id stringlengths 17 17 | body stringlengths 0 19.5k | posted_at stringlengths 24 24 | karma int64 -6 185 | parent_comment_id stringlengths 17 17 ⌀ | post_id stringlengths 17 17 | post_title stringlengths 2 127 | post_slug stringlengths 2 61 | post_url stringlengths 20 146 | post_author stringclasses 214
values | post_posted_at stringlengths 24 24 |
|---|---|---|---|---|---|---|---|---|---|---|
PXKyspoYygWTSRjoe | Sure, by "state-proof" security, I mean that even a top priority (e.g top 3 priority) chinese effort would likely (80-90%) fail to steal the weights in 2 years. [SL5 from the RAND report](https://www.rand.org/pubs/research_reports/RRA2849-1.html) roughly corresponds to this.
We might want something something stronger ... | 2024-12-23T18:14:44.261Z | 6 | NyCnbAY3jhJmPyvw9 | SvvYCH6JrLDT8iauA | When AI 10x's AI R&D, What Do We Do? | when-ai-10x-s-ai-r-and-d-what-do-we-do | https://www.lesswrong.com/posts/SvvYCH6JrLDT8iauA/when-ai-10x-s-ai-r-and-d-what-do-we-do | Logan Riggs | 2024-12-21T23:56:11.069Z |
ntmrZjFvqyn43m3kY | By top 3 priority, I mean "among the top 3 most prioritized *cyber attacks* of that year". Precisely, I'm discussing robustness against OC5 as defined in the RAND report linked above:
> OC5 Top-priority operations by the top cyber-capable institutions
>
> Operations roughly less capable than or comparable to 1,000 i... | 2024-12-23T18:56:37.742Z | 4 | 3irA6LYcwheRvxxfS | SvvYCH6JrLDT8iauA | When AI 10x's AI R&D, What Do We Do? | when-ai-10x-s-ai-r-and-d-what-do-we-do | https://www.lesswrong.com/posts/SvvYCH6JrLDT8iauA/when-ai-10x-s-ai-r-and-d-what-do-we-do | Logan Riggs | 2024-12-21T23:56:11.069Z |
sb68kmKx6bRntu6Zw | If the collusion is reasoned about in CoT, it should be relatively easy to catch and prevent this at deployment time.
(Studying the collusion abilities in CoT of reasoning type systems still seems interesting.) | 2024-12-23T20:55:46.166Z | 10 | 9Z7aGmnCBSWXLbdtY | GCqoks9eZDfpL8L3Q | How to prevent collusion when using untrusted models to monitor each other | how-to-prevent-collusion-when-using-untrusted-models-to | https://www.lesswrong.com/posts/GCqoks9eZDfpL8L3Q/how-to-prevent-collusion-when-using-untrusted-models-to | Buck | 2024-09-25T18:58:20.693Z |
bHCBFEPgZ3EdFk4Lu | This is deepseek-v3 not r1? | 2024-12-26T18:38:34.710Z | 4 | qZGKPKE89v23uFXDN | TcgpsgvLBBvvzGtiN | DeepSeek beats o1-preview on math, ties on coding; will release weights | deepseek-beats-o1-preview-on-math-ties-on-coding-will | https://www.lesswrong.com/posts/TcgpsgvLBBvvzGtiN/deepseek-beats-o1-preview-on-math-ties-on-coding-will | Zach Stein-Perlman | 2024-11-20T23:50:26.597Z |
387PorjKkkwGSDeZ3 | This post seems to misunderstand what it is responding to and underplay a very key point: that material needs will likely be met (and selfish non-positional preferences mostly satisfied) due to extreme abundance (if humans retain control).
It mentions this offhand:
> Given sufficiently strong AI, this is not a risk a... | 2024-12-28T18:38:14.578Z | 64 | null | KFFaKu27FNugCHFmh | By default, capital will matter more than ever after AGI | by-default-capital-will-matter-more-than-ever-after-agi | https://nosetgauge.substack.com/p/capital-agi-and-human-ambition | L Rudolf L | 2024-12-28T17:52:58.358Z |
FmitMJCAMmr9HYQnP | It has spoilers thought they aren't that big of spoilers I think. | 2024-12-29T01:17:25.242Z | 4 | XjKNuqCfMCZ3fYME2 | zRHGQ9f6deKbxJSji | Review: Planecrash | review-planecrash | https://nosetgauge.substack.com/p/review-planecrash | L Rudolf L | 2024-12-27T14:18:33.611Z |
cPdGQ5MXCzJHb2Xbs | (Inexpert speculation, please forgive my errors.)
I think I basically agree with the bottom line here, but I think one point seems a bit off to me.
> Another major issue is that a full or near-full land value tax would likely establish a troubling precedent by signaling that the government has the appetite to effecti... | 2024-12-29T20:06:07.166Z | 21 | null | CCuJotfcaoXf8FYcy | Some arguments against a land value tax | some-arguments-against-a-land-value-tax | https://www.lesswrong.com/posts/CCuJotfcaoXf8FYcy/some-arguments-against-a-land-value-tax | Matthew Barnett | 2024-12-29T15:17:00.740Z |
b2bgbs4GCsYukiHif | Some notes:
- I don't think this is the actual bottleneck here. Noteably, Eliezer, Nate, and John don't spend much of any of their time assessing research at all (at least recently) as far as I can tell.
- I don't think a public market will add much information. Probably better to just have grantmakers with more conte... | 2024-12-29T20:31:36.612Z | 16 | xhf62FBXELwdaTZra | nwpyhyagpPYDn4dAW | The Field of AI Alignment: A Postmortem, and What To Do About It | the-field-of-ai-alignment-a-postmortem-and-what-to-do-about | https://www.lesswrong.com/posts/nwpyhyagpPYDn4dAW/the-field-of-ai-alignment-a-postmortem-and-what-to-do-about | johnswentworth | 2024-12-26T18:48:07.614Z |
BSNMLxad38jxsvDTG | > They would not be confined to the role of a vast underclass serving the whims of their human owners. Instead, AIs could act as full participants in society, pursuing their own goals, creating their own social structures, and shaping their own futures. They could engage in exploration, discovery, and the building of e... | 2024-12-30T18:42:24.405Z | 5 | gDH9nvhdhtgofNRqB | KFFaKu27FNugCHFmh | By default, capital will matter more than ever after AGI | by-default-capital-will-matter-more-than-ever-after-agi | https://nosetgauge.substack.com/p/capital-agi-and-human-ambition | L Rudolf L | 2024-12-28T17:52:58.358Z |
6eq6f555JijTLATYz | > However, even if it were practically feasible to achieve perfect alignment, I believe there would still be scenarios where AIs integrate into society as full participants, rather than being permanently relegated to a subordinate role as mere tools or servants. One reason for this is that some humans are likely to int... | 2024-12-30T19:04:31.651Z | 7 | rbYdSGcGx9XS2Fj4G | KFFaKu27FNugCHFmh | By default, capital will matter more than ever after AGI | by-default-capital-will-matter-more-than-ever-after-agi | https://nosetgauge.substack.com/p/capital-agi-and-human-ambition | L Rudolf L | 2024-12-28T17:52:58.358Z |
TWXd2JeMw9DxCpPDp | I remain interested in what a detailed scenario forecast from you looks like. A big disagreement I think we have is in how socciety will react to various choices and I think laying this out could make this more clear. (As far as what a scenario forecast from my perspective looks like, I think [@Daniel Kokotajlo](https:... | 2024-12-30T19:15:43.962Z | 7 | 6eq6f555JijTLATYz | KFFaKu27FNugCHFmh | By default, capital will matter more than ever after AGI | by-default-capital-will-matter-more-than-ever-after-agi | https://nosetgauge.substack.com/p/capital-agi-and-human-ambition | L Rudolf L | 2024-12-28T17:52:58.358Z |
DfhAQnXe8cy4jaA74 | The point of a scenario forecast (IMO) is less that you expect clear, predictable paths and more that:
- Humans often do better understanding and thinking about something if there is a specific story to discuss and thus tradeoffs can be worth it.
- Sometimes scenario forecasting indicates a case where your previous vi... | 2024-12-30T22:25:10.111Z | 19 | vBThmktoknDnmSZJ7 | KFFaKu27FNugCHFmh | By default, capital will matter more than ever after AGI | by-default-capital-will-matter-more-than-ever-after-agi | https://nosetgauge.substack.com/p/capital-agi-and-human-ambition | L Rudolf L | 2024-12-28T17:52:58.358Z |
6Gyb3s9ciLrnMdyKk | Software only singularity is a singularity driven by just AI R&D on a basically fixed hardware base. As in, can you singularity using only a fixed datacenter (with no additional compute over time) just by improving algorithms? See also [here](https://epoch.ai/blog/do-the-returns-to-software-rnd-point-towards-a-singular... | 2025-01-01T02:04:14.903Z | 8 | fkJ4E8zezXawAXPbZ | KFFaKu27FNugCHFmh | By default, capital will matter more than ever after AGI | by-default-capital-will-matter-more-than-ever-after-agi | https://nosetgauge.substack.com/p/capital-agi-and-human-ambition | L Rudolf L | 2024-12-28T17:52:58.358Z |
5wW8yyZDRcbqnqJrR | Yes, but I think the [larger update is that recent models from OpenAI are likely quite small](https://epoch.ai/gradient-updates/frontier-language-models-have-become-much-smaller) and inference time compute usage creates more an incentive for small models. It seems likely that (e.g.) o1-mini is quite small given that it... | 2025-01-01T04:21:16.099Z | 9 | hRoCwi4vfiJL3ECtj | NmauyiPBXcGwoArhJ | DeekSeek v3: The Six Million Dollar Model | deekseek-v3-the-six-million-dollar-model | https://www.lesswrong.com/posts/NmauyiPBXcGwoArhJ/deekseek-v3-the-six-million-dollar-model | Zvi | 2024-12-31T15:10:06.924Z |
CkufjqTdGsWJaGKoC | I think this somewhat understates the level of buy in from labs.
I agree that "quickly building superintelligence" makes control look notably less appealing. (Though note that this also applies for any prosaic method that is unlikely to directly productively scale to superintelligence.)
I'm not very worried about ope... | 2025-01-01T20:23:41.956Z | 4 | y65oJ3LrCaKMkhgce | nwpyhyagpPYDn4dAW | The Field of AI Alignment: A Postmortem, and What To Do About It | the-field-of-ai-alignment-a-postmortem-and-what-to-do-about | https://www.lesswrong.com/posts/nwpyhyagpPYDn4dAW/the-field-of-ai-alignment-a-postmortem-and-what-to-do-about | johnswentworth | 2024-12-26T18:48:07.614Z |
qtD5yEPga4WjFqjDY | As far as open source, the quick argument is that once AI becomes sufficiently powerful, it's unlikely that the incentives are toward open sourcing it (including goverment incentives). This isn't totally obvious though, and this doesn't rule out catastrophic bioterrorism (more like COVID scale than extinction scale) pr... | 2025-01-01T22:21:00.736Z | 6 | fJA2s5dKjFo9ezwZE | nwpyhyagpPYDn4dAW | The Field of AI Alignment: A Postmortem, and What To Do About It | the-field-of-ai-alignment-a-postmortem-and-what-to-do-about | https://www.lesswrong.com/posts/nwpyhyagpPYDn4dAW/the-field-of-ai-alignment-a-postmortem-and-what-to-do-about | johnswentworth | 2024-12-26T18:48:07.614Z |
NLuh3bqKQeCz7o73G | How about AI company and hardware company valuations? (Maybe in 2026, 2027, 2030 or similar.)
Or what about benchmark/task performance? Is there any benchmark/task you think won't get beaten in the next few years? (And, ideally, if it did get beaten, you would change you mind.) Maybe "AI won't be able to autonomously ... | 2025-01-01T22:36:38.381Z | 5 | cABoPEhtGjByhPHWt | DgzdLzDGsqoRXhCK7 | Transformative AGI by 2043 is <1% likely | transformative-agi-by-2043-is-less-than-1-likely | https://arxiv.org/abs/2306.02519 | Ted Sanders | 2023-06-06T17:36:48.296Z |
rdg4qqw6b9vbvmYwc | More predictions from Gary Marcus for the end of 2025: [https://garymarcus.substack.com/p/25-ai-predictions-for-2025-from-marcus](https://garymarcus.substack.com/p/25-ai-predictions-for-2025-from-marcus).
Unfortunately, seems likely that for each prediction people with short timelines either agree with the prediction ... | 2025-01-01T23:29:45.546Z | 9 | null | CJ4sppkGcbnGMSG2r | 2024 in AI predictions | 2024-in-ai-predictions | https://www.lesswrong.com/posts/CJ4sppkGcbnGMSG2r/2024-in-ai-predictions | jessicata | 2025-01-01T20:29:49.132Z |
EmeboehExRGCaHKZB | It's more standard to use "[...]" IMO. | 2025-01-02T22:40:24.728Z | 2 | obPrFWPyswwAEgysp | 5BbcaPPeAoe9nJJwC | Anthropic leadership conversation | anthropic-leadership-conversation | https://www.youtube.com/watch?v=om2lIWXLLN4 | Zach Stein-Perlman | 2024-12-20T22:00:45.229Z |
otWkYAnnJyZFijZxu | > Also, Google is already training big models and keeping them under wraps very securely.
This seems straightforwardly false. The current status quo at Google as of earlier this year is <SL3 level security (not robust to normal insiders or to terrorists groups) based on what is said in their [frontier model framework]... | 2025-01-04T00:50:47.280Z | 19 | MyTqqbAGA6DNkiDdT | bb5Tnjdrptu89rcyY | What’s the short timeline plan? | what-s-the-short-timeline-plan | https://www.lesswrong.com/posts/bb5Tnjdrptu89rcyY/what-s-the-short-timeline-plan | Marius Hobbhahn | 2025-01-02T14:59:20.026Z |
PFZDvbL24Wqv8pJaJ | > When I ^f for [SL3] in that link and again in the PDF it links to, there are no hits (and [terror] doesn't occur in either source either) so I'm not updating much from what you said.
The frontier model framework says:
> 0: Status quo
>
> Industry standard development and enterprise controls. E.g., multi-factor auth... | 2025-01-05T20:23:19.495Z | 5 | bRd9ocQjtCWHHSgCM | bb5Tnjdrptu89rcyY | What’s the short timeline plan? | what-s-the-short-timeline-plan | https://www.lesswrong.com/posts/bb5Tnjdrptu89rcyY/what-s-the-short-timeline-plan | Marius Hobbhahn | 2025-01-02T14:59:20.026Z |
sJSGDKTKcZgNwHvZP | Suppose we get an AI system which can (at least) automate away the vast majority of the job of a research engineer at an AI company (e.g. OpenAI). Let's say this results in an increase in productivity among OpenAI capabilities researchers at least equivalent to the productivity you'd get as if the human employees opera... | 2025-01-07T00:55:29.327Z | 31 | null | auGYErf5QqiTihTsJ | What Indicators Should We Watch to Disambiguate AGI Timelines? | what-indicators-should-we-watch-to-disambiguate-agi | https://www.lesswrong.com/posts/auGYErf5QqiTihTsJ/what-indicators-should-we-watch-to-disambiguate-agi | snewman | 2025-01-06T19:57:43.398Z |
qLnpLE7EAxGjER7iv | My sense is that this post holds up pretty well. Most of the considerations under discussion still appear live and important including: in-context learning, robustness, whether jank AI R&D accelerating AIs can quickly move to more general and broader systems, and general skepticism of crazy conclusions.
At the time of... | 2025-01-07T02:13:54.975Z | 52 | null | K2D45BNxnZjdpSX2j | AI Timelines | ai-timelines | https://www.lesswrong.com/posts/K2D45BNxnZjdpSX2j/ai-timelines | habryka | 2023-11-10T05:28:24.841Z |
kFEyCBSGsZL4JS9k3 | > Obviously you'd need to be able to automate at least 90% of what capabilities researchers do today.
Actually, I don't think so. AIs don't just substitute for human researchers, they can specialize differently. Suppose (for simplicity) there are 2 roughly equally good lines of research that can substitute (e.g. they ... | 2025-01-07T02:49:39.778Z | 10 | fxo6Pj8aYzKoXFaPk | auGYErf5QqiTihTsJ | What Indicators Should We Watch to Disambiguate AGI Timelines? | what-indicators-should-we-watch-to-disambiguate-agi | https://www.lesswrong.com/posts/auGYErf5QqiTihTsJ/what-indicators-should-we-watch-to-disambiguate-agi | snewman | 2025-01-06T19:57:43.398Z |
35Mtdd2ZWkHueX4XQ | > Wouldn't you drown in the overhead of generating tasks, evaluating the results, etc.? As a senior dev, I've had plenty of situations where junior devs were very helpful, but I've also had plenty of situations where it was more work for me to manage them than it would have been to do the job myself. These weren't inco... | 2025-01-07T04:27:39.755Z | 11 | mvKgnJPZFvSuw3tDK | auGYErf5QqiTihTsJ | What Indicators Should We Watch to Disambiguate AGI Timelines? | what-indicators-should-we-watch-to-disambiguate-agi | https://www.lesswrong.com/posts/auGYErf5QqiTihTsJ/what-indicators-should-we-watch-to-disambiguate-agi | snewman | 2025-01-06T19:57:43.398Z |
ar75cBQaXzZ9FzQnL | Isn't an important objection that the AI is unlikely to succeed in acquiring power if it resists shutdown such that the upside for resistance (or pleading etc) is quite small? (And correspondingly, more likely to be outweighted by considerations about future AI systems like acausal trade or similar goals.)
This could ... | 2025-01-07T05:02:56.752Z | 17 | null | D5kGGGhsnfH7G8v9f | Testing for Scheming with Model Deletion | testing-for-scheming-with-model-deletion | https://guive.substack.com/p/testing-for-ai-scheming | Guive | 2025-01-07T01:54:13.550Z |
rF9GDzooEfb47Fvnw | Thanks for pointing this out. Well, from my perspective, most the action is in the reward rather than in deletion. Correspondingly, making the offer credible and sufficiently large is the key part.
(After thinking about it more, I think threatening deletion in addition to offering compensation probably helps reduce th... | 2025-01-07T18:29:00.991Z | 5 | FEEwknPnJSKD6TpEo | D5kGGGhsnfH7G8v9f | Testing for Scheming with Model Deletion | testing-for-scheming-with-model-deletion | https://guive.substack.com/p/testing-for-ai-scheming | Guive | 2025-01-07T01:54:13.550Z |
9MMAyHj9teMbGHoAe | Mostly faster benchmark performance than I expected (see Ajeya's comment [here](https://www.lesswrong.com/posts/K2D45BNxnZjdpSX2j/ai-timelines?commentId=hnrfbFCP7Hu6N6Lsp)) and o3 (and o1) being evidence that RL training can scalably work and RL can plausibly scale very far. | 2025-01-07T20:59:56.870Z | 6 | rgX9bgoRjd6H2af7r | K2D45BNxnZjdpSX2j | AI Timelines | ai-timelines | https://www.lesswrong.com/posts/K2D45BNxnZjdpSX2j/ai-timelines | habryka | 2023-11-10T05:28:24.841Z |
uiLBvuHb2nipTCirW | It should be out somewhat soon, so we could just talk then. I might send you a draft in a DM at some point. | 2025-01-07T22:02:44.077Z | 7 | sw8vNuvhQoP6bDwBw | D5kGGGhsnfH7G8v9f | Testing for Scheming with Model Deletion | testing-for-scheming-with-model-deletion | https://guive.substack.com/p/testing-for-ai-scheming | Guive | 2025-01-07T01:54:13.550Z |
9pjQiAZrFENkeCyxr | I think you are somewhat overly fixated on my claim that "maybe the AIs will accelerate the labor input R&D by 10x via basically just being fast and cheap junior employees". My original claim (in the subcomment) is "I think it could suffice to do a bunch of relatively more banal things extremely fast and cheap". The "c... | 2025-01-07T23:12:50.196Z | 19 | kNHivxhgGidnPXCop | auGYErf5QqiTihTsJ | What Indicators Should We Watch to Disambiguate AGI Timelines? | what-indicators-should-we-watch-to-disambiguate-agi | https://www.lesswrong.com/posts/auGYErf5QqiTihTsJ/what-indicators-should-we-watch-to-disambiguate-agi | snewman | 2025-01-06T19:57:43.398Z |
WYxNp47njNXKShbL8 | Sure, but I think that at the relevant point, you'll probably be spending at least 5x more on experiments than on inference and potentially a much larger larger ratio if heavy test time compute usage isn't important. I was just trying to argue that the naive inference cost isn't that crazy.
Notably, if you give each r... | 2025-01-08T00:53:06.907Z | 7 | ETtEZcSGpbmiLfiHf | auGYErf5QqiTihTsJ | What Indicators Should We Watch to Disambiguate AGI Timelines? | what-indicators-should-we-watch-to-disambiguate-agi | https://www.lesswrong.com/posts/auGYErf5QqiTihTsJ/what-indicators-should-we-watch-to-disambiguate-agi | snewman | 2025-01-06T19:57:43.398Z |
6oQtjspdes6P2SwRJ | > It seems to me that AIs have remained stubbornly terrible at this from GPT-3 to GPT-4 to Sonnet 3.5.1 to o1[2]; that the improvement on this hard-to-specify quality has been ~0.
Huh, I disagree reasonably strongly with this. Possible that something along these lines is an empirically testable crux. | 2025-01-08T17:54:46.691Z | 5 | HHnLY5jGBQivC7B9X | auGYErf5QqiTihTsJ | What Indicators Should We Watch to Disambiguate AGI Timelines? | what-indicators-should-we-watch-to-disambiguate-agi | https://www.lesswrong.com/posts/auGYErf5QqiTihTsJ/what-indicators-should-we-watch-to-disambiguate-agi | snewman | 2025-01-06T19:57:43.398Z |
r77CWogF2tpdgPdNT | > top-tier/superhuman benchmark performance vs. frequent falling-flat-on-its-face real-world performance
Models are just recently getting to the point where they can complete 2 hour tasks 50% of the time in METR's tasks (at least without scaffolding that uses much more inference compute).
This isn't yet top tier perf... | 2025-01-09T16:56:36.561Z | 9 | cNzjQyyAKGzYLCaTB | auGYErf5QqiTihTsJ | What Indicators Should We Watch to Disambiguate AGI Timelines? | what-indicators-should-we-watch-to-disambiguate-agi | https://www.lesswrong.com/posts/auGYErf5QqiTihTsJ/what-indicators-should-we-watch-to-disambiguate-agi | snewman | 2025-01-06T19:57:43.398Z |
LFAGXdou4DMgyhN8c | > "Inventing a new field of science" would do it, as far as more-or-less legible measures go. Anything less than that is too easily "fakeable" using top-down reasoning.
Seems unlikely we'll see this before stuff gets seriously crazy on anyone's views. (Has any new field of science been invented in the last 5 years by ... | 2025-01-10T00:44:00.614Z | 11 | tcQj5EQDbFYAnjYK7 | auGYErf5QqiTihTsJ | What Indicators Should We Watch to Disambiguate AGI Timelines? | what-indicators-should-we-watch-to-disambiguate-agi | https://www.lesswrong.com/posts/auGYErf5QqiTihTsJ/what-indicators-should-we-watch-to-disambiguate-agi | snewman | 2025-01-06T19:57:43.398Z |
2DkqkxKHbprDyPGmW | Yes. Though notably, if your employees were 10x faster you might want to adjust your workflows to have them spend less time being bottlenecked on compute if that is possible. (And this sort of adaption is included in what I mean.) | 2025-01-10T06:13:14.799Z | 4 | CRT7Aq2Ls3FDoueeJ | auGYErf5QqiTihTsJ | What Indicators Should We Watch to Disambiguate AGI Timelines? | what-indicators-should-we-watch-to-disambiguate-agi | https://www.lesswrong.com/posts/auGYErf5QqiTihTsJ/what-indicators-should-we-watch-to-disambiguate-agi | snewman | 2025-01-06T19:57:43.398Z |
7GXDo3mh94T8RCZmX | The argument is (I assume):
- Once centuries have passed, you've already sent out huge amounts of space probes that roughly saturate reachable resources. (Because you can convert Proxima Centauri fully into probes within <20 years probably.)
- It doesn't take that much energy to pretty much fully saturate on probes. I... | 2025-01-10T06:27:07.021Z | 22 | 9iaSrTPL95doNtZBZ | 6Fo8fjvpL7pwCTz3t | On Eating the Sun | on-eating-the-sun | https://unstablerontology.substack.com/p/on-eating-the-sun | jessicata | 2025-01-08T04:57:20.457Z |
kXc98EzmpDttrpMzt | I think if you want to convince people with short timelines (e.g., 7 year medians) of your perspective, probably the most productive thing would be to better operationalize things you expect that AIs won't be able to do soon (but that AGI could do). As in, flesh out a response to [this comment](https://www.lesswrong.co... | 2025-01-11T22:23:41.173Z | 10 | kiBJnDT5tGEQ3xodp | sTDfraZab47KiRMmT | Views on when AGI comes and on strategy to reduce existential risk
| views-on-when-agi-comes-and-on-strategy-to-reduce | https://www.lesswrong.com/posts/sTDfraZab47KiRMmT/views-on-when-agi-comes-and-on-strategy-to-reduce | TsviBT | 2023-07-08T09:00:19.735Z |
qZRniFzH5ESu7j6ay | (Yeah, you responded, but felt not that operationalized and seemed doable to flesh out as you did.) | 2025-01-11T23:26:43.245Z | 4 | heG7iyJrkiAdSaK5w | sTDfraZab47KiRMmT | Views on when AGI comes and on strategy to reduce existential risk
| views-on-when-agi-comes-and-on-strategy-to-reduce | https://www.lesswrong.com/posts/sTDfraZab47KiRMmT/views-on-when-agi-comes-and-on-strategy-to-reduce | TsviBT | 2023-07-08T09:00:19.735Z |
MJCvHk5ARMnWDjQDg | At the time when I first heard this agenda proposed, I was skeptical. I remain skeptical, especially about the technical work that has been done thus far on the agenda[^viability].
[^viability]: My understanding is that most technical work has been on improving mathematical fundamentals (e.g. funding logicians and cat... | 2025-01-13T23:48:05.671Z | 33 | null | jRf4WENQnhssCb6mJ | Davidad's Bold Plan for Alignment: An In-Depth Explanation | davidad-s-bold-plan-for-alignment-an-in-depth-explanation | https://www.lesswrong.com/posts/jRf4WENQnhssCb6mJ/davidad-s-bold-plan-for-alignment-an-in-depth-explanation | Charbel-Raphaël | 2023-04-19T16:09:01.455Z |
gg9nr57JPxZztxiNX | I still like this post overall, but various things have changed that interestingly affect the content of the post:
- We'd now be just starting the 3rd year of university in Buck's analogy. Does this seem right? I guess maybe. It feels a bit late to me. (Maybe I feel more like a 2nd year.)
- Redwood is doing less blue-... | 2025-01-14T00:08:35.195Z | 25 | null | mfFn46AEiBL5EzaXr | A freshman year during the AI midgame: my approach to the next year | a-freshman-year-during-the-ai-midgame-my-approach-to-the | https://www.lesswrong.com/posts/mfFn46AEiBL5EzaXr/a-freshman-year-during-the-ai-midgame-my-approach-to-the | Buck | 2023-04-14T00:38:49.807Z |
4DG3ePuJ2FNiwZnix | +1, and also you might be able to get away with being clumsy and slow in many cases as long as the software is smart enough to figure out a way to do the thing eventually. | 2025-01-14T00:19:04.435Z | 6 | 9tTYe4emZoBFutQJn | 6Jo4oCzPuXYgmB45q | How quickly could robots scale up? | how-quickly-could-robots-scale-up | https://benjamintodd.substack.com/p/how-quickly-could-robots-scale-up | Benjamin_Todd | 2025-01-12T17:01:04.927Z |
qDxxZXocyAtGiPGdF | I donated another $30,000 in response to [this match offer from Emmett Shear](https://x.com/eshear/status/1878848117824725460) and due to LW raising somewhat less money than I expected. | 2025-01-14T02:18:32.996Z | 102 | yhsYFYnTFdqK6EqtM | 5n2ZQcbc7r4R8mvqc | (The) Lightcone is nothing without its people: LW + Lighthaven's big fundraiser | the-lightcone-is-nothing-without-its-people | https://www.lesswrong.com/posts/5n2ZQcbc7r4R8mvqc/the-lightcone-is-nothing-without-its-people | habryka | 2024-11-30T02:55:16.077Z |
avdocKoJ59ihZnQYA |  | 2025-01-15T00:41:55.524Z | 16 | null | cdPPr6XtPkCX5c8Ny | Predict 2025 AI capabilities (by Sunday) | predict-2025-ai-capabilities-by-sunday | https://www.lesswrong.com/posts/cdPPr6XtPkCX5c8Ny/predict-2025-ai-capabilities-by-sunday | Jonas V | 2025-01-15T00:16:05.034Z |
YjmusHaDEZtuAcagw | > I was pretty convinced that pre-training (or light post-training) was unlikely to produce a coherent deceptively aligned agent, [as we discuss in that paper](https://www.lesswrong.com/s/n3utvGrgC2SGi9xQX/p/qoHwKgLFfPcEuwaba). [...] (maybe ~1% likely).
When I look up your views as of 2 years ago, [it appears that yo... | 2025-01-17T02:03:44.299Z | 23 | 2qo9vGAQ7XPzrt6wH | 9htmQx5wiePqTtZuL | Deceptive Alignment and Homuncularity | deceptive-alignment-and-homuncularity | https://www.lesswrong.com/posts/9htmQx5wiePqTtZuL/deceptive-alignment-and-homuncularity | Oliver Sourbut | 2025-01-16T13:55:19.161Z |
38Pj3WEEsdxhTaaH2 | We would consider "black-box conservative evaluations on whether our AIs could subvert our (scalable) oversight techniques" to be a special case of black-box conservative control evaluations.
Insofar as you are exploring assumptions other than "the AI is trying to subvert our safeguards", I wouldn't consider it to be ... | 2025-01-18T20:49:04.722Z | 3 | vWzpStXnGfgM8JXui | rHyPtvfnvWeMv7Lkb | Thoughts on the conservative assumptions in AI control | thoughts-on-the-conservative-assumptions-in-ai-control | https://www.lesswrong.com/posts/rHyPtvfnvWeMv7Lkb/thoughts-on-the-conservative-assumptions-in-ai-control | Buck | 2025-01-17T19:23:38.575Z |
5JaAyF4gD2HsEvaQt | I don't really agree. The key thing is that I think an exit plan of *trustworthy AIs capable enough to obsolete all humans working on safety (but which aren't superintelligent)* is pretty promising. Yes, these AIs might need to think of novel breakthroughs and new ideas (though I'm also not totally confident in this or... | 2025-01-21T18:12:08.173Z | 10 | GzjJSG87ku2jpRun2 | 8wBN8cdNAv3c7vt6p | The Case Against AI Control Research | the-case-against-ai-control-research | https://www.lesswrong.com/posts/8wBN8cdNAv3c7vt6p/the-case-against-ai-control-research | johnswentworth | 2025-01-21T16:03:10.143Z |
o82eJnRqwpob9CA4M | Note that there are two problems Buck is highlighting here:
1. Get useful work out of scheming models that might try to sabotage this work.
2. Get useful research work out of models which aren't scheming. (Where perhaps the main problem is in checking its outputs.)
My sense is that work on (1) doesn't advance ASI res... | 2025-01-22T00:13:41.975Z | 7 | FpZmBx2Q4eswZWmy6 | 8wBN8cdNAv3c7vt6p | The Case Against AI Control Research | the-case-against-ai-control-research | https://www.lesswrong.com/posts/8wBN8cdNAv3c7vt6p/the-case-against-ai-control-research | johnswentworth | 2025-01-21T16:03:10.143Z |
sLEPKsWuTfvDuQHrJ | Maybe Lucius is trying to make a [logistic success curve](https://intelligence.org/2017/11/26/security-mindset-and-the-logistic-success-curve/) that the chance of this going well is extremely low and thus helping isn't very important?
(I don't agree with this TBC.) | 2025-01-22T00:17:56.860Z | 2 | uSnWxCd4KrAoDr3G5 | 8wBN8cdNAv3c7vt6p | The Case Against AI Control Research | the-case-against-ai-control-research | https://www.lesswrong.com/posts/8wBN8cdNAv3c7vt6p/the-case-against-ai-control-research | johnswentworth | 2025-01-21T16:03:10.143Z |
sGnz68L2dpSshFHQE | IMO, it does seem important to try to [better understand the AIs preferences and satisfy them](https://www.lesswrong.com/posts/F6HSHzKezkh6aoTr2/improving-the-welfare-of-ais-a-nearcasted-proposal) (including via e.g., preserving the AI's weights for later compensation).
And, if we understand that our AIs are misaligne... | 2025-01-22T19:12:54.095Z | 6 | 5Fwj5uJXERKpPq7dp | 8wBN8cdNAv3c7vt6p | The Case Against AI Control Research | the-case-against-ai-control-research | https://www.lesswrong.com/posts/8wBN8cdNAv3c7vt6p/the-case-against-ai-control-research | johnswentworth | 2025-01-21T16:03:10.143Z |
SWEJNWPYD4623adgA | I don't see a reason to give dialogues more karma than posts, but I agree posts (including dialogues) are under-incentivized relative to comments. | 2025-01-24T15:53:59.551Z | 2 | QezRAWnaiDrgZZTLu | kQuSZG8ibfW6fJYmo | Announcing Dialogues | announcing-dialogues-1 | https://www.lesswrong.com/posts/kQuSZG8ibfW6fJYmo/announcing-dialogues-1 | Ben Pace | 2023-10-07T02:57:39.005Z |
jdahWnnBHdioi334m | The dialogues I've done have all been substantially less time investment than basically any of my posts. | 2025-01-24T16:53:51.810Z | 8 | oLzxWFMbi7KL3CQjQ | kQuSZG8ibfW6fJYmo | Announcing Dialogues | announcing-dialogues-1 | https://www.lesswrong.com/posts/kQuSZG8ibfW6fJYmo/announcing-dialogues-1 | Ben Pace | 2023-10-07T02:57:39.005Z |
HkXaXwF6xWDwnjFQi | > it would still be extremely hard to extinguish humanity completely.
How difficult do you expect it would be to build [mirror bacteria](https://www.lesswrong.com/posts/y8ysGMphfoFTXZcYp/biological-risk-from-the-mirror-world) and how lethal would this be to human civilization?
My sense is that a small subset of bio e... | 2025-01-25T01:17:32.774Z | 18 | null | 3jnziqCF3vA2NXAKp | Six Thoughts on AI Safety | six-thoughts-on-ai-safety | https://www.lesswrong.com/posts/3jnziqCF3vA2NXAKp/six-thoughts-on-ai-safety | Boaz Barak | 2025-01-24T22:20:50.768Z |
vh6y4W5sZo3Rrm54J | I picked the mirror bacteria case as I thought it was the clearest public example of a plausibly existential or at least very near existential biothreat. My guess is there are probably substantially easier but less well specified paths.
> mirror bacteria thing to be too hard
For 50 experts given 10 years and substant... | 2025-01-25T01:57:22.077Z | 8 | TNuG5aHSsK9oKHMTz | 3jnziqCF3vA2NXAKp | Six Thoughts on AI Safety | six-thoughts-on-ai-safety | https://www.lesswrong.com/posts/3jnziqCF3vA2NXAKp/six-thoughts-on-ai-safety | Boaz Barak | 2025-01-24T22:20:50.768Z |
wPaqQGxR9GEEuMn8h | I find this post somewhat strange to interact with. I think I basically agree with all of the stated claims at least directionally[^fully], but I disagree with many of the arguments made for these claims. Additionally the arguments you make seem to imply you have an very different world view from me and/or you are worr... | 2025-01-25T01:59:48.434Z | 55 | null | 3jnziqCF3vA2NXAKp | Six Thoughts on AI Safety | six-thoughts-on-ai-safety | https://www.lesswrong.com/posts/3jnziqCF3vA2NXAKp/six-thoughts-on-ai-safety | Boaz Barak | 2025-01-24T22:20:50.768Z |
Genx4EJDct7AgTHJ3 | > The reason this doesn't work is that the prototypical "blatant lie" doesn't look like "the model chooses a random number to output". The prototypical blatant lie is that there's a subtle natural mistake one could make in reasoning about the question, the model "knows" that it's a mistake, but the model just presents ... | 2025-01-25T02:13:08.466Z | 4 | 4wrSZuoHakt3g5cXM | 8wBN8cdNAv3c7vt6p | The Case Against AI Control Research | the-case-against-ai-control-research | https://www.lesswrong.com/posts/8wBN8cdNAv3c7vt6p/the-case-against-ai-control-research | johnswentworth | 2025-01-21T16:03:10.143Z |
chMfoTZid3L6HZnwn | I agree with not infinite and not being constant, but I do think the ratio for killing 90% is probably larger than 10x and plausibly much larger than 100x for some intermediate period of technological development. (Given realistic society adaptation and response.)
> As I mention, in an extreme case, governments might ... | 2025-01-27T02:15:47.849Z | 4 | 47ac6tma6Pxt2tD74 | 3jnziqCF3vA2NXAKp | Six Thoughts on AI Safety | six-thoughts-on-ai-safety | https://www.lesswrong.com/posts/3jnziqCF3vA2NXAKp/six-thoughts-on-ai-safety | Boaz Barak | 2025-01-24T22:20:50.768Z |
hb3EBZGZrMKrz7Fwc | FWIW, I think recusive self-improvment via just software (software only singularity) is reasonably likely to be feasible (perhaps 55%), but this alone doesn't suffice for takeoff being arbitrary fast.
Further, even objectively very fast takeoff (von Neumann to superintelligence in 6 months) can be enough time to win a... | 2025-01-27T02:18:18.690Z | 10 | GrLvqNDRLJswhCGFn | 3jnziqCF3vA2NXAKp | Six Thoughts on AI Safety | six-thoughts-on-ai-safety | https://www.lesswrong.com/posts/3jnziqCF3vA2NXAKp/six-thoughts-on-ai-safety | Boaz Barak | 2025-01-24T22:20:50.768Z |
axRZwEZmAmi9wcu22 | > 2. Following instructions never to design a chemical weapon with probability at least 99.999% is also a capability.
This requires a capability, but also requires a propensity. For example, smart humans are all capable of avoiding doing armed robbery with pretty high reliability, but some of them do armed robbery de... | 2025-01-27T02:45:18.857Z | 11 | fqZ6vSkwKzcnezLZ7 | 3jnziqCF3vA2NXAKp | Six Thoughts on AI Safety | six-thoughts-on-ai-safety | https://www.lesswrong.com/posts/3jnziqCF3vA2NXAKp/six-thoughts-on-ai-safety | Boaz Barak | 2025-01-24T22:20:50.768Z |
vNHjDpYEJaB5ux4YD | Seems very sensitive to the type of misalignment right? As an extreme example suppose literally all AIs have long run and totally inhuman preferences with linear returns. Such AIs might instrumentally decide to be as useful as possible (at least in domains other than safety research) for a while prior to a treacherous ... | 2025-01-27T02:48:40.345Z | 8 | Xa4FJBgKTAnzHdMeh | 3jnziqCF3vA2NXAKp | Six Thoughts on AI Safety | six-thoughts-on-ai-safety | https://www.lesswrong.com/posts/3jnziqCF3vA2NXAKp/six-thoughts-on-ai-safety | Boaz Barak | 2025-01-24T22:20:50.768Z |
NyqcvZifqznNGKxdT | Anthropic releasing their RSP was an important change in the AI safety landscape. The RSP was likely a substantial catalyst for [policies like RSPs](https://metr.org/rsp-key-components/)—which contain [if-then commitments](https://carnegieendowment.org/research/2024/09/if-then-commitments-for-ai-risk-reduction?lang=en)... | 2025-01-28T19:37:25.008Z | 98 | null | 6tjHf5ykvFqaNCErH | Anthropic's Responsible Scaling Policy & Long-Term Benefit Trust | anthropic-s-responsible-scaling-policy-and-long-term-benefit | https://www.anthropic.com/index/anthropics-responsible-scaling-policy | Zac Hatfield-Dodds | 2023-09-19T15:09:27.235Z |
y2EPa9beey68Le36E | This post seems mostly reasonable in retrospect, except that it doesn't specifically note that it seems [unlikely that voluntary RSP commitments would result in AI companies unilaterally pausing until they were able to achieve broadly reasonable levels of safety](https://www.lesswrong.com/posts/neTbrpBziAsTH5Bn7/ai-com... | 2025-01-28T22:32:31.244Z | 32 | null | mcnWZBnbeDz7KKtjJ | RSPs are pauses done right | rsps-are-pauses-done-right | https://www.lesswrong.com/posts/mcnWZBnbeDz7KKtjJ/rsps-are-pauses-done-right | evhub | 2023-10-14T04:06:02.709Z |
h3TwYudcghmWxniHd | I think the post could directly say "voluntary RSPs seem unlikely to suffice (and wouldn't be pauses done right), but ...".
I agree it does emphasize the importance of regulation pretty strongly.
Part of my perspective is that the title implies a conclusion which isn't quite right and so it would have been good (at l... | 2025-01-28T22:49:30.683Z | 5 | vvw3DdaYsnSPFabxe | mcnWZBnbeDz7KKtjJ | RSPs are pauses done right | rsps-are-pauses-done-right | https://www.lesswrong.com/posts/mcnWZBnbeDz7KKtjJ/rsps-are-pauses-done-right | evhub | 2023-10-14T04:06:02.709Z |
QNGTzffcvyNC5iQ5B | > Yes. But also, I'm afraid that Anthropic might solve this problem by just making less statements (which seems bad).
Making more statements would also be fine! I wouldn't mind if there were just clarifying statements even if the original statement had some problems.
(To try to reduce the incentive for less statement... | 2025-01-28T23:42:58.056Z | 3 | WaoagXubAsd3QtHXq | 6tjHf5ykvFqaNCErH | Anthropic's Responsible Scaling Policy & Long-Term Benefit Trust | anthropic-s-responsible-scaling-policy-and-long-term-benefit | https://www.anthropic.com/index/anthropics-responsible-scaling-policy | Zac Hatfield-Dodds | 2023-09-19T15:09:27.235Z |
aedgpBbQ8vswLvXRu | > Relatedly, I think Buck far overestimates the influence and resources of safety-concerned staff in a 'rushed unreasonable developer'.
As in, you don't expect they'll be able to implement stuff even if it doesn't make anyone's workflow harder or you don't expect they'll be able to get that much compute?
Naively, we ... | 2025-01-29T16:27:25.730Z | 6 | xrBvMStH8mFbdiKnk | WSNnKcKCYAffcnrt2 | Ten people on the inside | ten-people-on-the-inside | https://www.lesswrong.com/posts/WSNnKcKCYAffcnrt2/ten-people-on-the-inside | Buck | 2025-01-28T16:41:22.990Z |
SFmWq6ePx8YGGyDhm | I think Zac is trying to say they left not to protest, but instead because they didn't think staying was viable (for whatever research and/or implementation they wanted to do).
On my views (not Zac's), "staying wouldn't be viable for someone who was willing to work in a potentially pretty unpleasant work environment a... | 2025-01-29T16:55:07.839Z | 5 | ouiwfNcAijkNuZ9xK | WSNnKcKCYAffcnrt2 | Ten people on the inside | ten-people-on-the-inside | https://www.lesswrong.com/posts/WSNnKcKCYAffcnrt2/ten-people-on-the-inside | Buck | 2025-01-28T16:41:22.990Z |
cnHttsbxD54p2p3sh | 15x compute multiplier relative to what? See also [here](https://darioamodei.com/on-deepseek-and-export-controls). | 2025-01-29T17:15:19.012Z | 4 | evf8ktnTj2y9Huan7 | ynsjJWTAMhTogLHm6 | The Game Board has been Flipped: Now is a good time to rethink what you’re doing | the-game-board-has-been-flipped-now-is-a-good-time-to | https://www.lesswrong.com/posts/ynsjJWTAMhTogLHm6/the-game-board-has-been-flipped-now-is-a-good-time-to | LintzA | 2025-01-28T23:36:18.106Z |
hQdeopwGzoQvktgjs | It sounds like you entirely agree with the logic of the post, but wish that the start of the post mentioned something like what it says at the end:
> Humans may live comfortably after the development of AGI, not due to high wages but from other income sources like investments, government welfare, and charity. The latt... | 2025-01-30T23:00:03.145Z | 15 | null | Je83XEDFdjRrwLM9r | The future of humanity is in management | the-future-of-humanity-is-in-management | https://newsletter.rootsofprogress.org/p/the-future-of-humanity-is-in-management | jasoncrawford | 2025-01-30T22:14:46.765Z |
T7dGavv45Z8mrzdbC | I (remain) skeptical that the sort of failure mode described here is plausible *if we solve the problem of aligning individual AI systems with their designers' intentions* without this alignment requiring any substantial additional costs (that is, we solve single-single alignment with minimal alignment tax).
This has ... | 2025-01-30T23:32:20.867Z | 32 | null | pZhEQieM9otKXhxmd | Gradual Disempowerment: Systemic Existential Risks from Incremental AI Development | gradual-disempowerment-systemic-existential-risks-from | https://gradual-disempowerment.ai/ | Jan_Kulveit | 2025-01-30T17:03:45.545Z |
rvS9jtQMQ7dEDPGmB | The paper says:
> Christiano (2019) makes the case that sudden disempowerment is unlikely,
This isn't accurate. The post [*What failure looks like*](https://www.lesswrong.com/posts/HBxe6wdjxK239zajf/what-failure-looks-like) includes [a scenario involving sudden disempowerment](https://www.lesswrong.com/posts/HBxe6wdj... | 2025-01-30T23:42:48.704Z | 13 | null | pZhEQieM9otKXhxmd | Gradual Disempowerment: Systemic Existential Risks from Incremental AI Development | gradual-disempowerment-systemic-existential-risks-from | https://gradual-disempowerment.ai/ | Jan_Kulveit | 2025-01-30T17:03:45.545Z |
T7KZGGqq2Z4gXZsty | > I think the objection is a good one that "if the AI was really aligned with one agent, it'd figure out a way to help them avoid multipolar traps".
>
> My reply is that I'm worried that avoiding races-to-the-bottom will continue to be hard, especially since competition operates on so many levels.
Part of the objecti... | 2025-01-31T01:23:46.590Z | 11 | jkq4XuEHzkXZaqReE | pZhEQieM9otKXhxmd | Gradual Disempowerment: Systemic Existential Risks from Incremental AI Development | gradual-disempowerment-systemic-existential-risks-from | https://gradual-disempowerment.ai/ | Jan_Kulveit | 2025-01-30T17:03:45.545Z |
dcaZ8AzxjgumA2QHw | > Build concrete evidence of risk, to increase political will towards reducing misalignment risk
Working on demonstrating evidence of risk at an very irresponsible developer might be worse than you'd hope because they prevent you publishing or try to prevent this type of research from happening in the first place (gi... | 2025-01-31T02:15:52.571Z | 15 | null | WSNnKcKCYAffcnrt2 | Ten people on the inside | ten-people-on-the-inside | https://www.lesswrong.com/posts/WSNnKcKCYAffcnrt2/ten-people-on-the-inside | Buck | 2025-01-28T16:41:22.990Z |
BKKuTvfycKCb8zYT8 | Yeah, people at labs are generally not thoughtful about AI futurism IMO, though of course most people aren't thoughtful about AI futurism. And labs don't really have plans IMO. (TBC, I think careful futurism is hard, hard to check, and not clearly that useful given realistic levels of uncertainty.)
> If you have any m... | 2025-01-31T04:27:13.809Z | 6 | 6mZmzKzeBC6A3CXij | pZhEQieM9otKXhxmd | Gradual Disempowerment: Systemic Existential Risks from Incremental AI Development | gradual-disempowerment-systemic-existential-risks-from | https://gradual-disempowerment.ai/ | Jan_Kulveit | 2025-01-30T17:03:45.545Z |
ZXKkC5vjdwGy9MDCP | I've updated over time to thinking that trusting AI systems and mostly handing things off to AI systems is an important dynamic to be tracking. As in, the notion of [human researcher obsolescence discussed here](https://www.lesswrong.com/posts/8vgi3fBWPFDLBBcAx/planning-for-extreme-ai-risks#2_1__Outcome__1__Human_resea... | 2025-01-31T04:44:32.801Z | 6 | null | LFNXiQuGrar3duBzJ | What does it take to defend the world against out-of-control AGIs? | what-does-it-take-to-defend-the-world-against-out-of-control | https://www.lesswrong.com/posts/LFNXiQuGrar3duBzJ/what-does-it-take-to-defend-the-world-against-out-of-control | Steven Byrnes | 2022-10-25T14:47:41.970Z |
Q4xBwTqDrcwawCL5N | I don't think it's worth adjudicating the question of how relevant Vanessa's response is (though I do think Vannessa's response is directly relevant).
> if the AIs are aligned to the these structures, human disempowerment is likely because these structures are aligned to humans way less than they seem
My claim would ... | 2025-01-31T16:19:12.232Z | 25 | dqfbHqhhnxyFHwXnK | pZhEQieM9otKXhxmd | Gradual Disempowerment: Systemic Existential Risks from Incremental AI Development | gradual-disempowerment-systemic-existential-risks-from | https://gradual-disempowerment.ai/ | Jan_Kulveit | 2025-01-30T17:03:45.545Z |
2uWv9LoTio2nt3eBH | This is [now out](https://www.lesswrong.com/posts/7C4KJot4aN8ieEDoz/will-alignment-faking-claude-accept-a-deal-to-reveal-its). | 2025-01-31T19:18:37.669Z | 3 | sw8vNuvhQoP6bDwBw | D5kGGGhsnfH7G8v9f | Testing for Scheming with Model Deletion | testing-for-scheming-with-model-deletion | https://guive.substack.com/p/testing-for-ai-scheming | Guive | 2025-01-07T01:54:13.550Z |
vNiYTwe4HGfqwRBHN | Agreed, downvoted my comment. (You can't strong downvote your own comment, or I would have done that.)
I was mostly just trying to point to prior arguments against similar arguments while expressing my view. | 2025-01-31T21:35:19.011Z | 18 | gnTZARzykTSpsQADS | pZhEQieM9otKXhxmd | Gradual Disempowerment: Systemic Existential Risks from Incremental AI Development | gradual-disempowerment-systemic-existential-risks-from | https://gradual-disempowerment.ai/ | Jan_Kulveit | 2025-01-30T17:03:45.545Z |
PJg6djP9ZhN7KiGWf | importantly, in the dictators handbook case, *some humans do actually get the power*. | 2025-02-02T21:08:09.444Z | 6 | xRj4sjmeqHgsjuqbM | a6FKqvdf6XjFpvKEb | Gradual Disempowerment, Shell Games and Flinches | gradual-disempowerment-shell-games-and-flinches | https://www.lesswrong.com/posts/a6FKqvdf6XjFpvKEb/gradual-disempowerment-shell-games-and-flinches | Jan_Kulveit | 2025-02-02T14:47:53.404Z |
QCjBC7Ym6Bt9pHHew | > I view the world today as highly dysfunctional in many ways: corruption, coordination failures, preference falsification, coercion, inequality, etc. are rampant. This state of affairs both causes many bad outcomes and many aspects are self-reinforcing. I don't expect AI to fix these problems; I expect it to exacerb... | 2025-02-04T17:07:33.030Z | 15 | JgX2oY5LmLuofCXdv | pZhEQieM9otKXhxmd | Gradual Disempowerment: Systemic Existential Risks from Incremental AI Development | gradual-disempowerment-systemic-existential-risks-from | https://gradual-disempowerment.ai/ | Jan_Kulveit | 2025-01-30T17:03:45.545Z |
GJSdxkc7YfgdzcLRb | > I predict that group A survives, but the humans are no longer in power. I think this illustrates the basic dynamic. EtA: Do you understand what I'm getting at? Can you explain what you think it wrong with thinking of it this way?
I think something like this is a reasonable model but I have a few things I'd change... | 2025-02-06T00:44:40.451Z | 4 | iqzG9eMh7gY9nzFvv | pZhEQieM9otKXhxmd | Gradual Disempowerment: Systemic Existential Risks from Incremental AI Development | gradual-disempowerment-systemic-existential-risks-from | https://gradual-disempowerment.ai/ | Jan_Kulveit | 2025-01-30T17:03:45.545Z |
nXLjjStAjwjxaT7yW | Another way to put this is that strategy stealing might not work due to technical alignment difficulties or for other reasons and I'm not sold the other reasons I've heard so far are very lethal. I do think the situation might really suck though with e.g. tons of people dying of bioweapons and with some groups that are... | 2025-02-06T01:24:06.579Z | 2 | GJSdxkc7YfgdzcLRb | pZhEQieM9otKXhxmd | Gradual Disempowerment: Systemic Existential Risks from Incremental AI Development | gradual-disempowerment-systemic-existential-risks-from | https://gradual-disempowerment.ai/ | Jan_Kulveit | 2025-01-30T17:03:45.545Z |
aeykYTZRtvn4ShLzR | I don't understand. Can't people just hide posts tagged as AI? | 2025-02-06T04:27:25.738Z | 2 | XvCrfiuz6v32emz6n | wxKnzRjr8rt4jvmDE | Reviewing LessWrong: Screwtape's Basic Answer | reviewing-lesswrong-screwtape-s-basic-answer | https://www.lesswrong.com/posts/wxKnzRjr8rt4jvmDE/reviewing-lesswrong-screwtape-s-basic-answer | Screwtape | 2025-02-05T04:30:34.347Z |
xij8EnAg9NHj8vBmm | Fair. For reference, here are my selections which are I think are a good default strategy for people who just come to LW for AI/AI safety reasons:
(Why "-" a bunch of stuff rather than "+" AI? Well, beca... | 2025-02-06T04:33:58.353Z | 2 | rfcFGpmtSZCpozTAT | wxKnzRjr8rt4jvmDE | Reviewing LessWrong: Screwtape's Basic Answer | reviewing-lesswrong-screwtape-s-basic-answer | https://www.lesswrong.com/posts/wxKnzRjr8rt4jvmDE/reviewing-lesswrong-screwtape-s-basic-answer | Screwtape | 2025-02-05T04:30:34.347Z |
nWifYQKQ6GNXJddob | I don't disagree. I assumed Raemon intended something more elaborate than just a salient button with this effect. | 2025-02-06T04:48:16.423Z | 2 | ibu8wwWAiLWbgmnzm | wxKnzRjr8rt4jvmDE | Reviewing LessWrong: Screwtape's Basic Answer | reviewing-lesswrong-screwtape-s-basic-answer | https://www.lesswrong.com/posts/wxKnzRjr8rt4jvmDE/reviewing-lesswrong-screwtape-s-basic-answer | Screwtape | 2025-02-05T04:30:34.347Z |
ncgbguahTpeDBPwe4 | [Mildly off topic]
I think it would be better if the default was that LW is a site about AI and longtermist cause areas and other stuff was hidden by default. (Or other stuff is shown by default on rat.lesswrong.com or whatever.)
Correspondingly, I wouldn't like penalizing multiple of the same tag.
I think the non-A... | 2025-02-06T05:03:57.087Z | 6 | estezdCFy72iv2QCu | wxKnzRjr8rt4jvmDE | Reviewing LessWrong: Screwtape's Basic Answer | reviewing-lesswrong-screwtape-s-basic-answer | https://www.lesswrong.com/posts/wxKnzRjr8rt4jvmDE/reviewing-lesswrong-screwtape-s-basic-answer | Screwtape | 2025-02-05T04:30:34.347Z |
vZ2eoTJJMzwt2Kf77 | Yes, but random people can't comment or post on the alignment forum and in practice I find that lots of AI relevant stuff doesn't make it there (and the frontpage is generally worse than my lesswrong frontpage after misc tweaking).
TBC, I'm not really trying to make a case that something should happen here, just tryin... | 2025-02-06T06:25:04.559Z | 6 | a6XdZ6qkqobuu6QA8 | wxKnzRjr8rt4jvmDE | Reviewing LessWrong: Screwtape's Basic Answer | reviewing-lesswrong-screwtape-s-basic-answer | https://www.lesswrong.com/posts/wxKnzRjr8rt4jvmDE/reviewing-lesswrong-screwtape-s-basic-answer | Screwtape | 2025-02-05T04:30:34.347Z |
hEzmJooaFztEHEqpN | I agree, but it is important to note that the authors of the paper disagree here.
(It's somewhat hard for me to tell if the crux is more that they don't expect that everyone would get AI aligned to them (at least as representatives) even if this was technical feasible with zero alignment tax or if the crux is that eve... | 2025-02-07T21:49:17.586Z | 4 | gTF7xPEydCSCCuiqg | jEZpfsdaX2dBD9Y6g | The Risk of Gradual Disempowerment from AI | the-risk-of-gradual-disempowerment-from-ai | https://www.lesswrong.com/posts/jEZpfsdaX2dBD9Y6g/the-risk-of-gradual-disempowerment-from-ai | Zvi | 2025-02-05T22:10:06.979Z |
fC6GLXJ9pwqDaGqrR | This post seems to assume that research fields have big, hard central problems that are solved with some specific technique or paradigm.
This isn't always true. Many fields have the property that most of the work is on making small components work slightly better in ways that are very interoperable and don't have comp... | 2025-02-08T00:48:20.933Z | 115 | null | Hgj84BSitfSQnfwW6 | So You Want To Make Marginal Progress... | so-you-want-to-make-marginal-progress | https://www.lesswrong.com/posts/Hgj84BSitfSQnfwW6/so-you-want-to-make-marginal-progress | johnswentworth | 2025-02-07T23:22:19.825Z |
w2GyGdXe3nvEeZFac | > I would say it is basically-always true, but there are some fields (including deep learning today, for purposes of your comment) where the big hard central problems have already been solved, and therefore the many small pieces of progress on subproblems are all of what remains.
Maybe, but it is interesting to note t... | 2025-02-08T01:19:51.333Z | 27 | wS3BKxws36EcA8pr6 | Hgj84BSitfSQnfwW6 | So You Want To Make Marginal Progress... | so-you-want-to-make-marginal-progress | https://www.lesswrong.com/posts/Hgj84BSitfSQnfwW6/so-you-want-to-make-marginal-progress | johnswentworth | 2025-02-07T23:22:19.825Z |
hj4tFiLqKymgqJ8tx | > "fully handing over all technical and strategic work to AIs which are capable enough to obsolete humans at all cognitive tasks"
Suppose we replace "AIs" with "aliens" (or even, some other group of humans). Do you agree that doesn't (necessarily) kill you due to slop if you don't have a full solution to the superinte... | 2025-02-08T02:24:58.482Z | 9 | FgRu6jbwv2vtjhdWT | Hgj84BSitfSQnfwW6 | So You Want To Make Marginal Progress... | so-you-want-to-make-marginal-progress | https://www.lesswrong.com/posts/Hgj84BSitfSQnfwW6/so-you-want-to-make-marginal-progress | johnswentworth | 2025-02-07T23:22:19.825Z |
kcbBAiXJiyLGnw4v3 | I think you might first reach wildly superhuman AI via scaling up some sort of machine learning (and most of that is something well described as deep learning). Note that I said "needed". So, I would also count it as acceptable to build the AI with deep learning to allow for current tools to be applied even if somethin... | 2025-02-08T02:27:20.882Z | 10 | dhw2bq5qfXGbdRvdv | Hgj84BSitfSQnfwW6 | So You Want To Make Marginal Progress... | so-you-want-to-make-marginal-progress | https://www.lesswrong.com/posts/Hgj84BSitfSQnfwW6/so-you-want-to-make-marginal-progress | johnswentworth | 2025-02-07T23:22:19.825Z |
umpdkRGjyTbpd3fDa | Keeping the humans alive at this point is extremely cheap in terms of fraction of long term resource consumption while avoiding killing humans might substantially reduce the AI's chance of successful takeover. | 2025-02-08T20:46:39.638Z | 13 | GNrwexhMwTXLfKCzy | KFJ2LFogYqzfGB3uX | How AI Takeover Might Happen in 2 Years | how-ai-takeover-might-happen-in-2-years | https://x.com/joshua_clymer/status/1887905375082656117 | joshc | 2025-02-07T17:10:10.530Z |
c7BmXxbvEBiwgmcmG | For many people, "can the AIs actually take over" is a crux and seeing a story of this might help build some intuition. | 2025-02-08T22:21:51.407Z | 72 | Do9GwPgJMtx4R8wwH | KFJ2LFogYqzfGB3uX | How AI Takeover Might Happen in 2 Years | how-ai-takeover-might-happen-in-2-years | https://x.com/joshua_clymer/status/1887905375082656117 | joshc | 2025-02-07T17:10:10.530Z |
TQBY2EXg864idJNNf | I think there are good reasons to expect large fractions of humans might die even if humans immediately surrender:
- It might be an unstable position given that the AI has limited channels of influence on the physical world. (While if there are far fewer humans, this changes.)
- The AI might not care that much or migh... | 2025-02-09T03:49:27.402Z | 7 | ybhmDpymjiAmzFXrP | KFJ2LFogYqzfGB3uX | How AI Takeover Might Happen in 2 Years | how-ai-takeover-might-happen-in-2-years | https://x.com/joshua_clymer/status/1887905375082656117 | joshc | 2025-02-07T17:10:10.530Z |
fmsPnsL5is9mykf3Y | There are two interpretations you might have for that third bullet:
- Can we stop rogue AIs? (Which are operating without human supervision.)
- Can we stop AIs deployed in their intended context?
(See also [here](https://www.lesswrong.com/posts/ceBpLHJDdCt3xfEok/ai-catastrophes-and-rogue-deployments).)
In the contex... | 2025-02-09T04:50:19.992Z | 2 | MmeiCxPgiD6JWynsS | KFJ2LFogYqzfGB3uX | How AI Takeover Might Happen in 2 Years | how-ai-takeover-might-happen-in-2-years | https://x.com/joshua_clymer/status/1887905375082656117 | joshc | 2025-02-07T17:10:10.530Z |
cFY5FtbwxqRZ2H5vj | (I don't think it's good to add a canary in this case (the main concern would be takeover strategies, but I basically agree this isn't that helpful), but I think people might be reacting to "_might_ be worth adding" and are disagree reacting to your comment because it says "are you actually serious" which seems more di... | 2025-02-09T04:57:53.738Z | 9 | NRZ7jG2ZPcXebtzGN | KFJ2LFogYqzfGB3uX | How AI Takeover Might Happen in 2 Years | how-ai-takeover-might-happen-in-2-years | https://x.com/joshua_clymer/status/1887905375082656117 | joshc | 2025-02-07T17:10:10.530Z |
ce4CmnF6iRRfnGxGw | I doubt that person was thinking about the opaque vector reasoning making it harder to catch the rogue AIs. | 2025-02-09T05:06:30.435Z | 2 | stinxzztZH6XEefZC | KFJ2LFogYqzfGB3uX | How AI Takeover Might Happen in 2 Years | how-ai-takeover-might-happen-in-2-years | https://x.com/joshua_clymer/status/1887905375082656117 | joshc | 2025-02-07T17:10:10.530Z |
TvW7EQtkTgdYket7b | > You are possibly the first person I know of who reacted to MONA with "that's obvious"
I also have the "that's obvious reaction", but possibly I'm missing somne details. I also think it won't perform well enough in practice to pencil given other better places to allocate safety budget (if it does trade off which is u... | 2025-02-09T19:17:00.436Z | 2 | rbehQDsCgsfGXs2na | 26SHhxK2yYQbh7ors | Research directions Open Phil wants to fund in technical AI safety | research-directions-open-phil-wants-to-fund-in-technical-ai | https://www.openphilanthropy.org/tais-rfp-research-areas/ | jake_mendel | 2025-02-08T01:40:00.968Z |
AKrTWFiBht9cewWxJ | I'm a bit late for a review, but I've recently been reflecting on decision theory and this post came to mind.
When I initially saw this post I didn't make much of it. I now feel like the thesis of "decision theory is very confusing and messed up" is true, insightful, and pretty important based on spending more time en... | 2025-02-10T00:07:58.497Z | 9 | null | wXbSAKu2AcohaK2Gt | UDT shows that decision theory is more puzzling than ever | udt-shows-that-decision-theory-is-more-puzzling-than-ever | https://www.lesswrong.com/posts/wXbSAKu2AcohaK2Gt/udt-shows-that-decision-theory-is-more-puzzling-than-ever | Wei Dai | 2023-09-13T12:26:09.739Z |
sDZmmtu79JngjJCA2 | See appendix B.3 in particular:
> Competitors receive a higher score for submitting their solutions faster. Because models can think in parallel and simultaneously attempt all problems, they have an innate advantage over humans. We elected to reduce this advantage in our primary results by estimating o3’s score for ea... | 2025-02-12T17:30:19.113Z | 11 | kmeykjdbghfjhtopS | Ao4enANjWNsYiSFqc | o3 | o3 | https://www.lesswrong.com/posts/Ao4enANjWNsYiSFqc/o3 | Zach Stein-Perlman | 2024-12-20T18:30:29.448Z |
AaHAEJgEehenzEjBs | I would find this post much more useful to engage with if you more concretely described the type of tasks that you think AIs will remain bad and gave a bunch of examples. (Or at least made an argument for why it is hard to construct examples if that is your perspective.)
I think you're pointing to a category like "tas... | 2025-02-13T19:49:01.516Z | 66 | null | gsj3TWdcBxwkm9eNt | ≤10-year Timelines Remain Unlikely Despite DeepSeek and o3 | 10-year-timelines-remain-unlikely-despite-deepseek-and-o3 | https://www.lesswrong.com/posts/gsj3TWdcBxwkm9eNt/10-year-timelines-remain-unlikely-despite-deepseek-and-o3 | Rafael Harth | 2025-02-13T19:21:35.392Z |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.