Claims

2,653 claims by 33 public figures and organisations, checked against the public record. Each has its quote, source, verdict and evidence. Read them without a key; browse by person.

18,133 claims · page 97 of 182

18,133 claims, sorted by newest said. Choose a claim to read it in full.
VerdictPersonKindSaidClaimTopic
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 The second thing looks unworkable (less so than CEV, but still lethally unworkable) because corrigibility runs actively counter to… Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 I figured this stuff out using the null string as input Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 Reading this document cannot make somebody a core alignment researcher. Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 I have a hard time myself feeling hopeful about getting real alignment work out of somebody who previously sat around waiting for somebody… Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 This is sufficient on its own, even ignoring many other items on this list, to trash entire categories of naive alignment proposals which… Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 The first thing generally, or CEV specifically, is unworkable because the complexity of what needs to be aligned or meta-aligned for our… Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 So most organizations don’t have plans, because I haven’t taken the time to personally yell at them. Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 Human beings cannot inspect an AGI’s output to determine whether the consequences will be good. Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 This field is not making real progress and does not have a recognition function to distinguish real progress if it took place. Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 Corrigibility is anti-natural to consequentialist reasoning Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 This is a very lethal problem, it has to be solved one way or another, it has to be solved at a minimum strength and difficulty level… Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 When you explicitly optimize against a detector of unaligned thoughts, you’re partially optimizing for more aligned thoughts, and partially… Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 But I’d have more hope – not significant hope, but more hope – in separating the concerns of (a) credibly promising to pay big money… Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 You don’t get any cynical old veterans, in this case, because everybody on Earth is dead. Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 when it comes to a choice of utility function, there are unbounded degrees of freedom and multiple reflectively coherent fixpoints. Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 So far as I’m concerned, if you can get a powerful AGI that carries out some pivotal superhuman engineering task, with a less than fifty… Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 It does not appear to me that the field of ‘AI safety’ is currently being remotely productive on tackling its enormous lethal problems. Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 Nobody knows what the hell GPT-3 is thinking Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 Anticipating all of the really dangerous ones, and then successfully materializing them, in the correct form for early solutions to… Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 Human beings can figure out pretty difficult things over time, when they get lots of tries; when a failed guess kills literally everyone… Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 If all you need is an object that doesn’t do dangerous things, you could try a sponge; a sponge is very passively safe. Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 It’s not like people will yell at themselves about prospective alignment difficulties, they don’t have an internal voice of caution. Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 I have several times failed to write up a well-organized list of reasons why AGI will kill you. Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 If you perfectly learn and perfectly maximize the referent of rewards assigned by human operators, that kills them. Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 Surviving worlds, by this point, and in fact several decades earlier, have a plan for how to survive. Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 And become that cynical old veteran right away, before reality whaps you upside the head in the form of everybody dying and you not getting… Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 All of these kill you if optimized-over by a sufficiently powerful intelligence Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 The best and easiest-found-by-optimization algorithms for solving problems we want an AI to solve, readily generalize to problems we’d… Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 There is no pivotal output of an AGI that is humanly checkable and can be used to safely save the world but only after checking it Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 We’ve got no idea what’s actually going on inside the giant inscrutable matrices and tensors of floating-point numbers. Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 The big ask from AGI alignment, the basic challenge I am saying is too difficult, is to obtain by any strategy whatsoever a significant… Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 I do not know how to train it into others. Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 Therefore, capabilities generalize further out-of-distribution than alignment, once they start to generalize at all. Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 This general problem is a fact about the territory, not the map Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Eliezer Yudkowsky Not checkable 10 Jun 2022 They probably do not know where the real difficulties are, they probably do not understand what needs to be done, they cannot tell the… Eliezer Yudkowsky · Not checkable · 10 Jun 2022
Not checkable Nate Silver Not checkable 9 Jun 2022 The problem with a lot of Democratic messaging about Jan. 6 is that they want to persuade people that it was uniquely bad but then they… Nate Silver · Not checkable · 9 Jun 2022
Not checkable Nate Silver Not checkable 8 Jun 2022 "Crime decreased, other than murder" is sort of a literal "Other than that, Mrs. Lincoln—". Nate Silver · Not checkable · 8 Jun 2022
Not checkable Nate Silver Not checkable 6 Jun 2022 "Moderate" and "centrist" are reasonable terms to describe this behavior. Nate Silver · Not checkable · 6 Jun 2022
Not checkable Nate Silver Not checkable 5 Jun 2022 One of the lowkey worst arguments is when someone complains that society is handling new emerging thing A inconsistently with established… Nate Silver · Not checkable · 5 Jun 2022
Not checkable Nate Silver Not checkable 4 Jun 2022 IDK on Twitter I'm exposed to "extreme" left-wing and "extreme" right-wing views more often than IRL, so there's a greater diversity of… Nate Silver · Not checkable · 4 Jun 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 And actually, to pass the Turing test, meaning it’s equivalent to humans, we’re actually going to have to dumb them down, because if it… Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 In terms of what we’re using, computers are not that expensive. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 We actually need more intelligent weapons on the side of truth and so on. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 you actually have to dumb it down because if it showed its fantastic knowledge of every different field, you’d know it’s a machine. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 So it’s really a writing tool. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 I see no reason to increase my estimate. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 the one reason I think we’re living in a simulation is this is such the most extraordinary time ever to be alive Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 that’s one of the requirements of being able to connect to the cloud is that this is your portion of the cloud and other people can’t… Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 No human being today can do that. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 I think people will be OK, even if they choose the wrong thing, because it’s going to be a lot of money. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 So the future of human history is not set. I’m optimistic about it. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 Again, these computers is not something we need to compete with. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 I mean, large language models, it’s quite consistent with what I’ve said, but I’m still amazed by it. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 So the way we’ve used technology has been very beneficial to human beings so far. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 They can actually measure where a student is and help them to learn, overcome their barriers. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 I am optimistic, but I’m also worried about it. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 It is democratizing. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 I mean, we create these machines to make ourselves smarter. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 Meaning, it will make us to be better humans. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 But we use the same kind of connections to deal with emotional intelligence, know how to react to someone else and get into their frame of… Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 If your heart is in it, you’re likely to do more creative work and appreciate it more. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 But I don’t think we would want to get rid of AI. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 But people are very eager to retire because they really don’t like their jobs. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 So it’s actually a philosophical question. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 Its benefits are already quite enormous. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 the answer is no unless they’re in horrible pain. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 It is not just AI. AI is part of humanity. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 So, we’re gonna see tremendous progress by the end of this decade Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 I’m a little bit skeptical of quantum computing. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 we can have sensors that are even thousands of miles away that are really part of who we are. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 I feel that I actually retired when I was five because I decided to be an inventor. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 You want to be courageous. You want to follow your dreams despite the skeptics. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 So we have to be very mindful of avoiding these types of perils. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 Human trials are slow, they’re risky, they’re expensive, they take a long time. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 One of the biggest applications of LLM is to help education. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 But we really can’t predict what these large language models will do. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 Well, right now we have really just one body we can be. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 Because I deal with human beings and I can’t always account for what they might be doing. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 I think large language models, because it actually incorporates all of the learning of humans, we can actually begin to appreciate that. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 And most of the things that are coming out are positive for humanity. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 So I think that’s a good thing because I think information is extremely important. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 I mean, every six months, it’s going to be a whole new world. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 It is something that humans have created. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 But if you actually look at history, you see it’s constantly changing. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 I think that’s where feelings come from. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 it’s actually a good thing that these are decentralized, so you’re not giving too much power to large companies. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 Now things are happening every month. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 And the bad news is true. But we completely ignore the good news. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 I think we need a little bit more work, for example, on virtual reality. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 It knows everything that anybody has ever dealt with. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 I really believe we shouldn’t do that anymore because there’s nothing to do with Moore’s. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 We’re exactly where I would expect to be in 2022 to meet the Turing test by 2029. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 But someone could actually design that. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 And it’s actually an optimistic book. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 I think we have the correct methods: neural nets. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 You want to see an exponential future. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 So I think we have to actually export certain values. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 So I think it’s a manageable problem, at least by the big company, large language models. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 But these large language models, this is not an alien invasion of intelligent machines coming from Mars. Ray Kurzweil · Not checkable · 26 May 2022
Not checkable Ray Kurzweil Not checkable 26 May 2022 We don’t want to have a separate thing. Ray Kurzweil · Not checkable · 26 May 2022
Back to the list

Not checkable

“The second thing looks unworkable (less so than CEV, but still lethally unworkable) because corrigibility runs actively counter to instrumentally convergent behaviors within a core of general intelligence”

Eliezer Yudkowsky 10 June 2022 AGI Ruin: A List of Lethalities (MIRI)

The second thing looks unworkable (less so than CEV, but still lethally unworkable) because corrigibility runs actively counter to…

Not checkable An opinion, value judgement or statement with nothing to check.

Cite this claim

Eliezer Yudkowsky, “The second thing looks unworkable (less so than CEV, but still lethally unworkable) because corrigibility runs actively counter to…” (said 10 June 2022). Verdict: not checkable. Reliability API, claim c_b7ccef9e5892, method 0.1.0+c876030c. https://epistacklabs.com/claims/c_b7ccef9e5892

Verdicts are made automatically; how verdicts and scores are made. The same data is on Hugging Face under CC BY 4.0.