"AI: Forecasters on the community forecasting platform Metaculus think that artificial intelligent systems that are better than humans at all relevant tasks will be created in 2042."
How do you get this from the questions' operationalization?
Is there a description of the desk-reject policy and/or statistics on how many applications were desk rejected?
It does not really seem to address the reasoning from my second paragraph. You say:
"Similarly, if people refused to consume any goods or services that were associated with net-positive greenhouse gas emissions, then those industries would rapidly decarbonize or go out of business.",
but it seems to me that this would be way more costly for individuals than giving up on meat, in addition to leading to way larger economic damage in the short to medium term (without enough time for investments into replacement technologies to pay off).
There seems to be a clear disanalogy in that if every individual stopped eating meat tomorrow, factory farming would be history very quickly. On the other hand, if everyone tried very hard to reduce their personal CO2 consumption, the effect seems more limited (unless people are really serious about it, in which case this would probably lead to major economic damage).
The key difference seems to be that CO2 emissions are embedded in our current societies and economies in such a deep way, that we can only get out via long-term investment into rep...
Thank you!
5% does sound very alarming to me, and is definitely a lot higher than I would have said at the beginning of the crisis (without having thought about it much, then).
Also, beyond the purely personal, are there any actions that could be taken by individuals right now that would have a positive impact on humanity's chances to recover, conditional on nuclear war?
Some (probably naive) ideas:
Putin seems to have ordered deterrence forces (which include nuclear arms) to be on high alert, roughly an hour ago. https://www.reuters.com/world/europe/biden-says-russian-attack-ukraine-unfolding-largely-predicted-2022-02-24/
Can someone weigh in about how unprecedented this is? Some media coverage has compared the severity of the current situation to the Cuba Crisis, which would be extremely alarming if remotely true.
I am starting to worry that the possibility of Russia using conventional or perhaps more likely tactical nuclear weapons in the Ukraine conflict is real. My concern is one largely based in this article by Francesca Giovannini : A hurting stalemate? The risks of nuclear weapon use in the Ukraine crisis.
For Mr. Putin, any kind of losing the war with Ukraine seems like a non-option, given his domestic situation and the possibility that his regime could come to an end.
The article outlines three assumptions that those who don't think Russia w...
I'm not overly concerned with the news from this morning. In fact I expected them to raise the nuclear force readiness prior to or simultaneously to commencing the invasion, not now, which is expected going into a time of conflict/high tension from normal peacetime readiness. I had about a 5% chance this will escalate to a nuclear war going into it, and it's not much different now, certainly not above 10% (For context, my odds of escalation to full countervalue exchange in a US intervention in a Taiwan reunification campaign would be about 75%). Virtually ...
Miscalibration might cut both ways...
On one hand, It seems quite plausible for forecasts like this to usually be underconfident about the likelihood of the null event, but on the other hand recent events should probably have substantially increased forecasters' entropy for questions around geopolitical events in the next few days and weeks.
(This risk is a greater risk than the risk of >50% of the public advocating for unethical policies out of self-interest, because in expectation, unethical policies in the self-interest of ">50% of the public" would be good for more people than unethical policies in the self-interest of experts)
This seems to have a bunch of hidden assumptions, including both about the relative capabilities of experts vs. the public to assess the effects of policies, as well as about the distribution of potential policies: While constitutions are not really a technocra...
Do you have thoughts on how potentially rising inflation could affect emission pathways and the relative cost of renweables? I have heard the argument that associated rises in the cost of capital could be pretty bad, because most costs associated with renewables are capital costs, while fuel costs dominate for fossil energy.
Huh? I did not like the double-page style for the non-mobile pdf, as it required some manual rescaling on my PC.
And the mobile version has the main table cut between two pages in a pretty horrible way. I think I would have much preferred a single pdf in the mobile/single page style that is actually optimized for that style, rather than this.
Maybe I should have used the HTML version instead?
More detailed action points on safety from page 32:
The Office for AI will coordinate cross-government processes to accurately assess long term AI safety and risks, which will include activities such as evaluating technical expertise in government and the value of research infrastructure. Given the speed at which AI developments are impacting our world, it is also critical that the government takes a more precise and timely approach to monitoring progress on AI, and the government will work to do so.
The government will support the safe and ethic...
I don't think I get your argument for why the approximation should not depend on the downstream task. Could you elaborate?
I am also a bit confused about the relationship between spread and resiliency: a larger spread of forecasts does not seem to necessarily imply weaker evidence: It seems like for a relatively rare event about which some forecasters could acquire insider information, a large spread might give you stronger evidence.
Imagine is about the future enactment of a quite unusual government policy, and one of your forecaster...
This seems to connect to the concept of - means: If the utility for an option is proportional to , then the expected utility of your mixture model is equal to the expected utility using the -mean of the expert's probabilities and defined as , as the in the utility calculation cancels out the . If I recall correctly, all aggregation functions that fulfill some technical conditions on a generalized mean can be written as a -mean.
In the first example,&n...
I wanted to flag that many PhD programs in Europe might require you to have a Master's degree, or to essentially complete the coursework for Master's degree during your PhD (as seems to be the case in the US), depending on the kind of undergraduate degree you hold. Obviously, the arguments regarding funding might still partially hold in that case.
Do you have a specific definition of AI Safety in mind? From my (biased) point of view, it looks like large fractions of work that is explicitly branded "AI Safety" is done by people who are at least somewhat adjacent to the EA community. But this becomes a lot less true if you widen the definition to include all work that could be called "AI Safety" (so anything that could conceivably help with avoiding any kind of dangerous malfunction of AI systems, including small scale and easily fixable problems).
Relatedly, what is the likelihood that future iterations of the fellowship might be less US-centric, or include Visa sponsorship?
A large portion of the value from programs like this comes from boosting fellows into career paths where they spend at least some time working in the US government, and many of the most impactful government roles require US citizenship. We are therefore mainly focused on people who have (a plausible pathway to) citizenship and are interested in US government work. Legal and organizational constraints means it is unlikely that we will be able to sponsor visas even if we run future rounds.
This program is US-based because the US government is especially impor...
The job posting states:
"All participants must be eligible to work in the United States and willing to live in Washington, DC, for the duration of their fellowship. We are not able to sponsor US employment visas for participants; US permanent residents (green card holders) are eligible to apply, but fellows who are not US citizens may be ineligible for placements that require a security clearance."
So my impression would be that it would be pretty difficult to participate for non-US citizens who do not already live in the US.
https://en.wikipedia.org/wiki/Technological_transitions might be relevant.
The Geels book cited in the article (Geels, F.W., 2005. Technological transitions and system innovations. Cheltenham: Edward Elgar Publishing.) has a bunch of interesting case studies I read a while ago and a (I think popular) framework for technological change, but I am not sure the framework is sufficiently precise to be very predictive (and thus empirically validatable).
I don't have any particular sources on this, but the economic literature on the effects of regulation migh...
Facebook has at least experimented with using deep reinforcement learning to adjust its notifications according to https://arxiv.org/pdf/1811.00260.pdf . Depending on which exact features they used for the state space (i.e. if they are causally connected to preferences), the trained agent would at least theoretically have an incentive to change user's preferences.
The fact that they use DQN rather than a bandit algorithm seems to suggest that what they are doing involves at least some short term planning, but the paper does not seem to analyze the exp...
Interesting writeup!
Depending on your intended audience, it might make sense to add more details for some of the proposals. For example, why is scenario planning a good idea compared to other methods of decision making? Is there a compelling story, or strong empirical evidence for its efficacy?
Some small nitpicks:
There seems to be a mistake here:
"Bostrom argues in The Fragile World Hypothesis that continuous technological development will increase systemic fragility, which can be a source of catastrophic or existential risk. In the Precip...
So for the maximin we are minimizing over all joint distributions that are -close to our initial guess?
"One intuitive way to think about this might be considering circles of radius centered around fixed points, representing your first guesses for your options, in the plane. As becomes very large, the intersection of the interiors of these circles will approach 100% of their interiors. The distance between the centres becomes small relative to their radii. Basically, you can't tell the options apart anymore for h...
I wrote up my understanding of Popper's argument on the impossibility of predicting one's own knowledge (Chapter 22 of The Open Universe) that came up in one of the comment threads. I am still a bit confused about it and would appreciate people pointing out my misunderstandings.
Consider a predictor:
A1: Given a sufficiently explicit prediction task, the predictor predicts correctly
A2: Given any such prediction task, the predictor takes time to predict and issue its reply (the task is only completed once the reply is issued).
T1: A1,A2=> Given a self-predi...
They are, but I don't think that the correlation is strong enough to invalidate my statement. P(sun will exist|AI risk is a big deal) seems quite large to me. Obviously, this is not operationalized very well...
It seems like the proof critically hinges on assertion 2) which is not proven in your link. Can you point me to the pages of the book that contain the proof?
I agree that proofs are logical, but since we're talking about probabilistic predictions, I'd be very skeptical of the relevance of a proof that does not involve mathematical reasoning,
I don't think I buy the impossibility proof as predicting future knowledge in a probabilistic manner is possible (most simply, I can predict that if I flip a coin now, that there's a 50/50 chance I'll know the coin landed on heads/tails in a minute). I think there is some important true point behind your intuition about how knowledge (especially of more complex form than about a coin flip) is hard to predict, but I am almost certain you won't be able to find any rigorous mathematical proof for this intuition because reality is very fuzzy (in a ...
Ok, makes sense. I think that our ability to make predictions about the future steeply declines with increasing time horizions, but find it somewhat implausible that it would become entirely uncorrelated with what is actually going to happen in finite time. And it does not seem to be the case that data supporting long term predictions is impossible to get by: while it might be pretty hard to predict whether AI risk is going to be a big deal by whatever measure, I can still be fairly certain that the sun will exist in a 1000 years; in part due to a lot of data collection and hypothesis testing done by physicist.
"The "immeasurability" of the future that Vaden has highlighted has nothing to do with the literal finiteness of the timeline of the universe. It has to do, rather, with the set of all possible futures (which is provably infinite). This set is immeasurable in the mathematical sense of lacking sufficient structure to be operated upon with a well-defined probability measure. "
This claim seems confused, as every nonempty set allows for the definition of a probability measure on it and measures on function spaces exist ( https://en.wikipedia.org/wiki/Dir...
I am confused about the precise claim made regarding the Hilbert Hotel and measure theory. When you say "we have no measure over the set of all possible futures", do you mean that no such measures exist (which would be incorrect without further requirements: https://en.wikipedia.org/wiki/Dirac_measure , https://encyclopediaofmath.org/wiki/Wiener_measure ), or that we don't have a way of choosing the right measure? If it is the latter, I agree that this is an important challenge, but I'd like to highlight that the situati...
I'm also not sure I follow your exact argument here. But frequency clearly matters whenever the forecast is essentially resolved before the official resolution date, or when the best forecast based on evidence at time t behaves monotonically (think of questions of the type "will event Event x that (approximately) has a small fixed probability of happening each day happen before day y?", where each day passing without x happening should reduce your credence).
I guess you're right (I read this before and interpreted "active foreast" as "forecast made very recently").
If they also used this way of scoring things for the results in Superforecasting, this seems like an important caveat for forecasting advice that is derived from the book: For example the efficacy of updating your beliefs might mostly be explained by this. I previously thought that the results meant that a person who forecasts a question daily will make better forecasts on sundays than a person who only forecasts on sundays.
Do you have a source for the "carrying forward" on gjopen? I usually don't take the time to update my forecasts if I don't think I'd be able to beat the current median but might want to adjust my strategy in light of this.
Claims that people are "unabashed racists and sexists" should at least be backed up with actual examples. Like this, I cannot know whether you have good reasons for that believe that I don't see (to the very least not in all of the cases), or whether we have the same information but fundamentally disagree about what constitutes "unabashed racism".
I agree with the feeling that the post undersells concerns about the right wing, but I don't think you will convince anybody without any arguments except for a weakly supported claim...
"While Trump’s policies are in some ways more moderate than the traditional Republican platform". I do not find this claim self-evident (potentially due to biased media reporting affecting my views) and find it strange that no source or evidence for it is provided, especially given the commendable general amount of links and sources in the text.
Relatedly, I noticed a gut feeling that the text seems more charitable to the right-wing perspective than to the left (specific "evidence" included the statement from the previous paragra...
If you go by GDP per capita, most of europe is behind the US but ahead of most of Asia https://en.wikipedia.org/wiki/List_of_countries_by_GDP_(nominal)_per_capita (growth rates in Asia are higher though, so this might change at some point in the future.)
In terms of the Human Develompment Index https://en.wikipedia.org/wiki/List_of_countries_by_Human_Development_Index (which seems like a better measure of "success" than just GDP), some countries (including large ones like Germany and the UK) score above the US but others score lower. Most of Asi...
While I am unsure about how good of an idea it is to map out more plausible scenarios for existential risk from pathogens, I agree with the sentiment that the top level post seems seems to focus too narrowly on a specific scenario.
Re bonus section: Note that we are (hopefully) taking expectations over our estimates for importance, neglectedness and tractability, such that general correlations between the factors between causes do not necessarily cause a problem. However, it seems quite plausible that our estimation errors are often correlated because of things like the halo effect.
Edit: I do not fully endorse this comment any more, but still belief that the way we model the estimation procedure matters here. Will edit again, once I am less confused.
Maybe having a good understanding of Quantum Computing and how it could be leveraged in different paradigms of ML might help with forecasting AI-timelines as well as dominant paradigms, to some extend?
If that was true, while not necessarily helpful for a single agenda, knowledge about quantum computing would help with the correct prioritization of different agendas.
"The combination of these vastly different expressions of scale together with anchoring makes that we should expect people to over-estimate the probability of unlikely risks and hence to over-estimate the expected utility of x-risk prevention measures. "
I am not entirely sure whether i understand this point. Is the argument that the anchoring effect would cause an overestimation, because the "perceived distance" from an anchor grows faster per added zero than per increase of one to the exponent?
Directly relevant quotes from the articles for easier reference:
Paul Christiano:
"This story seems consistent with the historical record. Things are usually preceded by worse versions, even in cases where there are weak reasons to expect a discontinuous jump. The best counterexample is probably nuclear weapons. But in that case there were several very strong reasons for discontinuity: physics has an inherent gap between chemical and nuclear energy density, nuclear chain reactions require a large minimum scale, and the dynamics of war are very ...
Another point against the content overhang argument: While more data is definitely useful, it is not clear, whether raw data about a world without a particular agent in it will be similarly useful to this agent as data obtained from its own (or that of sufficiently similar agents) interaction with the world. Depending on the actual implementation of a possible superintelligence, this raw data might be marginally helpful but far from being the most relevant bottleneck.
"Bostrom is simply making an assumption that such rapid rates of progress could occur...
Thanks for writing this!
I think you are pointing out some important imprecisions, but i think that some of your arguments aren't as conclusive as you seem to present them to be:
"Bostrom therefore faces a dilemma. If intelligence is a mix of a wide range of distinct abilities as in Intelligence(1), there is no reason to think it can be ‘increased’ in the rapidly self-reinforcing way Bostrom speaks about (in mathematical terms, there is no single variable which we can differentiate and plug into the differential equation, as Bostrom does in hi...
Very interesting!
In the your literature review you summarize the Smith and Winkler (2006) paper as "Prove that nonrandom, non-Bayesian decision strategies systematically overestimate the value of the selected option."
On first sight, this claim seems like it might be stronger than the claim i have taken away from the paper (which is similar to what you write later in the text): if your decision strategy is to just choose the option you (naively) expect to be best, you will systematically overestimate the value of the selected option.
If you think t...
I think that the assumption of the existence of a Funnel shaped distribution with undefined expected value of things we care about is quite a bit stronger than assuming that there are infinitely many possible outcomes.
But even if we restrict ourselves to distributions with finite expected value, our estimates can still fluctuate wildly until we have gathered huge amounts of evidence.
So while i am sceptical of the assumption that there exists a sequence of world states with utilities tending to infinity and even more sceptical of extremely high/low utility ...
I think the argument is that additional information showing that a cause has high marginal impact might divert causes away towards it from causes with less marginal impact. And getting this kind of information does seem more likely for causes without a track record allowing for a somewhat robust estimation of their (marginal) impact.
I think this strongly depends on how much weight you expect forcasters on metaculus to put onto the actual operationalization rather than the question's "vibe". I personally expect quite a bit of weight on the exact operationalization, so I am generally not very happy with how people have been talking about this specific forecast (the term "AGI" often seems to invoke associations that are not backed by the forecast's operationalization), and would prefer a more nuanced statement in the report.
(Note, that you might believe that the gap between the res... (read more)