tlevin

Karma: 1,068

(Posting in a personal capacity unless stated otherwise.) I work at Coefficient Giving with a focus on helping humanity better navigate transformative AI, especially through US public policy. Formerly co-founder of the Cambridge Boston Alignment Initiative, which supports AI alignment/safety research and outreach programs at Harvard, MIT, and beyond, co-president of Harvard EA, Director of Governance Programs at the Harvard AI Safety Team and MIT AI Alignment, and occasional AI governance researcher.

Not to be confused with the user formerly known as trevor1.

tlevin 21 Nov 2025 21:18 UTC
20 points
9
in reply to: J Bostock’s comment on: Jemist’s Shortform
Other commenters have said most of what I was going to say, but a few other points in defense:
- On it sounding bad, I think time will tell. We are biased towards liking familiar stimuli.
- Coefficient Giving doesn’t tell you what we do, but neither did Open Philanthropy. And fwiw, neither does Good Ventures, IMO—or many nonprofits, e.g. Lightcone Infrastructure, Redwood Research, the Red Cross. Agreed that GiveWell is a very descriptive and good name though.
- Setting aside whether “coefficient” is a weird word, I don’t think having an unusual word in your name “burns weirdness points” in a costly way. Take some of the world’s biggest companies—Nvidia (“invidious,”) Google (“googol”), Meta—these aren’t especially common words, but it seems to have worked out for them.
- The emphasis of “coefficient” is on the “fish.” So it’s not an “awkward four-two,” it’s three sets of two, which seems melodic enough (cf. 80,000 Hours, Lightcone Infrastructure, European Union, Make-A-Wish Foundation, etc).
- On the no possible shortening, again, time will tell, but my money is on “CG,” which seems fine.

tlevin 6 Nov 2025 2:19 UTC
2 points
0
on: LLM-generated text is not testimony
I agree with some direction of this, but it seems to massively depend on the process by which the LLM text has reached your eyes.
At one extreme, a bot on social media, given some basic prompt and programmed to reply to random tweets, has basically zero content about the “mental elements” behind it, as you put it.
On the other, if someone writes “I asked an LLM to summarize this document, and upon closely reviewing it, I think it did a great job,” this has lots of content about a human’s mental elements. The human’s caption is obviously testimony, but the quoted LLM text also seems pretty much like testimony to me.
(There are plenty of intermediate cases, e.g. someone writes “I asked an LLM to summarize this document, which I personally skimmed, and it seems roughly right to me but caveat lector.”)

tlevin 6 Nov 2025 2:02 UTC
17 points
6
on: Heroic Responsibility
The entire thing about heroic responsibility, as I understand it, and as is apparent in the original HPMOR context, is that feeling responsibility for outcomes specifically when social expectations do not hold you responsible for them. I think the type of responsibility that the owner of a car ownership has for the profitability of said car dealership is called “responsibility.”

tlevin 21 Oct 2025 4:00 UTC
5 points
1
in reply to: Max H’s comment on: Considerations around career costs of political donations
I still think that the precise kind of optics considerations described and recommended in this post (and other EA-ish circles) are subtly but importantly different from what those staffers are doing.
It’s true that LessWrong readers would be doing a subtly but importantly different thing from what the staffers are doing. But the way that it’s different is that Congressional staffers, of all political persuasions, are much more intuitively and automatically doing these kinds of considerations because they’re pursuing careers in policy and politics in DC, whereas LessWrong readers tend to be technical people largely in the Bay Area who might someday later consider a career in policy and politics, and therefore they need to have these considerations explicitly laid out, as would anyone who’s considering a career pivot into an industry with very different norms.

tlevin 21 Oct 2025 0:43 UTC
7 points
6
in reply to: Max H’s comment on: Considerations around career costs of political donations
No. I’m more saying that the act of carefully weighing up career capital / PR considerations, and then not donating to a democrat based on a cost-benefit analysis of those considerations, feels to me like very stereotypical democrat / blue-tribe behavior.
Strongly disagree with the implication that Republicans/conservatives don’t carefully weigh up career capital and PR considerations when making decisions like this! The vast majority of elected Republicans and even more of their staff are comparably strategic in this regard as their Democratic counterparts. Of course, the exceptions are much higher-profile, which I think could be leading to an availability bias. (Again, notably Hanania is not employed in the government.)
And further, that some people could have a visceral negative reaction to that kind of PR sensitivity more so than the donations themselves.
“Some people,” sure. Federal government hiring managers? No.
Sure, but if he were the kind of person who would do that, he probably would not have gotten as popular as he is in the first place.
I mean, depends on if your goal is serving in the government or becoming a widely read Substacker.
And even then I’m not sure it’s true; many, many media figures with huge followings on both sides of the aisle are hardcore partisans. See for example (most of) the hosting lineups of MSNBC and Fox. LessWrong is an extreme outlier in how much readers intentionally consume heterodox and disagreeable content; the vast majority of political media consumers trust and prefer to listen to their co-partisans.

tlevin 20 Oct 2025 18:13 UTC
7 points
−4
in reply to: Max H’s comment on: Considerations around career costs of political donations
Is the idea that Hanania is evidence that being very public about your contrarian opinions is helpful for policy influence? If so, that seems wrong:
- As you said, Hanania would almost surely not get appointed to an actual position by either administration.
- To the extent that he has influence/reach on the right, I doubt he has more influence/reach because he pivoted to strong criticisms of Trump and the MAGA-sphere. I would rather guess that this pivot has been really costly to his influence on the right, and if he had self-censored, he’d be more influential.
- Also, even if this were the case, it seems unreliable to update more on the evidence from one unusual individual than the many cases (as noted in this post) of people losing out on jobs because of their donations and publicly expressed opinions.
That’s not to say that nobody should be doing the loud contrarian approach; the world would be worse if everyone were self-censoring to the degree incentivized by DC policy careers. But I think people should be clear-eyed about the costs and benefits.

tlevin 20 Oct 2025 18:04 UTC
2 points
0
in reply to: Annabelle’s comment on: tlevin’s Shortform
Interesting. Yeah rather than “entirely driven” I guess I should say: it seems like the direct effects of suppressing information probably usually outweigh the second-order Streisand Effect, and the exceptions are more salient than the non-exceptions due to survivorship bias?

tlevin 20 Oct 2025 15:19 UTC
2 points
2
in reply to: robo’s comment on: tlevin’s Shortform
Maybe we should distinguish between the “Weak Streisand Effect” (in some cases, the act of attempting to suppress information adds to the virality of a story) and the “Strong Streisand Effect” (attempting to suppress information increases the virality of a story in expectation). WSE seems definitely true, SSE seems pretty unlikely on average, though depends massively on the details.

tlevin 20 Oct 2025 6:00 UTC
9 points
0
on: tlevin’s Shortform
Hot take: the Streisand Effect might be entirely driven by survivorship bias. Probably happens all the time that people delete stuff and then the public never hears about it, but the rare exceptions are sufficiently ironic that they get a whole effect named after them!

tlevin 1 Oct 2025 23:14 UTC
1 point
0
in reply to: Vladimir_Nesov’s comment on: What SB 53, California’s new AI law, does
Makes sense! Yeah I can see that “that would be a problem” can easily be read as saying I don’t think this incentive effect even exists in this case; as you’re now saying, I meant “that would make the provision a problem, i.e. net-negative.” I think conditional on having to say sufficiently detailed things about catastrophic risk (which I think 53 probably does require, but we’ll see how it’s implemented), the penalty for bad-faith materially false statements is net positive.

tlevin 30 Sep 2025 0:36 UTC
10 points
2
in reply to: Vladimir_Nesov’s comment on: What SB 53, California’s new AI law, does
If that were the only provision of the bill, then yes, that would be a problem, but the bill requires them to publish summaries of (1) their up-to-date framework for assessing and mitigating catastrophic risks and (2) their assessments of catastrophic risks for specific models.

tlevin 29 Sep 2025 23:47 UTC
3 points
0
in reply to: Vladimir_Nesov’s comment on: What SB 53, California’s new AI law, does
There’s an exception for statements “made in good faith and reasonable under the circumstances”; I would guess it’s pretty hard to prove the contrary in court?

What SB 53, California’s new AI law, does

tlevin29 Sep 2025 23:29 UTC

106 points

12 comments4 min readLW link

tlevin 20 Aug 2025 1:58 UTC
3 points
2
on: Models Don’t “Get Reward”
Coming back >2.5 years later to say this is among the most helpful pieces of AI writing I’ve ever read—I remember it being super clarifying at the time, and I still link people to it, cite it in conversation, and use similar analogies. (Even though I now also caveat it with ”...but also maybe really sophisticated agents will actively seek it, because the training environment might reward it, and maybe they’ll ‘experience’ something like fear/pain/etc for things correlated with negative reward, if they experience things...”) Thank you for writing it up!!

tlevin 3 Aug 2025 0:26 UTC
4 points
2
on: The Observer Effect for belief measurement
Half-baked thought: this also creates another problem for using prediction markets to make decisions, as bettors should update on the question being asked in the first place (though this could be mitigated a lot by e.g. the existence of the problem being required by law or typical practice etc, as in the original futarchy proposal).

tlevin 12 Jul 2025 1:43 UTC
2 points
−1
in reply to: habryka’s comment on: tlevin’s Shortform
Alas, was hoping the smiley at the end would give it away...

tlevin 11 Jul 2025 18:21 UTC
3 points
0
in reply to: johnswentworth’s comment on: tlevin’s Shortform
Interesting point! Do you have any ideas for how I could read more about potential problems in air conditioner ratings (by Wirecutter and others)? Like, way more? :)

tlevin 10 Jul 2025 22:43 UTC
25 points
2
on: tlevin’s Shortform
Prime Day (now not just an amazon thing?) ends tomorrow, so I scanned Wirecutter’s Prime Day page for plausibly-actually-life-improving purchases so you didn’t have to (plus a couple others I found along the way; excludes tons of areas that I’m not familiar with, like women’s clothing or parenting):
Seem especially good to me:
- Their “budget pick” for best office chair $60 off
- Whoop sleep tracker $40 off
- Their top pick for portable computer monitor $33 off (I personally endorse this in particular)
- Their top pick for CO2 (and humidity) monitor $31 off
- Crest whitening strips $14 off (I personally endorse this in particular)
- 3-pack of their top pick for umbrellas, $12 off
- Their top pick for sleep mask $12 off
- Their top pick for electric toothbrush $10 off
- 6-pair pack of good and super-affordable socks $4 off (I personally endorse this in particular; see my previous enthusiasm for bulk sock-buying in general and these in particular here)
Could also be good:
- A top hybrid mattress $520 off
- A top inner-spring mattress pick $400 off
- Their top pick for large carry-on $59 off
- Their “budget pick” for weighted blanket $55 off
- Their top pick for best air conditioner $50 off
- Their top pick for laptop backpacks $45 off
- 3-in-1 travel charging station $24 off
- Top pick for face sunscreen $16 off
- Uber/UberEats gift card $15 off (basically free $15 if you ever use Uber or UberEats; see my previous enthusiasm for these gift cards as sold at 80% face value at Costco here)
- 4-pack of Apple AirTags $15 off
- Their “budget pick” for bidets $12 off
- Their top pick for towels $10 off
- Good and super-affordable portable Bluetooth speaker $8 off
- Top pick for portable mosquito repellent $7 off
- 3-pack of good and super-affordable sunglasses $4 off
- Many, many mechanical keyboards, headphones, smart watches, wifi extenders, chargers/cables, and gaming mice if you’re in the market for any of those

[linkpost] One Year in DC

tlevin19 May 2025 19:46 UTC

112 points

6 comments1 min readLW link

(www.greentape.pub)

tlevin 12 May 2025 4:05 UTC
50 points
14
on: tlevin’s Shortform
Found this passage from Schelling’s Strategy of Conflict to be useful in understanding how norms react, and how they should react, to behavior that is not necessarily harmful but is often associated with harm, cf. “decoupling.” Emphasis mine:
It is essential, therefore, for maximum credibility, to leave as little room as possible for judgment or discretion in carrying out the threat. If one is committed to punish a certain type of behavior when it reaches certain limits, but the limits are not carefully and objectively defined, the party threatened will realize that when the time comes to decide whether the threat must be enforced or not, his interest and that of the threatening party will coincide in an attempt to avoid the mutually unpleasant consequences.
In order to make a threat precise, so that its terms are visible both to the threatened party and to any third parties whose reaction to the whole affair is of value to the adversaries, it may be necessary to introduce some arbitrary elements. The threat must involve overt acts rather than intentions; it must be attached to the visible deeds, not invisible ones; it may have to attach itself to certain ancillary actions that are of no consequence in themselves to the threatening party. It may, for example, have to put a penalty on the carrying of weapons rather than their use; on suspicious behavior rather than observed misdemeanors; on proximity to a crime rather than the crime itself. And, finally, the act of punishment must be one whose effect or influence is clearly discernible.
In order that one be able to pledge his reputation behind a threat, there must be continuity between the present and subsequent issues that will arise. This need for continuity suggests a means of making the original threat more effective; if it can be decomposed into a series of consecutive smaller threats, there is an opportunity to demonstrate on the first few transgressions that the threat will be carried out on the rest. Even the first few become more plausible, since there is a more obvious incentive to fulfill them as a “lesson.” This principle is perhaps most relevant to acts that are inherently a matter of degree. In foreign aid programs the overt act of terminating assistance may be so obviously painful to both sides as not to be taken seriously by the recipient, but if each small misuse of funds is to be accompanied by a small reduction in assistance, never so large as to leave the recipient helpless nor to provoke a diplomatic breach, the willingness to carry it out will receive more credulity; or if it does not at first, a few lessons may be persuasive without too much damage.
The threatening party may not, of course, be able to divide the act into steps. (Both the act to be deterred and the punishment must be divisible.) But the principle at least suggests the un-wisdom of defining aggression, or transgression, in terms of some critical degree or amount that will be deemed intolerable. When the act to be deterred is inherently a sequence of steps whose cumulative effect is what matters, a threat geared to the increments may be more credible than one that must be carried out either all at once or not at all when some particular point has been reached. It may even be impossible to define a “critical point” with sufficient clarity to be persuasive. To make the threatened acts divisible, the acts themselves may have to be modified. Parts of an act that cannot be decomposed may have to be left out; ancillary acts that go with the event, though of no interest in themselves, may be objects to which a threat can effectively be attached. For example, actions that are only preparatory to the main act, and by themselves do no damage, may be susceptible of chronological division and thus be effective objects of the threat. The man who would kick a dog should be threatened with modest punishment for each step toward the dog, even though his proximity is of no interest in itself.

tlevin

What SB 53, Cal­ifor­nia’s new AI law, does

[linkpost] One Year in DC

What SB 53, California’s new AI law, does