I’m less active on LW since this incident.
Most important
I am looking for a cofounder to cyberattack the US govt and US AI companies
I will pay you $1000 after a year if you help me find a cofounder
See also
samuelshadrach
Karma: −30
I probably understand why you are taking this stance. If I was living in the US and a moderator of Lesswrong, I can imagine myself taking a very similar decision as you. Therefore I don’t hold it against you that you are doing this.
barring some pretty truly exceptional circumstances
I think the circumstances are exceptional. You disagree. I don’t see any easy way to resolve this.
I am not interested in wasting even more time trying to word things in the exact way that will avoid the threat of a ban, but yet argue in favour of the specific plan I am pursuing. If someone (including you) wants to seriously discuss this topic with me, doing it off Lesswrong seems like the best bet.
Lastly, as you probably already know, most people with open DMs on LW also have either twitter DMs open or a public email open. So this doesn’t actually stop coordination.
<redacted by mod>
I am also open to some initial feedback, although I might take some weeks before properly responding to it. I am not actively in feedback-collecting mode right now.
The document is incomplete. However, I don’t plan to complete it any time soon, so I figured it is better to post an incomplete document than to post nothing.
P.S. To those of you who are still friends with, or accepting funding from, or working with, the mass murderers at the AI companies, consider this is a very strong warning from my side. The people working at the AI companies are on track to slaughter more people than the Nazis, the Stalinists and the Maoists combined. Cut these people off from your life today. Otherwise, don’t be surprised if you end up as collateral damage in my plans tomorrow. One of my aims is to minimise collateral damage, but I can’t promise there will be zero of it.
Ugh. Fair enough
Short answer
US govt whistleblowers—Read my US govt whistleblower database
Manhattan project—Read the wikipedia pages, read Making of atomic bomb by Richard Rhodes. Just go ask some historians in academia, I think what I am stating here is the consensus viewpoint, and is not that controversial or politicised.
US Cold War history—This stuff is a bit more politicised by the leftists (partly because counterculture movement took off in 60s). You can read Devil’s chessboard by David Talbot for left-leaning view on this.
Journalists, politicians, mass protest movements against the US IC today—I don’t have any good resources on this, but you’ll easily get plenty of (extremely politicised) sources on No Kings and anti-ICE protests and so on. Both left-leaning and right-leaning groups in US have organised various protests, there’s a long list of YouTubers and so on. Ask gpt-5.5 to compile it.
Cypherpunk activists—I probably do have the list of cypherpunk activists compiled somewhere in my notes, but I can’t find it at the moment. Again just go ask gpt-5.5, and actually go read the original blogs, not what the media wrote about them. Or go read the cypherpunks mailing list itself. This index has organised it by popular authors. The short version is that people working on encryption (PGP, Monero, Zcash, Tornado cash, NIST crypto standards) and leaking of classified information (Wikileaks, Blacknet, etc) are the ones who have been the most relentlessly targeted and imprisoned. If you’re just working on a secure OS or on blockchain more broadly or on dark web cloud infra more broadly, you’re less likely to be targeted. There is a long history here. For instance, civilian usage of encrypted radio was banned during or after world war 2, after both Axis and Allies used Numbers Stations to send messages, and it remains banned to this day. Digital encryption was initial regulated as a munition, but due to courage and effort of the PGP authors and a few other cypherpunk, PGP atleast exists now and is legal to use. Almost nobody uses it, and people working on such tech are at higher risk even today.
Public statements by heads of US IC in the last few years—Read my Theory of change for US govt whistleblower database and guide, in particular the red-teaming section
Why I am not writing a long answer
I am busy doing actual object-level work, and I have already wasted a year of my life arguing on lesswrong with random people with no skin in the game. One year is extremely costly given how close we are to the end of the world.
If someone who actually has expertise or power can commit some actual skin-in-the-game, I can spend more time talking to them. For example, if someone says they will pre-commit to devoting X hours of their life giving feedback, or someone will pre-commit to devoting Y amount of dollars for my writeup or as a grant to me more broadly, or someone will seriously consider working with me full-time.
I think the US intelligence community has had absolute unaccountable power to take key decisions of the US govt for the last 80 years (since world war 2), and I don’t see this changing by default.
Key decisions they unilaterally take include decisions on sanctions (on equipment, reagents, etc) to stay at the frontier on any technological race in last 80 years. Another key decision they unilaterally take includes decisions to start wars.
They are not accountable to the Congress / Senate, or to the Supreme Court, or to any regulatory body or similar, when it comes to these key decisions.
Heads of US intelligence already have active relationships with the heads of all major US AI companies. They are one of the primary reasons the US AI companies will remain unaccountable to the US public, unless something big changes.
I think social media influencers are most accountable to the public. Politicians and billionaires are less unaccountable than them. Heads of intelligence are the least accountable to the public. There are structural reasons why this unaccountability exists.
Persuading heads of US intelligence community to stop the race towards ASI is too hard, they are likely to accelerate towards ASI even if they have some understanding of the risks.
I have a fairly strong belief in this direction. Why?
I have studied every US govt whistleblower leaking classified info in the past 50 years as part of the work I published.
I have read atleast a couple books about who actually took the decisions during the Manhattan project.
I have read a bit about US cold war history, and a bit on more recent wars started by the US, to see who took the decisions.
I have also looked at some of the journalists, politicians and mass protest movements in the US trying to (unsuccessfully) hold the US intelligence community more accountable, in more recent times.
I have read the blogs of almost every famous cypherpunk activist, and have carefully tracked who got arrested versus not, and why.
I have gone through public statements of many recent heads of US intelligence on topics including AI, cybersecurity, social media, surveillance, drone warfare, and so on.
(Incomplete list) Examples where I think the US intelligence community was held even a little bit accountable to the public
I think the Vietnam war stopped sooner than it would have otherwise, because of mass protests.
My guess is that one factor for why Truman didn’t support von Nuemann’s plan of doing a pre-emptive first strike and creating a nuclear monopoly for the US, was that he was afraid of mass protests and the authoritarianism he would have to engage in to suppress these protests.
How can you change my mind here?
The best way to change my mind is obviously to actually get US IC members in support of an AI pause.
I have very little hope in the “inside game” strategy that EA billionaires are attempting, such as putting Jason Matheny in RAND or Paul Christiano in NIST or similar.
I have slightly more hope in being Loud and Annoying. Actually go organise protests at Fort Meade. Actually go figure out which social media all these US IC people use and persistently call them out there, for not taking a stance on AI pause.
The second best way to change my mind is to offer historical evidence like I have shared above. Although, given that I already have a large number of datapoints on my side, just one datapoint on the opposite side won’t be enough, you’ll need either many datapoints or a datapoint with a lot of weight.
I am considering precommitting to the following:
If I ever believe ASI is less than 1 year away with above 50% probability, I will commit suicide in front of a govt building as a desperate final act of protest.
Why?
I want to demonstrate skin-in-the-game, that actually I do consider a world that builds ASI to be worse than a world where I die, as long as my death helps reduce probability of ASI being built.
This comment is not the precommitment itself, I am just considering it at this stage.
My current timelines are 25% ASI by 2030, 40% ASI by 2035 by the way, so I am unlikely to commit suicide atleast in 2026 or 2027, even if I did precommit to this.
If you know any pros/cons, feel free to share.
I am a lot more hesitant to update on the opinions of random strangers after I have spent a lot of time thinking about and reading about a topic. So, if you want me to update my opinion based on yours, the best time to inform me of your opinions is right now, while I haven’t yet spent much time thinking about this topic.
Audio for the Musk x Altman lawsuit
Sharing this violates US law
2026-04-29
Maybe especially focus on persuading Taiwanese govt and public about AI pause
Disclaimer
Quick Note
Contains politically sensitive info
target audience—lesswrong
I haven’t verified all these claims a lot.
Main
Some LW discourse seems to implicitly assume that Taiwanese politicians and billionaires, and the Taiwanese public don’t have much say in what happens to TSMC.
There is some writing on lesswrong that assumes either the leaders of the CCP or the US executive branch will inevitably take control of Taiwan via military invasion. See also: This LW post on Taiwan war timelines
There is also some writing that talks about how “middle powers” can intervene, which seems to me to be a euphemism for asking either the Indian or Pakistani or Russian executive branch to drop ballistic missiles on TSMC. See also: ControlAI on middle powers. It’s not clear to me who “middle powers” even refers to here and this lack of naming names seems intentional. Non-nuclear states don’t matter IMO. Executive branches of Israel, UK and France are unlikely to ever defect against the US executive branch. The North Korean leader is unlikely to ever defect against the CCP. This basically leaves executive branches of India, Russia and Pakistan as the only “middle powers” who can pull off something as radical as this.
However, whatever happened to just convincing people in Taiwan directly? TSMC has a democractically elected govt and an independent military, that gives them atleast some minimum amount of time to decide what to do with TSMC, before anyone successfully invades them.
Some options that are possible if you could actually convince people in Taiwan about an AI pause:
I think it is possible, atleast in theory, for a handful of politicians and billionaires in Taiwan to decide to shut down TSMC themselves, if they were convinced about the need for an AI pause.
I think it is possible for them to successfully install a deadman’s switch in TSMC, such that they could blow up TSMC before it falls into hands of either the US executive branch or the CCP. Credibly committing to destroying your own infrastructure to repel invaders is called scorched earth strategy, and has been observed in many previous military conflicts.
I also think it is possible, atleast in theory, for the Taiwanese public to try to shut down TSMC, either via a nonviolent referendum or election, or via an armed revolution, if they were convinced about the need for an AI pause.
Lastly, I do think it is possible, atleast in theory, for a militia (probably composed of Taiwanese citizens) to attempt to blow up TSMC without attempting to get consent of either the public or the elites. You could call such a militia a terrorist group if you like.
All these plans require actually persuading people in Taiwan about the risks of ASI.
Note that what persuasion strategy works in Taiwan probably won’t be the same as what works in the US, due to cultural differences. I don’t expect a translated version of Yudkowsky’s book to sell as many copies in Taiwan for instance.
I am not an expert on persuasion or on what these cultural differences are. My guess (from very little data) is that it overlaps with the social left versus social right divide, altruism being directly towards family/community/etc versus the world.
I think someone should actually try to understand Taiwanese culture a lot and figure out effective persuasion strategy for them.
Disclaimer
Contains politically sensitive info
I had temporarily deactivated my account for safety reasons. However, this topic seems very important and hence I feel compelled to respond to you regardless.
I wrote this quickly, but it is based on an older post I did think through a bit more.
My reply
Plan 1: Run a private militia that can maintain trust, chain of command, obtain weapons, etc. Use this militia to attempt hundreds of simultaneous assassinations to wipe out of the entire chains of command of US executive branch, US intelligence, and all US AI companies. When I say chain of command of US exec branch I mean president, vice president, and so on. When I say chain of command of US AI company I mean the entire C suite at minimum.
The assassinations need to be simultaneous because otherwise you’ll get retaliated soon after. Project Valkyrie and similar projects understand Himmler and Himmler’s subordinate and so on had to be assassinated too, not just Hitler. Lone terrorists have bad thinking because of social isolation, but organised militias tend to not be as naive about all this.
I agree Plan 1 is too hard and there are probably easier ways to solve ASI risk.
Plan 2: Incite a mob to actually storm into TSMC chip fabs and destroy all the equipment.
It is important to target the chip fabs, not random datacenters, for obvious reasons.
If you understand Chinese/Taiwanese politics at all, you also understand lower bound of how hard this is. I agree Plan 2 is also very hard and there may be easier ways to solve ASI risk. I am not sure, however, and I might still be open to this as a last resort.
Plan 3: Make a failed assassination attempt for propaganda reasons.
I think Gwen would agree with me that a major value of attempting assassination is propaganda value. It polarises the conflict, people’s emotional temperature raises on both sides.
I don’t yet have strong opinion on Plan 3. Is the propaganda values of an assassination attempt net positive or net negative, and what is the optimal timing of this? I don’t think my political skills are that good.
Quite bluntly, I think you, Yudkowsky, also are pretty mediocre at politics. Proof—There are 100,000 followers on the top ASI risk YouTube channels, not 100,000,000 followers. I have very little trust in your judgment on what is actually good or bad way to reach the masses on this topic.
Since I am not an expert politician/propagandist, I’ll just share my own personal feelings. I want the temperature of the conflict (between pro-ASI and anti-ASI factions) to rise. There is no morally neutral way of accusing someone of causing a thousand holocausts all at once, and I instinctively have (a bit) less respect for people who talk about extinction probabilities in such an emotionally detached manner. One of the best ways to increase temperature of the conflict is to actually be violent. I have some aesthetic preference in favour of violence and bloodshed, “kill or be killed” mindset, and I increasingly have distaste for a bunch of high-minded moralising from people in EA/LW circles who have no actual power. I agree with Naval Ravikant that something like 80% of social media discourse on morality is just a self-interested status game. People don’t want you to be violent against them for basically self-interested reasons, and then they come up with all sorts of (fake) arguments on why you should not be violent because some (fake) altruistic reason.
All that being said, I will reiterate that I don’t actually know what the propaganda value of an assassination attempt as of 2026-04-14 actually is, if I leave aside all my personal feelings.
I also have a fear that you too, are laundering some of your own personal feelings, when you write a post like the above. You have some aesthetic preference in favour of niceness and civility, just as I have aesthetic preference in favour of some violence.
There are other violent plans besides these but I won’t discuss those here unless it comes up as relevant context.
Crosspost: Learnings from past month of reading about politics
I am interested in: Constructive feedback on morality of cyberattacking US govt and US AI companies to leak their secrets
I am not interested in: Debating timelines or xrisk. Debating whether mass movement against US AI companies is a good idea. “Boo, go away”-type comments that reinforce groupthink on LW.
2026-03-25
Learnings from past month of reading about politics
Disclaimer
Quick Note
Contains politically sensitive info
target audience—strictly myself
Background context
I was hesitating working on the plan to cyberattack the US govt and US AI companies and leak their information. I used multiple ways to cope with lack of clarity on such topics.
I wanted to figure out my morality and politics, so that I work on this with more clarity, instead of at half-speed. Hence I started reading even more books.
My views on morality were previously a very Might makes Right view on history. But I was still hesitating acting in line with these views.
Technological offence/defence balances decide the routes to obtain power, people who obtain power are eventually respected by many, and these people then invent rhetoric on morality that justified the routes they took.s
In the surveillance/privacy/transparency context, I had a fairly strong view that multiple offence/defence balances around information are shifting towards offence, and most of the world’s information is going to come out publicly anyway, and this is more good than bad (although there is bad), and hence I should just act to take advantage of this.
However I noticed I still seemed to losing self-esteem by working on it. Just because you can get away with XYZ, and later invent rhetoric to morally justify it to the public, doesn’t mean you’ll feel good doing it. I realised I need moral rhetoric to justify it to myself first (not others), and also this had to be in line with my past self and past identity. Morality is part of identity and I can’t 180 degree my entire identity overnight.
Main learnings from the past month (technically some of these are from preceding months, but whatever)
Personal
Internalising that the world is ending helps me a lot
Death
One way of internalising the world ending (as opposed to treating it as an intellectual game) that I found super valuable is to go to a tall building and look outside, and visualise everyone die.
I visualise people dying in almost every conversation I have nowadays 24x7.
I don’t know if I advocate everyone do it, or even future Samuel do it, but it is what I do often nowadays.
It helps me take the stakes seriously.
“Blame yourself, preserve your agency” is this Naval Ravikant quote I take seriously now.
Fuck lesswrong, fuck indian politics, fuck whatever. If everyone else jumps off a bridge doesn’t mean I will jump off one too. If the world ends it was my fault.
It helps me connect with people a little bit more.
If other people don’t want to take my ASI risk concerns seriously in a conversation, that is now their problem not mine. Most people know jackshit about this topic and hence I don’t take their opinions too seriously.
You can hear out my concerns, and then move on to a separate topic, that is fine.
You can hear my concerns, have a respectful disagreement, that is also fine.
But if you don’t want to hear my concerns, and you already have pre-conceived notion I am crazy or politically annoying or whatever, then that’s it, we are done here, get out of my life, meet me after a few months at best.
Liking people, versus considering them political allies (I learned this some months back itself)
I like some of the people at AI the companies because of shared traits. Atheist, high openness to experience, intellectually curious, high trust, kind, support technology, capitalism and democracy, technical people. However, inspite of that, if they threaten to kill me then I will threaten to kill them first. This is the same way I look at my friends too. If my friend threatened to kill me out of some deranged beliefs in utopia, I’d kill my friend too.
I dislike many of the anti-ASI political demographics. I dislike religion due to a lot of personal experiences I can’t share here. I dislike most leftist thought. However, these people are among the biggest demographics in the anti-ASI political coalition.
Self-interest versus greater good
I was quite self-interested in teenage, leaned more altruist in college after my basic needs were fulfilled (loved ones, sufficient money, etc), now again self-interested after realising ASI risk could get me killed. Saving my life again seems more important to me than saving anyone else’s life, although I also care non-zero about the latter.
I will feel better about myself if I tell myself I am doing all this for greater good, saving humanity, blah blah, as opposed to saying I am doing it to save my own life. But that doesn’t make it true.
I know most people don’t want to go extinct or live under permanent dictatorship. However, beyond these basics, I have no fucking clue what most of humanity actually wants, so how can I pretend to represent them?
I know all humans want basics like they want food, shelter, safety from violence, love and respect of loved ones. They want atleast some certainty (but not complete certainty) about their past, present and future. They want free time to pursue “enlightenment”, whatever the hell that means to each person.
If I was given access to private diary notes of millions of people, and asked to predict what was instead of them without looking, I would be very bad at predicting it. The basic needs are boring and simple, but humans build a complex web of nuances on top, and I don’t think I’d be good at predicting it.
Many ideologies have this idea of “what is in people’s best interest, if they knew better, had better experiences, etc”. Hinduism might call it escape from reincarnation, Buddhism might call it Enlightenment, Yudkowsky might call it CEV, Marx will say overthrow upper class, etc. I am suspicious of most ideology’s answers to this question. I also don’t know what is in most people’s enlightened interest, and most people themselves don’t know what is in their own enlightened interest. People’s stated and revealed preferences are both shit, for predicting their actual enlightened interest.
I love violence?
This is a weird thing to say, but at some level, people who are too high-minded idealist and hence non-violent, often irk me. I think Thomas Sowell’s contrained versus unconstrained visions helps better explain why too much idealism irks me as a person. Or Viktor Suvorov’s reasoning. Or the scriptwriting for the Joker in the Dark Knight. “When the chips are down, these people will eat each other”-type rhetoric.
I have always been a sort of violent person, atleast in the privacy of my mind. Over many years of my life, I have at many times thought of killing people for all sorts of reasons, I just never did it irl because it wasn’t necessary to solve the problem. For the first time I am being confronted with a problem big enough that violence seems actually required to solve it.
I am not a sadist or masochist. But I also don’t think I have as much aversion to violence as some people around me seem to. I don’t have much problem looking at murder or rape footage for example.
Violence being necessary doesn’t make it an absolute good.
I think it is important I don’t Other or dehumanise my outgroup. “Look at a man in their eyes before you put a bullet through their skull.”
I agree with Duncan Sabien’s take on necessary violence. I think it is important that future Samuel doesn’t forget about this.
Political
It is depressing how little progress there has been in political science, and how little contact with reality most people’s political beliefs make:
People have been failing to build a world government for a literal century at this point. League of Nations, the UN, even Nehru and Gandhi write about their dreams for world government (atleast until the Indo-China war in 60s)
Significant overlap between 1960s counterculture and Julia Galef’s SF Bay Area map of today. Same hedonism, same social justice and environmentalist stuff, similar spread of political ideologies. Anti-nuclear protests in 1950s were just as unpopular then as anti-ASI protests are today, but if you bring in musicians and weed and talk about more normie causes that affect day-to-day life (like sexuality or parent-child relationship or whatever), then the crowd grows much larger.
The Federalist Papers from friggin late 1700s are discussing all the questions that people debate on reddit today. Power to states versus the centre, standing armies versus not, republic versus direct democracy.
My belief in both capitalism and democracy seem to have weakened due to my plans to fix ASI risk, and this makes me sad. I am not yet over it (as of 2026-03-25).
There are levels of belief in capitalism and democracy.
Example of ideological defender of capitalism: Naval Ravikant, Paul Graham. Example of ideological defender of democracy (atleast as per how federalist papers defined it): Tucker Carlson, Joe Rogan
Example of ideological defender of capitalism who violated some laws but not the “spirit” of capitalism: Ross Ulbricht. Example of ideological defender of democracy who may have violated some laws but not the “spirit” of democracy: Glenn Greenwald
Examples of capitalists who violate the ideals of free market capitalism, and are happy to purchase politicians and intelligence community members and cement their own monopoly: Marc Andreessen, Peter Thiel. Examples of politicians who violate the ideals of democracy: basically every US presidential candidate in last 30 years.
I don’t respect the people at the bottom of this list as much, and I will be sad if I also have to become more like them. I’d rather be in the middle category at best.
I might respect myself less as a person. Or, if I successfully change, I might start respecting other people who violate the spirit of capitalism and democracy more, and start respecting people who follow the spirit less as they will look like suckers to me.
I do think that the more people violate the spirit of these ideals, the more they contribute to polarisation, and people on the other side of the political isle also look at the them and choose to violate the spirit in turn.
I have important ideological differences with most of the people in the reference class of “software developers turned activists who used geopolitical arbitrage”
Examples: Aaron Swartz (progressive, anti-copyright), Snowden and Assange (anti-war anti-surveillance anti-US IC), Satoshi (anti-Keynesian), Roman Storm or Wilko Zookox (libertarian), Anna’s archive / Alexandra Elbakyan / torrent / napster (anti-copyright), Ross Ulbricht (libertarian, anti-drug law), many youtubers living outside their home country to criticise the regime there (pro-democracy, anti-defamation / pro-free speech). Richard Stallman (deontological anti-everything lmao)
I am not a libertarian or a progressive. I don’t have a very strong position on anti-war or anti-surveillance or anti-copyright or anti-taxation or anti-drug regulation or whatever. I have positions on these topics, but they’re weaker, and sometimes in direct conflict with other people who have worked on them. (For instance I am a lot more pro-transparency, whereas some of these people are a lot more pro-privacy.)
I have a very strong anti-ASI position. I have built up this position over many years. It makes me a bit sad that I need to devote my life to working on this, instead of something else I care more about. But, I have to do it now, there is no choice.
I also have a strong position in favour of more transparency (and less privacy) in society, both among the elite and the general public. I have built up this position over many years. I see a world with zero privacy as a platonic ideal in the same way I see a world with zero nukes or guns as a platonic ideal or a world with world government federal democracy as a platonic ideal.
I still haven’t completely given up hope that ASI risk can be fixed within the moral and US legal constraints of democracy and capitalism, but I don’t feel comfortable gambling the fate of all of civilisation on that hope.
See above—I will genuinely dislike myself as a person if I was some high-minded idealist and the world ended as a result.
Important—think more here
I dislike most of the people who advocate for violence for instrumental political reasons, such as anarchists and far-left activists. In general, a broad anti-ASI coalition is needed, but it is full of a lot of political demographics I dislike.
Interpersonal
Positive-sum versus zero-sum
Many silicon valley billionaires have heuristics that you’ll be more motivated to work, and other people will be more motivated to help you, if you work on a positive-sum thing involving creating new wealth or building technology, as opposed to a zero-sum game.
Managing motivation of you and your immediate set of people who will help (cofounder, friends, allies, etc) is one of the most important predictors of whether you succeed or fail. I actually agree strongly with all this.
However, I still think a zero-sum political battle is the only way ASI risk gets fixed. We need to fuck over the guys building ASI, and ensure they lose wealth and power (and status and girlfriends and loved ones and whatnot), and just, generally ensure they suffer. They have to lose power for the anti-ASI people to gain power.
Polarisation
It is straightforwardly true that if I use political violence, then people on the other side will also see this and be inspired to use political violence, and this too will further weaken what democracy and capitalism is still left in the US.
This is well-documented and many political youtubers talk about it, and I should not underestimate the effects of this. Moderate people (who aren’t terminally online thinking about politics) are likely to be polarised against you if you use political violence.
Only the outgroup
I have understood very clearly that I should not be violent against innocent bystanders if I can help it. Only be violent against the outgroup. You lose a lot of trust networks if you are violent against bystanders.
Moral feedback
I have made up my mind that if I go ahead with the cyberattack plan, we have to completely transparent regarding our estimates of how many people we have hurt and why. Most of the public’s feedback will be shit, but some of it will be useful. We will be entering a high-stakes adversarial domain, this will cut us off from most feedback including moral feedback, hence we have to get feedback this way.
Books
If someone else is reading this writeup, my top book recommendations to them:
Thomas Sowell—Conflict of visions
Srdja Popovick—Blueprint of revolution
Federalist Papers—I used AI to simplify the english and read it
Richard Rhodes—Making of the atomic bomb
My actual reading list is too long to share here. I also have a lot of books on tech, or on whistleblowers, for example.
A camera that can do facial recognition from outside of national borders doesn’t need to be a petapixel one. A mid-gigapixel camera with good optics can cover an entire city at once (or at least it could if it wasn’t for all the buildings in the way).
This is technically true. But yes, if you had the tech to build this it would also become trivial to built a petapixel camera too (for someone who can afford it). The hard part is doing 0.1 metre resolution from a 10,000 kilometre.
Thanks for this exchange btw, I guess in future I could be more precise.
The main barrier to petapixel cameras is that they don’t serve your goal of full public monitoring (regardless of whether it’s by the government or by everyone individually).
Why?
Assume we had the tech to manufacture petapixel cameras, and individuals worldwide could purchase them (i.e. a govt couldn’t just lock down the supply chain). Why does this not eventually to a world with zero privacy for everyone?
Thanks this comment is useful!
single aperture of ~500m
maintaining satellite relative positions to within a wavelength of light
There’s no law of physics that prevents humanity from building either of these things. I’m just pessimistic about the engineering advancing to the point that we can build this in next 10 years. (without help of superhuman intelligence, that is)
I am still not clear where we are disagreeing, sorry.
What do you think is the bottleneck to building a petapixel camera that lets you do facial recognition from outside national borders? I don’t think you can simply stitch a bunch of gigapixel cameras together and achieve this.
Okay. I think what I want is feedback on tactics, not strategy. I don’t want to debate why AI pause political movement is required, yet again. I don’t want to debate why US intelligence community will accelerate development of ASI, yet again. I can debate details of how gigapixel cameras work.
And I don’t have time for actual “proof” yes, I’m just gonna make a guess. Which could involve me projecting stuff based on reference class of similar comments or similar people commenting.
I understand all this. I am assuming a constant angular field of view, let’s just assume 120 degrees for now. I assuming a single photo from a single camera placed at a single location covering 10^15 pixels. I am not talking about multiple photos stitched together, or moving the camera around, and so on.
(Yes, the camera will necessarily have multiple sensor arrays and internally stitch the data together anyway)
And yes a petapixel camera with 120 degree (or some other large field of view) could cover the United States at 3 mm resolution.
I am not sure if we are actually disagreeing.
I am saying someone should be able to place a camera outside US borders and yet be able to do facial recognition of people inside from thousands of kilometres away.
I listed the plan above in short. Friendships are the biggest conflict of interest. If you are not willing to distance from people building ASI, you are unlikely to pursue actually effective plans to stopping ASI development.
If you want a longer version, here it is: Support the movement
Can you give an example in the real world? (Prefer historical examples if you dont wanna be too controversial) Both your comments are abstract so I’m unclear what you have in mind.
Lesswrong core members unwillingness to engage in conflict is directly leading to the end of the world.
By conflict I mean publicly humiliating developers at these companies (including Anthropic), cutting them out of your social circle, organising protests outside their offices, and running for election with an antiAI company message.
I am willing to go further by supporting whistleblowers and cyberattackers against AI companies. But the above is the minimum to become my ally.
In some hypothetical game theory puzzle sure. In the real world it does necessitate it with like >95% probability.
And here we are talking about positive sum stuff like growing a business.
Pause AI movement is explicitly a zero sum political battle.
Positive sum games still involve a lot of zero sum moves! Just because the pie is growing doesn’t mean it doesn’t matter who gets more of the pie. If you are a company CEO in a growing industry, you will end up taking adversarial moves against lots of people. You will sue people, you will fire your employees, you will take away profit from your competitors if you succeed, and so on.
This is the linked comment. Both of these people are me.
I can still message people on twitter or email.
I have slightly better quality (but not significantly better quality) writeups on all these topics on my website. I think I will avoid saying more about this topic in this thread to avoid being banned.