LintzA

• LintzAJan 31, 2025, 3:39 PM

  16 points

  3 ∶ 0

  in reply to: Joe Rogero’s comment on: The Game Board has been Flipped: Now is a good time to re­think what you’re doing

  That’s my bad, I did say ‘automated’ and should have been ‘automatable’. Have now corrected to clarify

• LintzAJan 31, 2025, 3:36 PM

  5 points

  0 ∶ 0

  in reply to: Buck’s comment on: The Game Board has been Flipped: Now is a good time to re­think what you’re doing

  Do you have anything you recommend reading on that?

  I guess I see a lot of the value of people at labs happening around the time of AGI and in the period leading up to ASI (if we get there). At that point I expect things to be very locked down such that external researchers don’t really know what’s happening and have a tough time interacting with lab insiders. I thought this recent post from you kind of supported the claim that working inside the labs would be good? - i.e. surely 11 people on the inside is better than 10? (and 30 far far better)
  
  I do agree OS models help with all this and I guess it’s true that we kinda know the architecture and maybe internal models won’t diverge in any fundamental way from what’s available OS. To the extent OS keeps going warning shots do seem more likely—I guess it’ll be pretty decisive if the PRC lets Deepseek keep OSing their stuff (I kinda suspect not? But no idea really).
  
  I guess rather than concrete implications I should indicate these are more ‘updates given more internal deployment’ some of which are pushed back against by surprisingly capable OS models (maybe I’ll add some caveats)

The Game Board has been Flipped: Now is a good time to re­think what you’re doing

• LintzASep 3, 2024, 5:51 PM

  4 points

  1 ∶ 1

  in reply to: Eric Neyman’s comment on: The value of a vote in the 2024 pres­i­den­tial election

  I think it’s very reasonable to say that 2008 and 2012 were unusual. Obama is widely recognized as a generational political talent among those in Dem politics. People seem to look back on, especially 2008, as a game-changing election year with really impressive work by the Obama team. This could be rationalization of what were effectively normal margins of victory (assuming this model is correct) but I think it matches the comparative vibes pretty well at the time vs now.

  As for changes over the past 20+ years, I think it’s reasonable to say that there’s been fundamental shifts since the 90s:

  • Polarization has increased a lot

  • The analytical and moneyball nature of campaigns has increased by a ton. Campaigns now know far more about what’s happening on the ground, how much adversaries spend, and what works.

  • Trump is a highly unusual figure which seems likely to lead to some divergence

  • The internet & good targeting have become major things

  Agree that 5-10% probability isn’t cause for rejection of the hypothesis but given we’re working with 6 data points, I think it should be cause for suspicion. I wouldn’t put a ton of weight on this but 5% is at the level of statistical significance so it seems reasonable to tentatively reject that formulation of the model.

  Trump vs Biden favorability was +3 for Trump in 2020, Obama was +7 on McCain around election day (average likely >7 points in Sept/​Oct 2008). Kamala is +3 vs Trump today. So that’s some indication of when things are close. Couldn’t quickly find this for the 2000 election.

• LintzASep 3, 2024, 5:06 PM

  12 points

  3 ∶ 2

  on: The value of a vote in the 2024 pres­i­den­tial election

  I think this is all very reasonable and I have been working under the assumption of one votes in PA leading to a 1 in 2 million chance of flipping the election. That said, I think this might be too conservative, potentially by a lot (and maybe I need to update my estimate).

  Of the past 6 elections 3 were exceedingly close. Probably in the 95th percentile (for 2016 & 2020) and 99.99th percentile (for 2000) for models based off polling alone. For 2020 this was even the case when the popular vote for Biden was +8-10 points all year (so maybe that one would also have been a 99th percentile result?). Seems like if the model performs this badly it may be missing something crucial (or it’s just a coincidental series of outliers).
  
  I don’t really understand the underlying dynamics and don’t have a good guess as to what mechanisms might explain them. However, it seems to suggest that maybe extrapolating purely from polling data is insufficient and there’s some background processes that lead to much tighter elections than one might expect.

  Some incredibly rough guesses for mechanisms that could be at play here (I suspect these are mostly wrong but maybe have something to them):

  • Something something polarization, steady voting blocs for Rep & Dem aren’t shifting much year to year. This means we should expect similar margins this year as 2016 & 2020.

  • Some balancing out process where politicians are adjusting their platform, messaging, etc to react to their adversary and this ends up increasing how close elections get.

  • Maybe something where voters have local information on whether the person they don’t like is more likely to win and they then feel more motivated to vote? Turns out, in aggregate, this local information is pretty accurate and leads to tighter-than-expected elections.

  • Maybe political parties/​donors observe how much their adversary spends in a given state and are consistently able to spend to counteract their efforts. This maybe provides a balancing effect that tightens the race. This would have the unfortunate consequence that visible spending is much less effective—but maybe implies that smaller, more under-the-radar, projects are better.

• LintzAAug 28, 2024, 2:42 PM

  7 points

  2 ∶ 0

  in reply to: richard_ngo’s comment on: The case for con­tribut­ing to the 2024 US elec­tion with your time & money

  I think it’s plausible that Dems turned a blind eye to some of this and that led to a few thousand extra votes here and there. US elections (and elections in general) always have issues like this and AFAIK there’s no reason to believe they played any larger or more important role in 2020 than any other election. In fact, given the amount of highly-motivated scrutiny applied to the 2020 election, I suspect it was cleaner than most previous elections.

  Even had Trump received any credible evidence of unusual tampering (you’d think he’d have laid it out by now if he had), his actions were beyond the pale. His own Attorney General refused to recognize any signs of fraud. He tried to cajole anyone he could into not certifying the results in any state or district he could despite no real evidence of wrong-doing. His scheme to create alternate slates of electors was an out-and-out attempt at election fraud. There’s no world in which that was intended to be representative of ground-truth.

  This article spells out a bunch of Trump’s actions around the 2020 election. I’m curious what you think of it.

  To be fair (kinda) to Trump, I think he really may have thought the election was stolen. He seems extremely capable of deluding himself about things like that. E.g. he just said that, if Jesus were counting the vote, he would win California easily. My hot take is that having a president who is actively trying to delude himself and his followers into believing 2020 was stolen (and that 2024 will be stolen) is bad, that it displays a weakness of character & epistemics that should be disqualifying. It should, e.g., make us question his ability to act reasonably in a crisis situation or when presented with a complicated new risk like AI.

• LintzAAug 26, 2024, 3:41 PM

  22 points

  4 ∶ 0

  in reply to: richard_ngo’s comment on: The case for con­tribut­ing to the 2024 US elec­tion with your time & money

  Generally feels like it’s primarily talking to an audience who already agrees that Trump is bad, and just needs to be persuaded about how bad he is

  This is true to some extent. I did not write this thinking it would be ‘the EA case for Kamala’ in response to Hammond’s piece. I also was wary about adding length to an already too-long piece so didn’t go into detail on various counterpoints to Kamala.

  Is framed it not just as a case for Kamala, but as a case for action (which, I think, requires a significantly higher bar than just believing that it’d be better on net if Kamala won).

  I personally see Trump’s anti-democratic behavior and demonstrably bad values as very-nearly disqualifying on their own (similar to, e.g., Scott Aaronson’s case against Trump). That’s why I focus so much on likely damage to liberal democracy. In my view these are crucial enough considerations that I would require some strong and clearly positive data points in Trump’s favor to override his obvious flaws. I am not aware of clear and strong positives on Trump’s side, only some points which seem closer to ‘maybe he would do this good thing. He hasn’t talked about it, but it seems more likely he’d do it than that Harris would’.

  Except where business-as-usual decisions would affect catastrophic risk scenarios I think they generally wash out when compared to Trump’s flaws.

  Doesn’t address the biggest concerns with another Democrat administration (some of which I lay out here).

  I address a good chunk of those concerns here. Agree that I could have talked about this more (though again, the piece was already very long).

  And yet it has been heavily upvoted. Very disappointing lack of consistency here, which essentially demonstrates that the criticisms of the previous post, while framed as criticisms of the post itself, were actually about the side chosen.

  I don’t see why this follows from the above. The claim seems to be that the only reason that post could have been downvoted and this post upvoted is because of bias. You’ve argued that there’s some content I didn’t address, and that it’s written for a Harris-leaning audience, but haven’t put forward a critique of the positions put forward in the post. It also seems clear that I’ve, on each cause area at least, attempted to present both sides of the argument. I’m curious why you see it as inconsistent? People disagree on object-level politics – and many people on here seem to strongly disagree with you – but one side is generally right, on net. Two posts advocating for different sides of an issue shouldn’t be treated the same just because it’s about politics. Also, this post has received its fair share of criticism (e.g. Larks’ comment, which I thought was useful and led me to update the post).

  one of the most harmful things that can be done for AI safety is to heavily politicize it

  Agreed, I don’t want to politicize AI safety. I really hope that, should Trump be elected, he’ll have good advisors and make good decisions on AI policy. I suspect he won’t, but I really hope he does.

  Here’s my thoughts on why it seems fine to post this:

  • There’s been pro-Trump content on the Forum already but virtually no pro-Harris content AFAIK.

  • This post doesn’t show up on the front page because it’s politics (at least that’s my understanding, I didn’t see it there personally despite the upvotes).

  • We’re not spreading this publicly in any ways that non-EAs are likely to see.

  • This kind of post seems like a drop in the bucket. Lots of EAs identify as Democrat, many as Republican. Having debates about who to elect seems perfectly reasonable. I’m glad there’s not a ton of posts like this on the forum, if there were I probably wouldn’t have written it. Adding one on the margin doesn’t seem like a big deal to me.

  • This post talks about AI safety fairly little and, what content there is, is mainly in the appendix.

  By default, we should expect that a lot more people will end up getting on the AI safety train over time; the main blocker to that is if they’re so entrenched in their positions that they fail to update even in the face of overwhelming evidence. We’re already heading towards entrenchment; efforts like this will make it worse.

  Not sure I fully understand this point but will attempt to answer. Again, I do not think this post or any of my other efforts are contributing meaningfully to politicizing/​polarizing AI safety or “entrenching” positions about it, and I really hope a Trump administration will make good decisions on AI policy in case he is elected (and I’ll support efforts to this end). However, this is fully compatible with believing that a Harris administration would be far better – or far less bad – in expectation for AI policy. I give several important reasons for believing this in the post, e.g.: Trump has vowed to repeal Biden’s executive order on AI on day 1; Trump generally favors non-regulation and plans to abolish various agencies (Vance favors tech non-regulation in particular); and the demographics and professions that make up the AI safety/​governance movement seem to have a far better chance at getting close to and influencing a Democratic administration than a MAGA administration, for several reasons.

• LintzAAug 25, 2024, 6:51 PM

  44 points

  11 ∶ 2

  in reply to: richard_ngo’s comment on: The case for con­tribut­ing to the 2024 US elec­tion with your time & money

  [This is part 1, I will get to foreign policy and AI-specific questions hopefully soon]

  I don’t think it’s fair to put an attempt to overthrow an election on par with biased media coverage (seems like both sides do this about equally, maybe conservative media is worse?) or dumping on opposition candidates (not great but also typical of both parties for many decades AFAIK). Scott Aaronson lays out some general concerns well here.

  Trump incited a violent coup/​insurrection attempt to prevent the 2020 election from being certified as well as other extremely norm-violating and likely illegal behavior to overturn the 2020 election (see long list here). The Republican party and supporting infrastructure (committees, media, influencers, fans, etc.) have since agreed to support his re-election attempt, punished members of the party for holding Trump accountable, and are touting January 6th insurrectionists as heroes.

  Lawfare seems more concerning (though also far from new or specific to Dems) - curious what examples you’re worried about here. FWIW, I think the Supreme Court’s Trump immunity ruling is far and away worse than anything Dems have done in the past several years.

  RFK Jr, a lifelong Democrat (and a Kennedy to boot), has now endorsed Trump because he considers Democrat behavior too undemocratic

  As far as comments from Stein or RFK Jr., don’t third party candidates always hate their mainstream counterparts? The Democratic party, like the GOP, is going to act in ways which help get their candidate elected. Boosting third party candidates who take votes from them is not something an American party will ever do. Maybe that’s not great but it’s more a systemic issue than an issue with the party itself.

  One thing I’ll acknowledge in this vein is the Dems failure to run a real primary this year. I think that was a big mistake. There’s nothing illegal about this though, parties are private entities and can do whatever they want to select a candidate. I consider it more of a strategic mistake than an outright betrayal of Democratic principles. To be clear, I think the DNC is kinda incompetent (probably the RNC too though, Trump’s daughter-in-law is leading it now).

  coordinated social media censorship

  I don’t think the link you provided on Reddit censorship demonstrates censorship? What I saw was mostly people expressing political views in a space where most people disagree with them getting downvoted as well as posts from subreddits with lots of very lefty people where very lefty posts get lots of upvotes. Non-lefty posts and comments get downvotes there. It’s not great epistemics, sure, but it’s extremely typical of both sides. There are very similar conservative spaces where Dem posters don’t even exist (e.g. patriots.win). Am I missing something on this? (I haven’t read the other link you posted which seems more substantial but very long, might read another time).

  Arguably Musk is doing something worse with Twitter right now (though I haven’t looked into it). FB is the go-to place for conservatives and conspiracy groups, I really don’t think it’s a haven for liberal censorship.

  From the linked comment:

  The strongest case for Trump is that the Democrat establishment is systematically deceiving the American people (e.g. via the years-long cover-up of Biden’s mental state

  I think it’s really bad that top Dems covered up Biden’s mental state (which is why I pushed hard to get him to step down) and it reduces my trust in the party. I think this pales in comparison to Trump’s willingness to silence critics (e.g. via hush money and threats).

  generally growing the power of unaccountable bureaucracies over all aspects of life

  To be honest, I’m sympathetic to this concern and I’d be happy to have a reasonable Republican (if we get one) take a swing at reducing over-regulation in 2028. To the extent this is a cost of electing Harris, I will happily pay it.

  What links here?

  • Ba­sic game the­ory and how you can do a bunch of good in ~3 Hours. (de­vel­op­ing ar­ti­cle.) by No longer EA-affiliated (Oct 10, 2024, 4:30 AM; -3 points)

• LintzAAug 24, 2024, 6:41 AM

  13 points

  3 ∶ 0

  in reply to: Larks’s comment on: The case for con­tribut­ing to the 2024 US elec­tion with your time & money

  I agree with the criticism. The quotes provided aren’t good evidence that she is personally concerned about x-risk. We just don’t have much information about her views on catastrophic risks. I’ve updated the text to reflect this and tried to encompass more of what Trump has said about AI as well. Also edited a few other parts of the piece.

  I’ve pasted the new text for Harris below:

  Harris tends to focus on present harms, but has expressed some concern about existential risk.

  Harris has generally put more emphasis on current harms, highlighting that local/​personal harms feel existential to individuals (and implicitly deprioritizing globally existential threats posed by AI) in a November 2023 speech. That said, in the same speech, she acknowledged that AI might “endanger the very existence of humanity”, citing “AI-formulated bioweapons” and “AI-enabled cyberattacks” as particular concerns. In general, it seems reasonable to expect that Harris will at least not reverse the Biden-Harris administration’s previous actions on AI safety. The Biden administration has made impressive progress on AI safety policy, including the establishment of the US AI Safety Institute, securing voluntary commitments on AI safety from many companies, and the 2023 AI Executive Order.

• LintzAAug 23, 2024, 4:19 PM

  5 points

  1 ∶ 0

  in reply to: River’s comment on: The case for con­tribut­ing to the 2024 US elec­tion with your time & money

  To be fair, there is at least one pro-Trump post up on the Forum. Neither this nor that post will go to front page AFAIK because it’s politics, but there are at least points on both sides.

• LintzAAug 21, 2024, 11:24 PM

  12 points

  1 ∶ 0

  in reply to: huw’s comment on: The case for con­tribut­ing to the 2024 US elec­tion with your time & money

  This is what ActBlue, the official Dem donation platform states: “International donors can give to nonprofits, including both 501c(3)s and 501c(4)s, through ActBlue Charities and ActBlue Civics, both part of the ActBlue family of organizations. Only U.S. citizens and permanent residents can give to political groups and candidates on ActBlue, per FEC guidelines.”

• LintzAFeb 15, 2023, 6:00 PM

  1 point

  0 ∶ 0

  in reply to: SammyDMartin’s comment on: Three pillars for avoid­ing AGI catas­tro­phe: Tech­ni­cal al­ign­ment, de­ploy­ment de­ci­sions, and co­or­di­na­tion

  Thanks for this!

  My thinking has moved in this direction as well somewhat since writing this. I’m working on a post which tells a story more or less following what you lay out above—in doc form here: https://​​docs.google.com/​​document/​​d/​​1msp5JXVHP9rge9C30TL87sau63c7rXqeKMI5OAkzpIA/​​edit#

  I agree this danger level for capabilities could be an interesting addition to the model.

  I do feel like the model remains useful in my thinking, so I might try a re-write + some extensions at some point (but probably not very soon)

Three pillars for avoid­ing AGI catas­tro­phe: Tech­ni­cal al­ign­ment, de­ploy­ment de­ci­sions, and co­or­di­na­tion

• LintzADec 7, 2021, 10:40 PM

  1 point

  0 ∶ 0

  in reply to: Mathilde da Rui’s comment on: A bunch of rea­sons why you might have low en­ergy (or other vague health prob­lems) and what to do about it

  Hey Mathilde! Thanks for your thoughtful comment. Curious to hear the mechanism behind eating too many healthy things leading to your issues.

  Also, interesting about the supplements, hadn’t heard that before. I am a bit ignorant on these things but try to offset that by buying the more expensive versions of supplements when trying them for the first time.

  Heartening to hear that you figured it out after a few years!

• LintzADec 3, 2021, 2:26 PM

  4 points

  0 ∶ 0

  on: What “defense lay­ers” should gov­ern­ments, AI labs, and busi­nesses use to pre­vent catas­trophic AI failures?

  In response to an earlier version of this question (since taken down) weeatquince responded with the following helpful comment:

  Regulatory type interventions (pre-deployment):

  • Regulatory restriction (rules on what can be done)

  • Regulatory oversight (regulators)

  • Industry self-regulation

  • Industry (& regulator) peer reviews systems

  • Fiduciary duties

  • Senior management regimes

  • Information sharing regimes

  • Whistleblowing regimes

  • Staff security clearances

  • Cybersecurity of AI companies

  • Standardisation (to support ease of oversight etc)

  • Clarity about liability & legal responsibility

  • Internal government oversight (all of the above applied internally by government to itself, e.g. internal military safety best practice)

  • Technical type interventions (pre-deployment):

  • AI safety research

  Defence in depth type interventions (post-deployment):

  • Windfall clauses etc

  • Shut-off switches for AI systems

  • AIs policing other AIs’ behaviours

  • Internet /​ technology shut-off systems

[Question] What “defense lay­ers” should gov­ern­ments, AI labs, and busi­nesses use to pre­vent catas­trophic AI failures?

A col­lec­tion of AI Gover­nance-re­lated Pod­casts, Newslet­ters, Blogs, and more

• LintzAAug 1, 2021, 9:48 AM

  6 points

  0 ∶ 0

  on: Digi­tal Peo­ple Would Be An Even Big­ger Deal

  I see a lot of talk with digital people about making copies but wouldn’t a dominant strategy (presuming more compute = more intelligence/​ability to multitask) be to just add compute to any given actor? In general, why copy people when you can just make one actor, who you know to be relatively aligned, much more powerful? Seems likely, though not totally clear, that having one mind with 1000 compute units would be strictly better for seeking power than 100 minds with 10 compute units each.

  For example, companies might compete with one another to have the smartest/​most able CEO by giving them more compute. The marginal benefit of more intelligence might be really high such that Tim Cook being 1% more intelligent than Mark Zuckerberg could mean Apple becomes dominant. This would trigger an intense race for compute. The same should go for governments. At some point we should have a multipolar superintelligence scenario but with human minds.

• LintzAJul 28, 2021, 12:38 PM

  2 points

  0 ∶ 0

  in reply to: kokotajlod’s comment on: Digi­tal Peo­ple Would Be An Even Big­ger Deal

  That seems true for many cases (including some I described) but you could also have a contingent of forward-looking digital people who are optimizing hard for future bliss (a potentially much more appealing prospect than expansion or procreation). Seems unclear that they would necessarily be interested in this being widespread.

  Could also be that digital people find that more compute = more bliss without any bounds. Then there is plenty of interest in the rat race with the end goal of monopolizing compute. I guess this could matter more if there were just one or a few relatively powerful digital people. Then you could have similar problems as you would with AGI alignment. E.g. infrastructure profusion in order to better reward hack. (very low confidence in these arguments)

• LintzAJul 28, 2021, 10:35 AM

  6 points

  0 ∶ 0

  on: Digi­tal Peo­ple Would Be An Even Big­ger Deal

  One thing that seems interesting to consider for digital people is the possibility of reward hacking. While humans certainly have quite a complex reward function, once we have full understanding of the human mind (having very good understanding could be a prerequisite to digital people anyway) then we should be able to figure out how to game it.

  A key idea here is that humans have built-in limiters to their pleasure. I.e. if we eat good food that feeling of pleasure must subside quickly or else we’ll just sit around satisfied until we die of hunger. Digital people need not have limits to pleasure. They would have none of the obstacles that we have to experiencing constant pleasure (e.g. we only have so much serotonin, money, and stamina so we can’t just be on heroin all the time). Drugs and video games are our rudimentary and imperfect attempts to reward hack ourselves. Clearly we already have this desire. By becoming digital we could actually do it all the time and there would be no downsides.

  This would bring up some interesting dynamics. Would the first ems have the problem of quickly becoming useless to humans as they turn to wireheading instead of interacting with humans? Would pleasure-seekers just be kind of socially evolved away from and some reality fundamentalists would get to drive the future while many ems sit in bliss? Would reward-hacked digital humans care about one another? Would they want to expand? If digital people optimize for personal ‘good feelings’ probably that won’t need to coincide with interacting with the real world except so as to maintain the compute substrate, right?