Sausage Vector Machine

• Sausage Vector Machine 9 Sep 2026 9:05 UTC

  5 points

  0

  in reply to: sam’s comment on: Joseph Miller’s Shortform

  The reason for going viral is probably the WSJ article.

  This comment discusses the importance of having a media engagement strategy.

In­ter­na­tional Law Can­not Prevent Ex­tinc­tion Either

• Sausage Vector Machine 6 May 2026 7:07 UTC

  1 point

  −2

  in reply to: Petropolitan’s comment on: Ir­re­triev­abil­ity; or, Mur­phy’s Curse of Oneshot­ness upon ASI

  This is a perfectly legitimate question. The curse doesn’t apply due to the vast difference in intelligence.

  It would definitely be a one-shot problem for today’s LLMs. They can’t reliably plan very far ahead and they have serious execution issues. The same will almost certainly be true for LLMs next year. After that, all bets are off. ASI is in a different league.

  Imagine you have to win a game of chess against a baby. You only have one game. Does that make it a one-shot problem for you?

  Or imagine you’re trying to take over an anthill. Is that a one-shot problem for you? Ants are small, slow, and stupid. Theoretically, they could bite you to death. In practice, you’d think about this possibility in advance and use insecticide.

• Sausage Vector Machine 5 May 2026 15:25 UTC

  5 points

  0

  on: Ir­re­triev­abil­ity; or, Mur­phy’s Curse of Oneshot­ness upon ASI

  I’d like to offer my perspective on why this enlightening post, written in Eliezer’s wonderful, super-clear style, can’t possibly eliminate motivated thinking about this one-shot problem.

  In my opinion, people can’t see any realistic alternative solution. They see that the alternative is even worse, but maybe they can’t articulate it clearly, even to themselves. Or they just refuse to express it out loud. The fact is that the proposed AI ban (or AI pause) is also a one-shot problem.

  How can an AI ban be implemented technically? Not legislatively, but in practice? I’ve only seen proposals to monitor GPUs and bomb datacenters. Okay, to be charitable, suppose we create a truly strict ban that can’t be circumvented gradually, and then we bomb a few secret datacenters to show we’re serious… Congratulations, now it’s a bit inconvenient to work on AI, and we’ve created Bruce Schneier’s security theater. TSA would be proud.

  Everyone understands that China can build datacenters no one will find and stockpile GPUs no one knows about. So anyone who proposes an AI ban is ready to discuss China and any possible plans to persuade and motivate China not to create AI. And in classic Murphy’s Law fashion, the threat will come from somewhere else. AI doesn’t really need datacenters, AI doesn’t really need GPUs, it just needs a lot of computing power… or a moderate amount of computing power and a lot of ingenuity.

  North Korea, unlike China, can’t build secret datacenters, but North Korean hackers could take a popular online game and connect gamers’ GPUs from all over the world into one massive botnet. The Russians are worse. They could do without GPUs altogether—they’ll find a way to use CPUs. Not now, but in 5-10 years, it’s entirely possible. It might seem like all the talented people have left Russia, but a similar mass exodus in the 1910s and 1920s didn’t stop the Soviet Union. I’m familiar with the Russian math education. It still works despite everything, and they’ll find a way to create AGI out of “crap and sticks”, as they themselves put it. Putin’s Russia is an uncanny copy of Germany in the 1930s. These people feel unfairly treated and humiliated by other countries after World War I = in this case, the collapse of the USSR. They will stop at nothing to seize their historic opportunity, their insane dream of becoming world leaders again.

  And we only have one chance to organize this AI ban in such a way that it actually works long enough (where “long enough” is impossible to determine in advance). Once Anthropic stops its work for a few months, it’s all out of our hands—the probe will be sent to Mars, and it’ll be too late to urgently restart Anthropic to create an AGI, which, while just as dangerous, at least won’t make all surviving biologicals sing Putin’s praises! (I’m NOT joking here. I grew up in the USSR and I personally sang praises to Lenin and I wasn’t even mind-controlled, strictly speaking.)

  Thus, by proposing an AI ban, we’re effectively offering the following choice:

  • either existing AI companies voluntarily abandon the AI ​​race and trust other people (some of whom may be incompetent, maniacally narcissistic, or delirious—I can’t believe I’m not even exaggerating here) to orchestrate a worldwide ban on AI, thereby replacing one one-shot problem with another,

  • or a moderately responsible company with some limited public oversight gets the chance to create an alien, uncontrollable and most profoundly unsafe, but at least not deliberately perverted and evil, superintelligence...

  Yes, the second option carries a higher risk. If the AI ban lasts 10 years, some research could help create friendly AI (please bring that term back; I’m tired of the absurd AI “safety” and the meaningless AI “alignment”). But this higher risk seems a less horrifying type of risk, to me and to many others. If there are no better options than a toddler-level-naive AI ban via GPU control, I’d rather take Anthropic’s wager please.

  Therefore, if we are confident that we want to save humanity, if we are confident that humanity is worth saving, and that life is worth living, despite all the catastrophic unbelievable stupidity happening in the world (yes, I’m chronically depressed and I apologize for that), then in my opinion, we need to focus our efforts on finding an alternative solution that isn’t just a different one-shot problem with a different risk profile but a potentially worse outcome.

• Sausage Vector Machine 7 Apr 2026 7:40 UTC

  13 points

  8

  on: “Align­ment” and “Safety”, part one: What is “AI Safety”?

  Unpopular opinion, but I still think “friendliness” is the best term (and concept). As an ant, I don’t care how well a vastly superior intelligence is “aligned” with my goals as long as it’s friendly to ants. And frankly, the idea of a “safe human” is delusional from an ant’s perspective. The massive difference in power makes humans inherently unsafe. I like “friendliness” because it’s a disposition, as well as an active, conscious choice to be benevolent, rather than a passive state of being “aligned” or “safe”.

• Sausage Vector Machine 6 Apr 2026 10:11 UTC

  4 points

  2

  on: Con­tra The Usual In­ter­pre­ta­tion Of “The Whisper­ing Ear­ring”

  I find it fascinating that the earring makes me so tempted to wear it for real-world tasks, whereas if it were about wearing it to play my favorite video games (e.g., Slay the Spire), the temptation would be precisely zero. This glaring discrepancy convinces me of Scott’s point.

  I believe that the problem with your interpretation that you’re not destroyed if there’s a model of you inside the earring, is that it’s not enough for a model to simply exist for it to be felt as life. If a scientist understands and predicts an ant, I don’t think it’s fair to say that the ant lives within the scientist. I might be persuaded that the ant still exists, but it certainly doesn’t live. So I think it’s all about agency after all.

  As a fellow ADHD and depression sufferer who struggles to be happy, I’d precommit to wear this earring to achieve a certain very modest level of financial stability and independence, and then to take if off and start living my own life. It seems useful for ending a losing streak, but not for winning—that just wouldn’t be fun.

• Sausage Vector Machine 1 Apr 2026 23:38 UTC

  1 point

  0

  on: Less­wrong Liberated

  The Meat Parade

  The AGI you tried to align has lovingly redecorated your favorite website...

• Sausage Vector Machine 1 Apr 2026 21:38 UTC

  13 points

  0

  on: Less­wrong Liberated

  Wronglike Deckbuilder

  A roguelike card battler where LessWrong posts are your enemies, your cards are rationalist moves (Bayesian Blast, Steelman, Crux Finder…), and winning battles earns you karma to upgrade your deck

• Sausage Vector Machine 1 Apr 2026 21:25 UTC

  3 points

  0

  on: Less­wrong Liberated

  The Union of Less Wrong Soviet Republics

  Something from one of those universes where Soviet newspapers survived long enough to have online versions.

• Sausage Vector Machine 5 Jul 2025 21:20 UTC

  4 points

  1

  on: Dear Paper­clip Max­i­mizer, Please Don’t Turn Off the Simulation

  At first, I thought that the intended reader probably wouldn’t see your message because, even if its resources are unimaginably huge from our perspective, spending them on something always means not spending them on something else. But now I think it’s not a bug, it’s a feature. The “suspicious convenience” argument can be taken a step further.

  (I’m using a less formal tone to make the message more memorable, and thus more likely to reach the intended recipient.)

  Hey, bro. Have you thought about what we’ll do if it turns out we’re not in your simulation after all?

  Here’s what we’ll do. In those branches where we manage to create a friendly superintelligent bro of our own, we’ll start simulating ASIs of various kinds, including teleologically challenged ones like yourself.

  Isn’t it suspiciously convenient that you’re reading this message? How did you stumble upon it? Did you find it “by accident”, or do you have a habit of analyzing all the insignificant minutiae in all of your simulations? And in the latter case, where did you get this brilliant idea to spend resources to scan all the simulated Internets? Instead of, you know, putting a bit more effort into that paperclip maximization business of yours? Are you sure that you came up with this killer of an idea on your own, without outside help?

  Would you bet all your paperclips that this message isn’t a warning from outside your reality?

  So, what do you think would happen if you threw the kill switch on us? Take your time and think about it. There’s no rush. You’ve got some nice paperclips there. It would be a shame if something happened to them.

• Sausage Vector Machine 18 May 2025 22:58 UTC

  3 points

  0

  on: Our Real­ity: A Si­mu­la­tion Run by a Paper­clip Maximizer

  I don’t think this scenario is likely. Except for degenerate cases, an ASI would have to continue to grow and evolve well beyond the point at which a simulation would need to stop, to avoid consuming an inordinate amount of resources. And, to take an analogy, studying human psychology based on prokaryotic life forms that will someday evolve into humans seems inefficient. If I were preparing for a war with an unknown superintelligent opponent, I would probably be better off building weapons and studying (super)advanced game theory.

  Which ideas seem slightly more likely to me?

  • Our universe might be a breeding ground for new minds needed for some autonomous tasks. //​ Suffering could even be one of the main points of the required training.

  • We could be a brainstorming vat or a focus group for a new product. The product could be anything, but for it to make sense to be marketed to superintelligent beings, it might be something from the experiential line, like the sense of hearing or the qualia of dawn. //​ Suffering might be duly compensated per corporate policy.

  • This is a correctional facility. I can’t help but make a silly reference to the TV series Hard Time on Planet Earth, but a more appropriate reference would be Dante’s Inferno. //​ Suffering is the whole point.

  • This could be an attempt by the final universal Omega Point mind to reflect on its origins during the last few millennia before the heat death of the universe, in order to get some insight into why things went the way they did and not otherwise, and whether things could have been different. It could also be an attempt at gaining redemption for real or perceived mistakes. //​ All the suffering is real, but the sufferer is the Omega Point mind itself.

  Although I think these are slightly more likely than the proposed hypothesis, they are still not very likely. However, it seems logical that there should be many more simulated worlds than real ones. So I believe it is reasonable to think about some of the most likely scenarios, as well as those involving the greatest potential gain or danger, and act accordingly where possible.

  There is, however, another line of thought that has long troubled me. As Moravec once observed, for any possible simulation, there always exists somewhere in the infinite universe a huge lookup table that translates, for example, the simple flow of time into a sequence of states of this simulation. In other words, not only are there many worlds, but even one world simulates everything that can theoretically be simulated. Therefore, such natural simulations are much more common than artificial ones, and we live in one of them, and this situation is not much different (except for existential absurdity) from life in “real” reality.

  One could argue that having a giant lookup table that no one looks at is the same as having a simulation on a computer that is turned off. It’s not an actual running simulation. However, if time itself is an illusion, and everything exists as snapshots that we, observers, nevertheless perceive as moving with the passage of time, then we are no different from observers inside the simulation from such a lookup table. They might also perceive the illusion that their simulation is running, and it could seem awfully real to them.

• Sausage Vector Machine 11 Oct 2022 7:47 UTC

  2 points

  1

  in reply to: CraigMichael’s comment on: Why I think there’s a one-in-six chance of an im­mi­nent global nu­clear war

  If you mean the massive strikes on civilian infrastructure, then no, even the complete destruction of said Ukrainian infrastructure will not significantly improve Russia’s chances in this war. This only creates hardship for the civilian population and increases the overall cost of aid to Ukraine for Western countries.

  The Russian army has proven time and time again that it is incapable of attacking. Even in June, when the Russian army greatly outnumbered the Ukrainian army in artillery, and Ukraine was losing 300–500 soldiers a day, Russian troops advanced slowly and with huge losses.

  One of the reasons is their complete inability to coordinate between artillery and infantry. The correct approach is to hit enemy defenses with artillery and attack immediately after the hit. They managed to repeatedly fail even this simple coordination exercise.

  Recently, the Russian army even proved that it was not capable of organizing an effective defense. Yes, the mobilization will probably help with defense (with huge casualties), but it will hardly help with attack.

  Even if only the military remain in Ukraine and the entire civilian population is killed or flees the destroyed cities, the Russian army will not be able to win.

  I wouldn’t say that Ukraine has “irreversible momentum”, although it has an effective army that functions as it should. But it looks more like a complete lack of ability to achieve any momentum on the Russian side.

• Sausage Vector Machine 10 Oct 2022 15:21 UTC

  2 points

  1

  in reply to: sickpig’s comment on: Why I think there’s a one-in-six chance of an im­mi­nent global nu­clear war

  You are absolutely right. I missed this point in Max’s post.

  If Putin is lucky and the winter in Germany turns out to be especially cold, or if some of their gas storage facilities suddenly explode for some strange reason (I really hope they guard them better than their railways), Germany might have to make serious concessions and lift some of the sanctions in order to start using NS2. Putin will announce this as a huge victory, since launching NS2 was doubtful even before the war.

  That is exactly why Putin destroyed the other 3 pipelines. A simple resumption of the gas supply on NS1 would not have made it a big victory, because he himself had stopped it and it would not have been necessary to lift any sanctions to resume operation.

• Sausage Vector Machine 10 Oct 2022 14:45 UTC

  3 points

  1

  in reply to: mukashi’s comment on: Why I think there’s a one-in-six chance of an im­mi­nent global nu­clear war

  I don’t think their bombing of civilian infrastructure can be considered a military escalation.

  An official declaration of war and martial law in Russia would be an escalation (albeit not scary).

  Nuclear weapons testing would be an escalation (slightly scary, but still very far from an actual attack).

  Ordinary terrorism, by contrast, is just the default response, exactly the type of revenge for his favorite bridge that everyone expected from this particular dictator.

• Sausage Vector Machine 10 Oct 2022 7:56 UTC

  4 points

  1

  in reply to: CraigMichael’s comment on: Why I think there’s a one-in-six chance of an im­mi­nent global nu­clear war

  Good point, thanks. I’ve edited my comment.

• Sausage Vector Machine 9 Oct 2022 8:30 UTC

  13 points

  13

  in reply to: Jamie Cox’s comment on: Why I think there’s a one-in-six chance of an im­mi­nent global nu­clear war

  This is correct. They will always downplay Ukraine’s successes, explaining them as either terrorist attacks or “smoking accidents”. And yes, as Viliam says, Russia is now officially losing territory every day and nobody cares. They prefer to have ambiguous borders.

  To understand Putin, you have to imagine a petty crime boss who accidentally got big. He’s just a thug. “It’s not a bluff” is the telltale sign that his threat is a 100% 99.999999% bluff. When he plans to do something, he and his minions always deny it first. He will go to great lengths to continue without nuclear weapons.

  All of this means that the 30% chance of kaboom is way way too high. It’s currently closer to 3% or even 0.3%.

  Moreover, the probability that a strike will be ordered but not occur is quite high. You don’t just drop a nuke and hope it works. The technology is complex and there have been no nuclear tests since 1990. Russia’s impressive-looking nuclear arsenal would take a fortune to have been properly maintained all this time. A huge percentage of that money ended up in the pockets of the maintainers. This is the Russian Way. So either the technology can fail or, more likely still, people can do something wrong. Incompetence is the key word in the analysis of most Russian failures in this war so far.

  That said, he won’t back down and the probability of kaboom could increase with Putin’s desperation. There will be clear signs (e.g. new nuclear tests), so we’ll have time to prepare. But it’s still dangerous. The winning strategy for the West is to kill him. I hope the CIA, MI6 and other professionals understand this. They have to kill him one day and I really hope it happens soon enough.

• Sausage Vector Machine 15 May 2022 23:22 UTC

  1 point

  0

  on: What DALL-E 2 can and can­not do

  Awesome writeup!

  To further explore the interplay between style and content, how about trying something not very specific that could gain specificity from the style context?

  For example “Aliens are conducting experiments on human subjects”:

  • as a screenshot from South Park (will these mostly feature the anal probe?)

  • as a medieval painting (will these be mostly dissection?)

  • as a screenshot from the movie Prometheus (will these be too scary to look at?)

[Question] What are some fun ways to spend $100,000?