It’s helpful, I think, to do so; I don’t know if the mods have declared it an explicit norm, but I do know some folks who’ve had negative experiences when folks thought they were trying to pull a fast one.
(And just to reaffirm—don’t think Max was doing that, that’s why I wrote “misconstrued” above—just trying to help everyone keep their trigger fingers nice and relaxed here)
davekasten
Karma: 2,103
Max, it might be useful here to disclose your relationship with the same employer as Alex.
(I do agree with some parts of this post and think they are good, disagree with other parts that are weak arguments, but think LessWrong has an intense allergy against anything that could be misconstrued as astroturfing karma. I know you and think you’re a nice dude, so it’s to everyone’s benefit to flag this.)
Perhaps a helpful synthesis: DC and Berkeley spend some of their weirdness points very similarly (e.g., being nerds about esoteric worldshaking topics) and some very differently (e.g., formality of attire, relationship styles)[1]
- ^
Note that a person from Berkeley would not have used this phrasing for either of these points.
- ^
What’s the current state of play on “good fairly quiet indoor air purifiers”? Like, is there a better model than “whatever Lighthaven has” or is that the best call, and if so, what is it called on Amazon?
I think this mostly tracks with how many military professionals talk about this stuff. It’s beneficial for all sides to have rights of surrender in the special case of surrender, even though it’s beneficial for all sides to be clever generally in war. So, we absolutely ban the act of falsely surrendering under a white flag (to, e.g., enable an ambush) as the war crime of perfidy, even though we allow many other “ruses de guerre” that involve deceit in other contexts.
Basic obvious point for some audiences here: if you’re not using LLMs for helping you with basic scutwork tasks, you’re losing leverage that you shouldn’t be on AI policy advocacy. Claude Code makes my job faster and more fun.
This shouldn’t be news, but sometimes people need a self-improvement tip.
(If you have an ethical objection to using them at all, obviously this wouldn’t apply to you.)
One thing I find myself doing increasingly in the AI policy community is connecting people to offices in DC. This is good, but I worry that it leads us to have path-dependency on who people like me know and the logic that we have about where to start, rather than prospecting for new offices that we don’t know as well. It’s plausible that everyone doing this should run rand(535) against a list of all the Senators and Representatives a few times a month and go meet a truly random office and see what happens.
I’d vote for it being a special subdomain or something. Or maybe it’s a post-only form, and then the outputs get shared privately to e.g., the company that made that particular model as well as researchers in private?
Thanks! If so, then I strongly object to the claim that good design generalists should be expected to do good policy within ~6 months.
Basic Q: am I a “design” or a “technical” person? Put another way, I think “go do policy” is either some blend of both skill trees, or possibly its own thing; there’s some very weird sense in which most people can’t make the kinds of coupled-with-just-the-right-degree-of-metaphorical-slippage work that policy analysis and writing entails.
(Many people can! Just see people who feel like excellent at either “design” or “technical” fail at this all the time.)
Hmm, Transformer reports: “That campaign was established in the UK at the end of last year by Harry Atkinson, a former story editor in the film industry, and Frieda Lurken, a co-founder of climate group Extinction Rebellion”.
see https://www.transformernews.ai/p/scream-if-you-want-to-move-slower-pause-ai-pull-the-plug
Does anyone know who runs Pull the Plug? Was there anyone at the event who represented themselves as a leader of Pull the Plug?
Their website has not a single organizer name on it, and I really dislike not knowing who runs any organization.
Who’s behind Pull the Plug? I don’t see any details about it on their website.
Idle musing: should we all be writing a Claude Constitution-esque set of posts about our hopes for how AIs help humans around the dangerous moments in takeoff, in hopes that this influences the models advise not only us, but people who are coming into these issues fresher than us?
(Yes, I know that from the outside I have exponentially less influence on model behavior than Anthropic, and for MIRI-like reasons maybe this doesn’t go well at all. But, you know, play to all of your outs.)
I would like, for obvious self-interested reasons, discussion of public policy ideas you think people should be pushing more on than they currently are.
It was a mild positive update for me that an Anthropic employee, in this case Zac, was able to clarify this.
I worry, a lot, that the true gloss on the American Way of War is, roughly the meme of “every Pacific naval encounter from late 1943 onward is like the IJN Golden Kirin, Glorious Harbinger of Eternal Imperial Dawn versus six identical copies of the USS We Built This Yesterday supplied by a ship that does nothing but make birthday cakes for the other ships.”
Or, put more generally, we show up in the 4th quarter with a shit ton of gratuitously over-the-top production of every possibly-vaguely-good idea, and manage to eke out a win. See, e.g., the Civil War, WW1, WW2 (as above), Korea (kind of, long story), the Gulf War (after we fucked up the pre-war diplomacy), and the post-Surge Iraq War.
”There is no pause, so we just gotta Get a Lot of Alignment Research Done Real REAL Fast” is plausibly the real world we end up in, and I think we should have more folks optimizing for it beyond Redwood (and Anthropic???), even as terrifying as it feels.
I also had this question and couldn’t find an obvious answer. I recognize that to some degree this might be proprietary, but this feels like a pretty obvious comms question. It doesn’t negatively impact my opinion of Zac or Drake if they’re unable to answer given their confidentiality obligations, but I would ask them to relay back internally that not indicating one way or another looks extremely weird from a comms perspective.
It’s like Ford announcing that they’ve added airbags to their car designs. Which designs, the ones available for purchase now, or future year models? Oh, you know, just...car designs.
This is awesome! It broadly aligns with my understanding of the situation, although it does miss some folks that are known to care a bunch about this from their public statements. Downloading the JSON to take a deeper look!
Strong upvoting the underlying post for Doing The Thing.
FWIW, I resonate with some parts of this post—e.g., that a near-ASI endgame is deeply unstable from a risk-of-war perspective.
But I also think it overall has a code smell of what you might call “politics as done by physicists” (https://xkcd.com/793/) : far too much confidence in the exhaustiveness of frameworks; far less respect for friction, uncertainty, and risk-averseness than exists in actual policymaking; and very little exploration of polities and societies as anything other than command-and-control actors.
And I note that, in general, politics as done by “physicists” has failed to work very well. (These are metaphorical “physicists” referencing the XKCD comic—actual literal physicists can become good politicians, if they learn the actual affordances of the societies they live in, however. Metaphorical “physicists” include things like the writings of Karl Marx, in addition to some literal physicists).