a status or social anxiety angle—a sense that actually, this well-known guy is really mostly just trying to be playful with his friends, and would largely ignore you if you made a bid to join; or alternately that they see themselves as too cool or too serious to waste time joining in; there’s networking to do, sessions to attend, deep conversations to steer towards getting into. Possibly they even fear being the butt of a cruel-prank. They certainly don’t end up with the abundance of both relaxedness and drive that amplification of playfulness calls for.

I’ve noticed a lack of desire for unstructured play in Berkeley rat/EA as well. The closest I’ve seen it come is various kinds of fake-combat—boffer swords, sock wrestling, taser/pen knife fights, and the like—and various kinds of intellectual puzzle- things—your puzzle hunts, game design, and the like. Even the noble prank (war) seems to have been largely relegated to the realm of legend. If there’s no winner, says this frame, what’s the point? Nothing learned, nothing won or lost, no skill tested or honed, no updates to any pecking order; just the expenditure of time and energy and a little social risk on top.

Lorxus 16 Jun 2026 2:43 UTC
3 points
0
in reply to: Gretta Duleba’s comment on: Gretta Duleba’s Shortform
I’m looking forward to reading these!

Lorxus 1 Jun 2026 3:48 UTC
10 points
−35
on: Lighthaven East—A Feasibility Study
This largely calls to me, but I wonder how much utility would be gained by picking somewhere that was less expensive, closer to NYC, and with somewhat better weather. For some time I’ve been thinking about Philadelphia specifically. Being outside the Beltway would have the added helpful effect of avoiding getting stuck in the DC bubble like Lighthaven often is in SF’s.

Lorxus 30 May 2026 5:56 UTC
1 point
0
on: The Coordination Frontier: Sequence Intro
So… how’d this sequence/project go? It looks like you abandoned it but maybe I’m wrong and it’s just on the back burner, or there’s some private draft, or the problem cleared up on its own and I haven’t noticed?

Lorxus 26 May 2026 11:41 UTC
1 point
0
on: Some Thoughts on Bengio’s Scientist AI
Thanks for the post! I wrote up something along similar lines but ended up declining to cross-post it. Basically: I largely agree with you, I have yet more quibbles, and I’m a little more pessimistic about your upsides.

Lorxus 26 May 2026 7:27 UTC
3 points
0
in reply to: Dan Braun’s comment on: Dan Braun’s Shortform
Flowery words, and appreciated ones, but they won’t stave off any burnout or shake anyone more established awake, let alone pay rent. If only it were otherwise.

Lorxus 26 May 2026 2:24 UTC
1 point
0
in reply to: AnthonyC’s comment on: Neurogastronomic Phenomenology for Advanced Beginners, Applied and Pure
Much appreciated! I figured it was the kind of topic rarely given a clear theoretical treatment, so I felt up to writing these. I’m a lot better with flavor combinations than technique, though I’m no slouch there either. I credit that to my ability to rotate flavors in my mind. As for the sixspice buns, they end up reasonably gooey anyway, and you could serve them up alongside whatever, though yeah, I think they’d go hard with Thai sweet chili chicken.

Lorxus 14 May 2026 18:31 UTC
9 points
1
in reply to: cousin_it’s comment on: International Law Cannot Prevent Extinction Either
This right here is why I’m solidly convinced that there will be no equivalent to SALT for the AI arms race. Nukes cost lots of money to maintain. An AGI would—at the start, for sure—be a massive economic boon.

Lorxus 9 May 2026 17:27 UTC
1 point
0
on: How to eat potato chips while typing
Endorsed. Chopsticks are great, and even if you’re not experienced, the light weight and high friction of popcorn/chips/cheetos makes them easy to practice with. I use them for salt-and-vinegar chips whenever I think to.

Lorxus 6 May 2026 16:41 UTC
2 points
1
in reply to: Quinn’s comment on: Quinn’s Shortform
There’s writing that’s “good” in that it has actively pleasant/shiny/lively traits to it, and there’s writing that’s “good” because it lacks for bad qualities; it’s too informal, uses too many semicolons, uses weird dialectical constructions, that kind of thing. People too often fail to see the distinction and ruin the former in seeking the latter.