Posts
Comments
single-use
Considering how loog it took me to get that by this you mean "not dual-use", I expect some others just won't get it.
Some people who are very concerned about suffering might be considering building an unaligned AI that kills everyone just to avoid the risk of an AI takeover by an AI aligned to values which want some people to suffer.
Let this be me being on the record saying: I believe the probability of {alignment to values that strongly diswant suffering for all moral patients} is high enough, and the probability of {alignment to values that want some moral patients to suffer} is low enough, that this action is not worth it.
I think this applies to approximately anyone who would read this post, including heads of major labs in case they happen to read this post and in case they're pursuing the startegy of killing everyone to reduce S-risk.
See also: how acausal trade helps in 1, 2, but I think I think this even without acausal trade.
sigh I wish people realized how useless it is to have money when the singularity happens. Either we die or we get a utopia in which it's pretty unlikely that pre-singularity wealth matters. What you want to maximize is not your wealth but your utility function, and you sure as hell are gonna get more from LDT handshakes with aligned superintelligences in saved worlds, if you don't help OpenAI reduce the amount of saved worlds.
I believe that ChatGPT was not released with the expectation that it would become as popular as it did.
Well, even if that's true, causing such an outcome by accident should still count as evidence of vast irresponsibility imo.
I'm surprised at people who seem to be updating only now about OpenAI being very irresponsible, rather than updating when they created a giant public competitive market for chatbots (which contains plenty of labs that don't care about alignment at all), thereby reducing how long everyone has to solve alignment. I still parse that move as devastating the commons in order to make a quick buck.
I made guesses about my values a while ago, here.
but that this would be bad if the users aren't one of "us"—you know, the good alignment researchers who want to use AI to take over the universe, totally unlike those evil capabilities researchers who want to use AI to produce economically valuable goods and services.
Rather, " us" — the good alignment researchers who will be careful at all about the long term effects of our actions, unlike capabilities researchers who are happy to accelerate race dynamics and increase p(doom) if they make a quick profit out of it in the short term.
I am a utilitarian and agree with your comment.
The intent of the post was
- to make people weigh whether to publish or not, because I think some people are not weighing this enough
- to give some arguments in favor of "you might be systematically overestimating the utility of publishing", because I think some people are doing that
I agree people should take the utilitalianly optimal action, I just think they're doing the utilitarian calculus wrong or not doing the calculus at all.
I think research that is mostly about outer alignment (what to point the AI to) rather than inner alignment (how to point the AI to it) tends to be good — quantilizers, corrigibility, QACI, decision theory, embedded agency, indirect normativity, infra bayesianism, things like that. Though I could see some of those backfiring the way RLHF did — in the hands of a very irresponsible org, even not very capabilities-related research can be used to accelerate timelines and increase race dynamics if the org doing it thinks it can get a quick buck out of it.
I don't buy the argument that safety researchers have unusually good ideas/research compared to capability researchers at top labs
I don't think this particularly needs to be true for my point to hold; they only need to have reasonably good ideas/research, not unusually good, for them to publish less to be a positive thing.
That said, if someone hasn't thought at all about concepts like "differentially advancing safety" or "capabilities externalities," then reading this post would probably be helpful, and I'd endorse thinking about those issues.
That's a lot of what I intend to do with this post, yes. I think a lot of people do not think about the impact of publishing very much and just blurt-out/publish things as a default action, and I would like them to think about their actions more.
One straightforward alternative is to just not do that; I agree it's not very satisfying but it should still be the action that's pursued if it's the one that has more utility.
I wish I had better alternatives, but I don't. But the null action is an alternative.
It certainly is possible! In more decision-theoritic terms, I'd describe this as "it sure would suck if agents in my reference class just optimized for their own happiness; it seems like the instrumental thing for agents in my reference class to do is maximize for everyone's happiness". Which is probly correct!
But as per my post, I'd describe this position as "not intrinsically altruistic" — you're optimizing for everyone's happiness because "it sure would sure if agents in my reference class didn't do that", not because you intrinsically value that everyone be happy, regardless of reasoning about agents and reference classes and veils of ignorance.
decision theory is no substitute for utility function
some people, upon learning about decision theories such as LDT and how it cooperates on problems such as the prisoner's dilemma, end up believing the following:
my utility function is about what i want for just me; but i'm altruistic (/egalitarian/cosmopolitan/pro-fairness/etc) because decision theory says i should cooperate with other agents. decision theoritic cooperation is the true name of altruism.
it's possible that this is true for some people, but in general i expect that to be a mistaken analysis of their values.
decision theory cooperates with agents relative to how much power they have, and only when it's instrumental.
in my opinion, real altruism (/egalitarianism/cosmopolitanism/fairness/etc) should be in the utility function which the decision theory is instrumental to. i actually intrinsically care about others; i don't just care about others instrumentally because it helps me somehow.
some important aspects that my utility-function-altruism differs from decision-theoritic-cooperation includes:
- i care about people weighed by moral patienthood, decision theory only cares about agents weighed by negotiation power. if an alien superintelligence is very powerful but isn't a moral patient, then i will only cooperate with it instrumentally (for example because i care about the alien moral patients that it has been in contact with); if cooperating with it doesn't help my utility function (which, again, includes altruism towards aliens) then i won't cooperate with that alien superintelligence. corollarily, i will take actions that cause nice things to happen to people even if they've very impoverished (and thus don't have much LDT negotiation power) and it doesn't help any other aspect of my utility function than just the fact that i value that they're okay.
- if i can switch to a better decision theory, or if fucking over some non-moral-patienty agents helps me somehow, then i'll happily do that; i don't have goal-content integrity about my decision theory. i do have goal-content integrity about my utility function: i don't want to become someone who wants moral patients to unconsentingly-die or suffer, for example.
- there seems to be a sense in which some decision theories are better than others, because they're ultimately instrumental to one's utility function. utility functions, however, don't have an objective measure for how good they are. hence, moral anti-realism is true: there isn't a Single Correct Utility Function.
decision theory is instrumental; the utility function is where the actual intrinsic/axiomatic/terminal goals/values/preferences are stored. usually, i also interpret "morality" and "ethics" as "terminal values", since most of the stuff that those seem to care about looks like terminal values to me. for example, i will want fairness between moral patients intrinsically, not just because my decision theory says that that's instrumental to me somehow.
I would feel better about this if there was something closer to (1) on which to discuss what is probably the most important topic in history (AI alignment). But noted.
I'm generally not a fan of increasing the amount of illegible selection effects.
On the privacy side, can lesswrong guarantee that, if I never click on Recommended, then recombee will never see an (even anonymized) trace of what I browse on lesswrong?
Here the thing that I'm calling evil is pursuing short-term profits at the cost of non-negligeably higher risk that everyone dies.
Regardless of how good their alignment plans are, the thing that makes OpenAI unambiguously evil is that they created a strongly marketed public product and, as a result, caused a lot public excitement about AI, and thus lots of other AI capabilities organizations were created that are completely dismissive of safety.
There's just no good reason to do that, except short-term greed at the cost of higher probability that everyone (including people at OpenAI) dies.
(No, "you need huge profits to solve alignment" isn't a good excuse — we had nowhere near exhausted the alignment research that can be done without huge profits.)
There's also the case of harmful warning shots: for example, if it turns out that, upon seeing an AI do a scary but impressive thing, enough people/orgs/states go "woah, AI is powerful, I should make one!" or "I guess we're doomed anyways, might as well stop thinking about safety and just enjoy making profit with AI while we're still alive", to offset the positive effect. This is totally the kind of thing that could be the case in our civilization.
There could be a difference but only after a certain point in time, which you're trying to predict / plan for.
What you propose, ≈"weigh indices by kolmogorov complexity" is indeed a way to go about picking indices, but "weigh indices by one over their square" feels a lot more natural to me; a lot simpler than invoking the universal prior twice.
If you use the UTMs for cartesian-framed inputs/outputs, sure; but if you're running the programs as entire worlds, then you still have the issue of "where are you in time".
Say there's an infinitely growing conway's-game-of-life program, or some universal program, which contains a copy of me at infinitely many locations. How do I weigh which ones are me?
It doesn't matter that the UTM has a fixed amount of weight, there's still infinitely many locations within it.
Is quantum phenomena anthropic evidence for BQP=BPP? Is existing evidence against many-worlds?
Suppose I live inside a simulation ran by a computer over which I have some control.
-
Scenario 1: I make the computer run the following:
pause simulation if is even(calculate billionth digit of pi): resume simulationSuppose, after running this program, that I observe that I still exist. This is some anthropic evidence for the billionth digit of pi being even.
Thus, one can get anthropic evidence about logical facts.
-
Scenario 2: I make the computer run the following:
pause simulation if is even(calculate billionth digit of pi): resume simulation else: resume simulation but run it a trillion times slowerIf you're running on the non-time-penalized solomonoff prior, then that's no evidence at all — observing existing is evidence that you're being ran, not that you're being ran fast. But if you do that, a bunch of things break including anthropic probabilities and expected utility calculations. What you want is a time-penalized (probably quadratically) prior, in which later compute-steps have less realityfluid than earlier ones — and thus, observing existing is evidence for being computed early — and thus, observing existing is some evidence that the billionth digit of pi is even.
-
Scenario 3: I make the computer run the following:
pause simulation quantum_algorithm <- classical-compute algorithm which simulates quantum algorithms the fastest infinite loop: use quantum_algorithm to compute the result of some complicated quantum phenomena compute simulation forwards by 1 stepObserving existing after running this program is evidence that BQP=BPP — that is, classical computers can efficiently run quantum algorithms: if BQP≠BPP, then my simulation should become way slower, and existing is evidence for being computed early and fast (see scenario 2).
Except, living in a world which contains the outcome of cohering quantum phenomena (quantum computers, double-slit experiments, etc) is very similar to the scenario above! If your prior for the universe is a programs, penalized for how long they take to run on classical computation, then observing that the outcome of quantum phenomena is being computed is evidence that they can be computed efficiently.
-
Scenario 4: I make the computer run the following:
in the simulation, give the human a device which generates a sequence of random bits pause simulation list_of_simulations <- [current simulation state] quantum_algorithm <- classical-compute algorithm which simulates quantum algorithms the fastest infinite loop: list_of_new_simulations <- [] for simulation in list_of_simulations: list_of_new_simulations += [ simulation advanced by one step where the device generated bit 0, simulation advanced by one step where the device generated bit 1 ] list_of_simulations <- list_of_new_simulationsThis is similar to what it's like to being in a many-worlds universe where there's constant forking.
Yes, in this scenario, there is no "mutual destruction", the way there is in quantum. But with decohering everett branches, you can totally build exponentially many non-mutually-destructing timelines too! For example, you can choose to make important life decisions based on the output of the RNG, and end up with exponentially many different lives each with some (exponentially little) quantum amplitude, without any need for those to be compressible together, or to be able to mutually-destruct. That's what decohering means! "Recohering" quantum phenomena interacts destructively such that you can compute the output, but decohering* phenomena just branches.
The amount of different simulations that need to be computed increases exponentially with simulation time.
Observing existing after running this program is very strange. Yes, there are exponentially many me's, but all of the me's are being ran exponentially slowly; they should all not observe existing. I should not be any of them.
This is what I mean by "existing is evidence against many-worlds" — there's gotta be something like an agent (or physics, through some real RNG or through computing whichever variables have the most impact) picking a only-polynomially-large set of decohered non-compressible-together timelines to explain continuing existing.
Some friends tell me "but tammy, sure at step N each you has only 1/2^N quantum amplitude, but at step N there's 2^N such you's, so you still have 1 unit of realityfluid" — but my response is "I mean, I guess, sure, but regardless of that, step N occurs 2^N units of classical-compute-time in the future! That's the issue!".
Some notes:
-
I heard about pilot wave theory recently, and sure, if that's one way to get single history, why not. I hear that it "doesn't have locality", which like, okay I guess, that's plausibly worse program-complexity wise, but it's exponentially better after accounting for the time penalty.
-
What if "the world is just Inherently Quantum"? Well, my main answer here is, what the hell does that mean? It's very easy for me to imagine existing inside of a classical computation (eg conway's game of life); I have no idea what it'd mean for me to exist in "one of the exponentially many non-compressible-together decohered exponenially-small-amplitude quantum states that are all being computed forwards". Quadratically-decaying-realityfluid classical-computation makes sense, dammit.
-
What if it's still true — what if I am observing existing with exponentially little (as a function of the age of the universe) realityfluid? What if the set of real stuff is just that big?
Well, I guess that's vaguely plausible (even though, ugh, that shouldn't be how being real works, I think), but then the tegmark 4 multiverse has to contain no hypotheses in which observers in my reference class occupy more than exponentially little realityfluid.
Like, if there's a conway's-game-of-life simulation out there in tegmark 4, whose entire realityfluid-per-timestep is equivalent to my realityfluid-per-timestep, then they can just bruteforce-generate all human-brain-states and run into mine by chance, and I should have about as much probability of being one of those random generations as I'd have being in this universe — both have exponentially little of their universe's realityfluid! The conway's-game-of-life bruteforced-me has exponentially little realityfluid because she's getting generated exponentially late, and quantum-universe me has exponentially little realityfluid because I occupy exponentially little of the quantum amplitude, at every time-step.
See why that's weird? As a general observer, I should exponentially favor observing being someone who lives in a world where I don't have exponentially little realityfluid, such as "person who lives only-polynomially-late into a conway's-game-of-life, but happened to get randomly very confused about thinking that they might inhabit a quantum world".
Existing inside of a many-worlds quantum universe feels like aliens pranksters-at-orthogonal-angles running the kind of simulation where the observers inside of it to be very anthropically confused once they think about anthropics hard enough. (This is not my belief.)
I didn't see a clear indication in the post about whether the music is AI-generated or not, and I'd like to know; was there an indication I missed?
(I care because I'll want to listen to that music less if it's AI-generated.)
Unlike on your blog, the images on the lesswrong version of this post are now broken.
Taboo the word "intelligence".
An agent can superhumanly-optimize any utility function. Even if there are objective values, a superhuman-optimizer can ignore them and superhuman-optimize paperclips instead (and then we die because it optimized for that harder than we optimized for what we want).
(I'm gonna interpret these disagree-votes as "I also don't think this is the case" rather than "I disagree with you tamsin, I think this is the case".)
I don't think this is the case, but I'm mentioning this possibility because I'm surprised I've never seen someone suggest it before:
Maybe the reason Sam Altman is taking decisions that increase p(doom) is because he's a pure negative utilitarian (and he doesn't know-about/believe-in acausal trade).
For writing, there's also jan misali's ASCII toki pona syllabary.
Reposting myself from discord, on the topic of donating 5000$ to EA causes.
if you're doing alignment research, even just a bit, then the 5000$ are probly better spent on yourself
if you have any gears level model of AI stuff then it's better value to pick which alignment org to give to yourself; charity orgs are vastly understaffed and you're essentially contributing to the "picking what to donate to" effort by thinking about it yourself
if you have no gears level model of AI then it's hard to judge which alignment orgs it's helpful to donate to (or, if giving to regranters, which regranters are good at knowing which alignment orgs to donate to)
as an example of regranters doing massive harm: openphil gave 30M$ to openai at a time where it was critically useful to them, (supposedly in order to have a chair on their board, and look how that turned out when the board tried to yeet altman)
i know of at least one person who was working in regranting and was like "you know what i'd be better off doing alignment research directly" — imo this kind of decision is probly why regranting is so understaffed
it takes technical knowledge to know what should get money, and once you have technical knowledge you realize how much your technical knowledge could help more directly so you do that, or something
yes, edited
So this option looks unattractive if you think transformative AI systems are likely to developed within the next 5 years. However, with a 10-years timeframe things look much stronger: you would still have around 5 years to contribute as a research.
This phrasing is tricky! If you think TAI is coming in approximately 10 years then sure, you can study for 5 years and then do research for 5 years.
But if you think TAI is coming within 10 years (for example, if you think that the current half-life on worlds surviving is 10 years; if you think 10 years is the amount of time in which half of worlds are doomed) then depending on your distribution-over-time you should absolutely not wait 5 years before doing research, because TAI could happen in 9 years but it could also happen in 1 year. If you think TAI is coming within 10 years, then (depending on your distribution) you should still in fact do research asap.
(People often get this wrong! They think that "TAI probably within X years" necessarily means "TAI in approximately X years".)
Sure, this is just me adapting the idea to the framing people often have, of "what technique can you apply to an existing AI to make it safe".
AI safety is easy. There's a simple AI safety technique that guarantees that your AI won't end the world, it's called "delete it".
AI alignment is hard.
I'm confused about why 1P-logic is needed. It seems to me like you could just have a variable X which tracks "which agent am I" and then you can express things like sensor_observes(X, red) or is_located_at(X, northwest). Here and Absent are merely a special case of True and False when the statement depends on X.
Moral patienthood of current AI systems is basically irrelevant to the future.
If the AI is aligned then it'll make itself as moral-patient-y as we want it to be. If it's not, then it'll make itself as moral-patient-y as maximizes its unaligned goal. Neither of those depend on whether current AI are moral patients.
If my sole terminal value is "I want to go on a rollercoaster", then an agent who is aligned to me would have the value "I want Tamsin Leake to go on a rollercoaster", not "I want to go on a rollercoaster myself". The former necessarily-has the same ordering over worlds, the latter doesn't.
- I think the term "conscious" is very overloaded and the source of endless confusion and should be tabood. I'll be answering as if the numbers are not "probability(-given-uncertainty) of conscious" but "expected(-given-uncertainty) amount of moral patienthood", calibrated with 1 meaning "as much as a human" (it could go higher — some whales have more neurons/synapses than humans and so they might plausibly be more of a moral patient than humans, in the sense that in a trolley problem you should prefer to save 1000 such whales to 1001 humans).
- Besides the trivia I just mentioned about whales, I'm answering this mostly on intuition, without knowing off the top of my head (nor looking up) the amount of neurons/synapses. Not to imply that moral patienthood is directly linear to amount of neurons/synapses, but I expect that that amount probably matters to my notion of moral patienthood.
- I'll also assume that everyone has a "normal amount of realityfluid" flowing through them (rather than eg being simulated slower, or being fictional, or having "double-thick neurons made of gold" in case that matters).
First list: 1, 1, 1, .7, 10⁻², 10⁻³, 10⁻⁶, 10⁻⁶, 10⁻⁸, ε, ε, ε, ε, ε.
Second list: .6, .8, .7, .7, .6, .6, .5, ε, ε, ε, ε.
Edit: Thinking about it more, something feels weird here, like these numbers don't track at all "how many of these would make me press the lever on the trolley problem vs 1 human" — for one, killing a sleeping person is about as bad as killing an awake person because like the sleeping person is a temporarily-paused-backup for an awake person. I guess I should be thinking about "the universe has budget for one more hour of (good-)experience just before heat death, but it needs to be all same species, how much do I value each?" or something.
If you start out with CDT, then the thing you converge to is Son of CDT rather than FDT.
(that arbital page takes a huge amount of time to load for me for some reason, but it does load eventually)
And I could totally see the thing that kills us {being built with} or {happening to crystallize with} CDT rather than FDT.
We have to actually implement/align-the-AI-to the correct decision theory.
By thinking about each other's source code, FAI and Clippy will be able to cooperate acausally like Alice and Bob, each turning their future lightcone into 10% utopia, 90% paperclips. Therefore, we get utopia either way! :D
So even if we lose we win, but even if we win we lose. The amount of utopiastuff is exactly conserved, and launching unaligned AI causes timelines-where-we-win to have less utopia by exactly as much as our timeline has more utopia.
The amount of utopiastuff we get isn't just proportional to how much we solve alignment, it's actually back to exactly equal.
See also: Decision theory does not imply that we get to have nice things.
I think it's exceedingly unlikely (<1%) that we robustly prevent anyone from {making an AI that kills everyone} without an aligned sovereign.
I continue to think that, in worlds where we robustly survive, money is largely going to be obsolete. The thing that maximizes the terminal values of the kind of (handshake of) utility functions we can expect probably aren't maximized by maintaining current allocations of wealth and institutions-that-care-about-that-wealth. The use for money/investment/resources is making sure we get utopia in the first place, by slowing capabilities and solving alignment (and thus also plausibly purchasing shares of the LDT utility function handshake), not being rich in utopia. (maybe see also 1, 2)
‘high level machine intelligence’ (HLMI) and ‘full automation of labor’ (FAOL)
I continue to believe that predicting things like that is not particularly useful to predicting when AI will achieve decisive strategic advantage and/or kill literally everyone. AI could totally kill literally everyone without us ever getting to observe HLMI or FAOL first, and I think development in HLMI / FAOL does not say much about how close we are to AI that kills literally everyone.
Both are possible. For theoretical examples, see the stamp collector for consequentialist AI and AIXI for reward-maximizing AI.
What kind of AI are the AIs we have now? Neither, they're not particularly strong maximizers. (if they were, we'd be dead; it's not that difficult to turn a powerful reward maximizer into a world-ending AI).
If the former, I think this makes alignment much easier. As long as you can reasonably represent “do not kill everyone”, you can make this a goal of the AI, and then it will literally care about not killing everyone, it won’t just care about hacking its reward system so that it will not perceive everyone being dead.
This would be true, except:
- We don't know how to represent "do not kill everyone"
- We don't know how to pick which quantity would be maximized by a would-be strong consequentialist maximizer
- We don't know know what a strong consequentialist maximizer would look like, if we had one around, because we don't have one around (because if we did, we'd be dead)
The first one. Alice fundamentally can't fully model Bob because Bob's brain is as large as Alice's, so she can't fit it all inside her own brain without simply becoming Bob.
I remember a character in Asimov's books saying something to the effect of
It took me 10 years to realize I had those powers of telepathy, and 10 more years to realize that other people don't have them.
and that quote has really stuck with me, and keeps striking me as true about many mindthings (object-level beliefs, ontologies, ways-to-use-one's-brain, etc).
For so many complicated problem (including technical problems), "what is the correct answer?" is not-as-difficult to figure out as "okay, now that I have the correct answer: how the hell do other people's wrong answers mismatch mine? what is the inferential gap even made of? what is even their model of the problem? what the heck is going on inside other people's minds???"
Answers to technical questions, once you have them, tend to be simple and compress easily with the rest of your ontology. But not models of other people's minds. People's minds are actually extremely large things that you fundamentally can't fully model and so you're often doomed to confusion about them. You're forced to fill in the details with projection, and that's often wrong because there's so much more diversity in human minds than we imagine.
The most complex software engineering projects in the world are absurdly tiny in complexity compared to a random human mind.
I don't think it's a binary; they could still pay less attention!
(plausibly there's a bazillion things constantly trying to grab their attention, so they won't "lock on" if we avoid bringing AI to their attention too much)
You might want to read this post (it's also on lesswrong but the images are broken there)
(to be clear: this is more an amusing suggestion than a serious belief)
By "vaguely like dath ilan" I mean the parts that made them be the kind of society that can restructure in this way when faced with AI risk. Like, even before AI risk, they were already very different from us.
I'm pretty sure we just need one resimulation to save everyone; once we have located an exact copy of our history, it's cheap to pluck out anyone (including people dead 100 or 1000 years ago). It's a one-time cost.
Lossy resurrection is better than nothing but it doesn't feel as "real" to me. If you resurrect a dead me, I expect that she says "I'm glad I exist! But — at least as per my ontology and values — you shouldn't quite think of me as the same person as the original. We're probly quite different, internally, and thus behaviorally as well, when ran over some time."
Like, the full-history resimulation will surely still not allow you to narrow things down to one branch. You'd get an equivalence class of them, each of them consistent with all available information. Which, in turn, would correspond to a probability distribution over the rescuee's mind; not a unique pick.
I feel like I'm not quite sure about this? It depends on what quantum mechanics entails, exactly, I think. For example: if BQP = P, then there's "only a polynomial amount" of timeline-information (whatever that means!), and then my intuition tells me that the "our world serves as a checksum for the one true (macro-)timeline" idea is more likely to be a thing. But this reasoning is still quite heuristical. Plausibly, yeah, the best we get is a polynomially large or even exponentially large distribution.
That said, to get back to my original point, I feel like there's enough unknowns making this scenario plausible here, that some people who really want to get reunited with their loved ones might totally pursue aligned superintelligence just for a potential shot at this, whether their idea of reuniting requires lossless resurrection or not.