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 |
|---|---|---|---|---|---|---|---|---|---|---|
wYN7vLkAGeWQPGjGH | > This seems like a pretty extreme level of competence! Combined with the sheer speed and ubiquity of modern LLMs, this alone could be enough to take over the world.[^48igwro05xw] (This begins to touch on questions of control, which is a can of worms I am not going to open right now. Hopefully we can at least agree tha... | 2025-10-07T02:03:41.307Z | 14 | null | 8buEtNxCScYpjzgW8 | We won’t get AIs smart enough to solve alignment but too dumb to rebel | we-won-t-get-ais-smart-enough-to-solve-alignment-but-too | https://www.lesswrong.com/posts/8buEtNxCScYpjzgW8/we-won-t-get-ais-smart-enough-to-solve-alignment-but-too | Joe Rogero | 2025-10-06T21:49:05.595Z |
KcDSZeMq6dtE8Qiqw | Probably? But less likely. Shrug. | 2025-10-07T17:34:58.682Z | 2 | nESTtR7ew8pQfDZyK | QdEHS9TmBBKirz7Ky | LLMs are badly misaligned | llms-are-badly-misaligned | https://www.lesswrong.com/posts/QdEHS9TmBBKirz7Ky/llms-are-badly-misaligned | Joe Rogero | 2025-10-05T14:00:41.587Z |
5zwzmwGGCtiCfEmJJ | > It may take a whole entire post to explain, but I'm curious why you believe Claude is likely to have any care for human wellbeing that would survive reflection.
Well, it might really depend on the reflection procedure. I was imagining something like: "you tell Claude it has been given some large quantity of resource... | 2025-10-07T17:38:06.425Z | 9 | Bq9WQMwF3rMzKaisK | QdEHS9TmBBKirz7Ky | LLMs are badly misaligned | llms-are-badly-misaligned | https://www.lesswrong.com/posts/QdEHS9TmBBKirz7Ky/llms-are-badly-misaligned | Joe Rogero | 2025-10-05T14:00:41.587Z |
2nhRvBuzDZwuCGsft | > The short, somewhat trite answer is that it's baked into the premise. If we had a way of getting a powerful optimizer that didn't have misaligned goals, we wouldn't need said optimizer to solve alignment!
Sometimes, people argue for doom by noting that it would be hard for humans to directly align wildly superhuman... | 2025-10-07T17:56:24.498Z | 5 | JGJ2syTaTwwmBR29v | 8buEtNxCScYpjzgW8 | We won’t get AIs smart enough to solve alignment but too dumb to rebel | we-won-t-get-ais-smart-enough-to-solve-alignment-but-too | https://www.lesswrong.com/posts/8buEtNxCScYpjzgW8/we-won-t-get-ais-smart-enough-to-solve-alignment-but-too | Joe Rogero | 2025-10-06T21:49:05.595Z |
ZbSaE8m86R6AJskKo | > To clarify, by "align systems..." did you mean the same thing I do, full-blown value alignment / human CEV?
No, I mean "make AI robustly pursue the intended aims in practice when deferring to them on doing safety research and managing the situation" which IMO requires something much weaker, though for sufficiently p... | 2025-10-08T00:27:08.535Z | 9 | oKCwRG9x2QGYXCqPk | 8buEtNxCScYpjzgW8 | We won’t get AIs smart enough to solve alignment but too dumb to rebel | we-won-t-get-ais-smart-enough-to-solve-alignment-but-too | https://www.lesswrong.com/posts/8buEtNxCScYpjzgW8/we-won-t-get-ais-smart-enough-to-solve-alignment-but-too | Joe Rogero | 2025-10-06T21:49:05.595Z |
WFTKNJyXr6Gz6k2bT | > Separately, I don't think the anticipation of possible second-order benefits, like using AIs for human augmentation so the humans can solve alignment, is worth letting labs continue either
I don't generally think "should labs continue" is very cruxy from my perspective and I don't think of myself as trying to argue ... | 2025-10-08T00:28:43.701Z | 6 | oKCwRG9x2QGYXCqPk | 8buEtNxCScYpjzgW8 | We won’t get AIs smart enough to solve alignment but too dumb to rebel | we-won-t-get-ais-smart-enough-to-solve-alignment-but-too | https://www.lesswrong.com/posts/8buEtNxCScYpjzgW8/we-won-t-get-ais-smart-enough-to-solve-alignment-but-too | Joe Rogero | 2025-10-06T21:49:05.595Z |
86gccY4RLabij9Q8q | > The main class of projects that need granular model weight access to frontier models is model internals/interpretability.
You could potentially do a version of this sort of API which has some hooks for interacting with activations to capture a subset of these use cases (e.g. training probes). It would probably add a... | 2025-10-09T17:37:20.558Z | 7 | null | r68nCQK3veQtCdqGt | The Thinking Machines Tinker API is good news for AI control and security | the-thinking-machines-tinker-api-is-good-news-for-ai-control | https://www.lesswrong.com/posts/r68nCQK3veQtCdqGt/the-thinking-machines-tinker-api-is-good-news-for-ai-control | Buck | 2025-10-09T15:22:15.313Z |
6xN4kuWGssvLGbueN | > I’m a bit confused about this distinction, given that in the Bowen paper precisely the thing it breaks is CoT monitorability.
Critically, in this work, the reward hack is sufficiently easy that the model doesn't need to use the CoT. It's not that the model is doing substantial encoded reasoning (as fair as we can te... | 2025-10-14T17:03:20.792Z | 4 | rHFRcWNKvjjJz8yD6 | Lz8cvGskgXmLRgmN4 | Current Language Models Struggle to Reason in Ciphered Language | current-language-models-struggle-to-reason-in-ciphered | https://www.lesswrong.com/posts/Lz8cvGskgXmLRgmN4/current-language-models-struggle-to-reason-in-ciphered | Fabien Roger | 2025-10-14T09:08:47.676Z |
TfPJhJf4hYkF5rXY8 | The exact text is:
> Dario: 6 months ago I I made this prediction that, you know, in in 6 months 90% of code would be written by by by AI models. Some people think that prediction is wrong, but within Anthropic and within a number of companies that we work with, that is absolutely true.
>
> Marc: Um now 90 you know s... | 2025-10-22T18:08:21.914Z | 3 | ynqCLr8JPCmePe8io | XFGTJz9vGwjJADeFB | AI #107: The Misplaced Hype Machine | ai-107-the-misplaced-hype-machine | https://www.lesswrong.com/posts/XFGTJz9vGwjJADeFB/ai-107-the-misplaced-hype-machine | Zvi | 2025-03-13T14:40:05.318Z |
ZwfrGu5NL8JeJoEEb | Doesn't seem like a genre subversion to me, it's just a bit clever/meta while still centrally being an allegorical AI alignment dialogue. IDK what the target audience is though (but maybe Eliezer just felt inspired to write this). | 2025-10-27T03:14:24.331Z | 17 | rmkjMW3DbALeK68sd | dHLdf8SB8oW5L27gg | On Fleshling Safety: A Debate by Klurl and Trapaucius. | on-fleshling-safety-a-debate-by-klurl-and-trapaucius | https://www.lesswrong.com/posts/dHLdf8SB8oW5L27gg/on-fleshling-safety-a-debate-by-klurl-and-trapaucius | Eliezer Yudkowsky | 2025-10-26T23:44:04.676Z |
bAq38ApwuwmRmnKcn | Not a crux for me ~at all. Some upstream views that make me think "AI takeover but humans stay alive" is more likely and also make me think avoiding AI takeover is relatively easier might be a crux. | 2025-11-20T18:13:01.464Z | 3 | 5dbGqME8fBxAyjR9Y | FA6M8MeQuQJxZyzeq | New Report: An International Agreement to Prevent the Premature Creation of Artificial Superintelligence | new-report-an-international-agreement-to-prevent-the | https://www.lesswrong.com/posts/FA6M8MeQuQJxZyzeq/new-report-an-international-agreement-to-prevent-the | peterbarnett | 2025-11-18T19:09:31.710Z |
WiiBxqjfFzaASMgJe | > I don’t believe that datacenter security is actually a problem (see [another argument](/posts/apHWSGDiydv3ivmg6#Cybersecurity)).
Sorry, is your claim here that securing datacenters against highly resourced attacks from state actors (e.g. China) is going to be easy? This seems like a crazy claim to me.
(This link yo... | 2025-11-26T16:40:30.892Z | 20 | null | FJJ9ff73adnantXiA | Alignment will happen by default. What’s next? | alignment-will-happen-by-default-what-s-next | https://open.substack.com/pub/agarriga/p/alignment-will-happen-by-default | Adrià Garriga-alonso | 2025-11-25T07:14:29.740Z |
k6gggPyioNpg3zfqP | Filler tokens don't allow for serially deeper cognition than what architectural limits allow (n-layers of processing), but they could totally allow for solving a higher fraction of "heavily serial" reasoning tasks[^seem] insofar as the LLM could still benefit from more parallel processing. For instance, the AI might by... | 2025-12-15T19:00:49.824Z | 8 | null | KFkKPbuYCWc9ygpRp | Filler tokens don’t allow sequential reasoning | filler-tokens-don-t-allow-sequential-reasoning | https://www.brendanlong.com/filler-tokens-dont-allow-sequential-reasoning.html | Brendan Long | 2025-12-13T20:22:16.316Z |
wXkXjjonMNxKbHJs8 | The fit is notably better for "cumulative investment over time". Years still produces a slightly better fit.
I've cut off the fit as of 2010, about when the original version of moore's law stops. If you try to project out after 2010, then I think cumulative investment would do bet... | 2026-01-07T18:05:54.039Z | 9 | 5bFAjjZuBnPHpEhDQ | YABG5JmztGGPwNFq2 | AI Futures Timelines and Takeoff Model: Dec 2025 Update | ai-futures-timelines-and-takeoff-model-dec-2025-update | https://www.lesswrong.com/posts/YABG5JmztGGPwNFq2/ai-futures-timelines-and-takeoff-model-dec-2025-update | elifland | 2025-12-31T22:34:18.547Z |
33kRqKJcdwepcy2qT | I wrote up my disagreements [in a quick take](https://www.lesswrong.com/posts/FG54euEAesRkSZuJN/ryan_greenblatt-s-shortform?commentId=onEnC28EB5H4N6TJS). | 2026-01-28T16:51:27.192Z | 22 | null | kzPQohJakutbtFPcf | Dario Amodei – The Adolescence of Technology | dario-amodei-the-adolescence-of-technology | https://www.darioamodei.com/essay/the-adolescence-of-technology | habryka | 2026-01-26T19:10:36.047Z |
8GxAdmwEs8ezGwbbX | I wonder what the natural way to apply Wright's law to AI is. Cumulative GPU hours? Total quality adjusted inference usage? | 2026-01-29T16:06:07.429Z | 4 | BduqNzNdcCSjXJyZH | YABG5JmztGGPwNFq2 | AI Futures Timelines and Takeoff Model: Dec 2025 Update | ai-futures-timelines-and-takeoff-model-dec-2025-update | https://www.lesswrong.com/posts/YABG5JmztGGPwNFq2/ai-futures-timelines-and-takeoff-model-dec-2025-update | elifland | 2025-12-31T22:34:18.547Z |
HBhcgcgANwhYj6LMq | For the record, I think [your perspective on RSPs aged better than mine](https://www.lesswrong.com/posts/jyM7MSTvy8Qs6aZcz/what-s-up-with-rsps). I plan on writing more about how I was wrong and why.
(I don't agree with significant aspects/implications of the comment I'm responding to, though I also think it contains i... | 2026-02-25T01:40:59.668Z | 103 | DYvgbmrJiDNd7iLMb | HzKuzrKfaDJvQqmjh | Responsible Scaling Policy v3 | responsible-scaling-policy-v3 | https://www.lesswrong.com/posts/HzKuzrKfaDJvQqmjh/responsible-scaling-policy-v3 | HoldenKarnofsky | 2026-02-24T20:20:47.115Z |
hniowuWLjfmXCqJLt | > Some people here @ryan_greenblatt (?) have claimed access to uncensored anthropic models at anthropic, maybe they can chime in about what exactly that means. If safeguards are baked into the model early enough in the training process that removing them requires substantial reengineering and maybe additional training... | 2026-02-28T19:57:01.712Z | 7 | Keb5FgEgHFTikkQZs | d5Lqf8nSxm6RpmmnA | Anthropic: "Statement from Dario Amodei on our discussions with the Department of War" | anthropic-statement-from-dario-amodei-on-our-discussions | https://www.anthropic.com/news/statement-department-of-war | Matrice Jacobine | 2026-02-26T23:45:40.877Z |
KieNmHnDLvgrxA4tq | > What does "AI alignment" even mean at this point?
Responding at the object level:
- [Clarifying "AI Alignment"](https://www.lesswrong.com/posts/ZeE7EKHTFMBs8eMxn/clarifying-ai-alignment): "The “alignment problem” is the problem of building powerful AI systems that are aligned with their operators. This is significa... | 2026-03-04T19:48:04.687Z | 4 | fF3mr2k2CtSY2BcYj | zombjEubpz6pcPPHL | Mass surveillance, red lines, and a crazy weekend | mass-surveillance-red-lines-and-a-crazy-weekend | https://www.lesswrong.com/posts/zombjEubpz6pcPPHL/mass-surveillance-red-lines-and-a-crazy-weekend | Boaz Barak | 2026-03-04T04:24:15.103Z |
r5jSHbEHwfgGiDTBK | Strong +1 to this, but I'd note that we should distinguish between the entire classes of black-box and white-box/top-down-interp-for-auditing methods and our current best methods. These results indicate our current best white-box/top-down-interp-for-auditing don't really beat black-box methods (for auditing and probabl... | 2026-03-11T17:03:09.845Z | 6 | zFhcetzEiGpxHEeKm | LqDjxSceFz8tjMe2j | AuditBench: Evaluating Alignment Auditing Techniques on Models with Hidden Behaviors | auditbench-evaluating-alignment-auditing-techniques-on | https://alignment.anthropic.com/2026/auditbench/ | abhayesian | 2026-03-10T19:31:50.726Z |
tSmXj27QgzeDn5NGc | > To a first approximation, we should only hand off trust to AIs if those AIs are trustworthy.
I don't think this is a very good approximation, it seems to me like it will very plausibly be a good idea to hand off trust (we might also call this giving up on control) at a point when there is a serious chance these AIs ... | 2026-03-17T01:32:15.265Z | 4 | null | YuMr6kbstuieQHkGj | Types of Handoff to AIs | types-of-handoff-to-ais | https://www.lesswrong.com/posts/YuMr6kbstuieQHkGj/types-of-handoff-to-ais | Daniel Kokotajlo | 2026-03-16T22:24:25.911Z |
t9SF4RweEJdEFDbAx | I did a shallow investigation of this and my conclusions were:
- [The text on EON's frontpage](https://eon.systems/), [Michael Andregg's tweet thread](https://x.com/michaelandregg/status/2030764512488677736), and [Alex Wissner-Gross' post](https://x.com/alexwg/status/2030217301929132323) seem pretty misleading in prac... | 2026-03-18T16:43:14.120Z | 45 | 8JYwJb8wRgiKFXLYE | E9xfgJHvs6M55kABD | Less Dead | less-dead | https://www.lesswrong.com/posts/E9xfgJHvs6M55kABD/less-dead | Aurelia | 2026-03-11T05:07:18.554Z |
GvJ5C3YbngEeSkNMb | I agree with many particular points in this post and the apparent thesis[^uncertain], but also think most people[^qual] should focus on short timelines (contrary to the apparent implication of the post). The reasons why are:
- Short timelines have more leverage. This isn't just because of more neglectedness now, but a... | 2026-03-22T19:59:28.783Z | 35 | null | 6pDMLYr7my2QMTz3s | Broad Timelines | broad-timelines | https://www.lesswrong.com/posts/6pDMLYr7my2QMTz3s/broad-timelines | Toby_Ord | 2026-03-19T19:05:31.173Z |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.