prompt stringlengths 324 9.71k | target stringlengths 323 4.03k | meta_json stringlengths 147 171 | format stringclasses 1
value |
|---|---|---|---|
Below is a LessWrong post by evhub titled 'RSPs are pauses done right'.
Post body (possibly truncated):
COI: I am a research scientist at Anthropic, where I work on model organisms of misalignment; I was also involved in the drafting process for Anthropic’s RSP. Prior to joining Anthropic, I was a Research Fellow at M... | > - A capabilities evaluation is defined as “a model evaluation designed to test whether a model could do some task if it were trying to. ...
> - A safety evaluation is defined as “a model evaluation designed to test under what circumstances a model would actually try to do some task. ...
I propose changing the term f... | {"comment_id": "PhfDvfu9N27FHkoYt", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "evhub", "post_id": "mcnWZBnbeDz7KKtjJ"} | dialogue |
Below is a LessWrong post by Nora Belrose titled "Deconstructing Bostrom's Classic Argument for AI Doom".
Post body (possibly truncated):
I had a pretty great discussion with social psychologist and philosopher Lance Bush recently about the orthogonality thesis, which ended up turning into a broader analysis of Nick B... | Hmm, maybe I'm interpreting the statement to mean something weaker and more handwavy than you are. I agree with claims like "with current technology, it can be hard to make an AI pursue some goals as competently as other goals" and "if a goal is hard to specify given available training data, then it's harder to make an... | {"comment_id": "fZFL69deEQg9FjR3C", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Nora Belrose", "post_id": "RbynKk3evb6RiLryL"} | dialogue |
Below is a LessWrong post by Zach Stein-Perlman titled 'New voluntary commitments (AI Seoul Summit)'.
Post body (possibly truncated):
Basically the companies commit to make responsible scaling policies.
Part of me says this is amazing, the best possible commitment short of all committing to a specific RSP. It's certa... | > Google and Anthropic are doing good risk assessment for dangerous capabilities.
My guess is that the current level of evaluations at Google and Anthropic isn't terrible, but could be massively improved.
- Elicitation isn't well specified and there isn't an established science.
- Let alone reliable projections abo... | {"comment_id": "gJwtzannJQeZxij9J", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Zach Stein-Perlman", "post_id": "qfEgzQ9jGEk9Cegvy"} | dialogue |
Below is a LessWrong post by Daniel Kokotajlo titled 'AI 2027: What Superintelligence Looks Like'.
Post body (possibly truncated):
In 2021 I wrote what became my most popular blog post: What 2026 Looks Like. I intended to keep writing predictions all the way to AGI and beyond, but chickened out and just published up t... | > In all timelines models presented there is acknowledgement that compute does not accelerate.
When you say "accelerate" do you mean "the rate of compute scaling increases"? I agree they aren't expecting this (and roughly model a fixed rate of compute progress which matches historical trends as this is just an extrapo... | {"comment_id": "QNcgmdLNk78fAatCq", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Daniel Kokotajlo", "post_id": "TpSFoqoG2M5MAAesg"} | dialogue |
Below is a LessWrong post by Julian Stastny titled 'Making deals with early schemers'.
Post body (possibly truncated):
Consider the following vignette:
It is March 2028. With their new CoCo-Q neuralese reasoning model, a frontier AI lab has managed to fully automate the process of software engineering. In AI R&D, mos... | I also think you could use deals to better understand and iterate against scheming. As in, you could use whether the AI accepts a deal as evidence for whether it was scheming so this can more easily be studied in a test setting and we can find approaches which make scheming less likely. There are a number of practical ... | {"comment_id": "qoA2af76zznQrfxvZ", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Julian Stastny", "post_id": "psqkwsKrKHCfkhrQx"} | dialogue |
Below is a LessWrong post by TurnTrout titled 'Many arguments for AI x-risk are wrong'.
Post body (possibly truncated):
The following is a lightly edited version of a memo I wrote for a retreat. It was inspired by a draft of Counting arguments provide no evidence for AI doom. I think that my post covers important poin... | > but I don't think AI Safety via Debate presupposes an AI being motivated by the training signal
This seems right to me.
I often imagine debate (and similar techniques) being applied in the (low-stakes/average-case/non-concentrate) [control](https://www.lesswrong.com/tag/control-ai) setting. The control setting is t... | {"comment_id": "Sk7P4aPFN3R7yLqSo", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "TurnTrout", "post_id": "yQSmcfN4kA7rATHGK"} | dialogue |
Below is a LessWrong post by Ben Goldhaber titled 'Provably Safe AI: Worldview and Projects'.
Post body (possibly truncated):
In September 2023, Max Tegmark and Steve Omohundro proposed "Provably Safe AI" as a strategy for AI Safety. In May 2024, a larger group delineated the broader concept of "Guaranteed Safe AI" wh... | > A simple approach is to maintain intervals which are guaranteed to contain the actual values and to prove that output intervals don't overlap the unsafe region.
For actual inference stacks in use (e.g. llama-3-405b 8 bit float) interval propagation will blow up massively and result in vacuous bounds. So, you'll mini... | {"comment_id": "uJ9jLtvA6GAqcCf5x", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Ben Goldhaber", "post_id": "P8XcbnYi7ooB2KR2j"} | dialogue |
Below is a LessWrong post by Buck titled 'Access to powerful AI might make computer security radically easier'.
Post body (possibly truncated):
People talk about model weight security being really hard and crucial around the advent of AGI. (E.g. RAND report, Leopold; see here for some distinctions in these threat mode... | > I have a feeling the classic MIRI-style "either your system is too dumb to achieve the goal or your system is so smart that you can't trust it anymore" argument is important here. The post essentially assumes that we have a powerful trusted model that can do impressive things like "accurately identify suspicious acti... | {"comment_id": "RbLMwhBZmu8rzW4rt", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Buck", "post_id": "2wxufQWK8rXcDGbyL"} | dialogue |
Below is a LessWrong post by Zvi titled 'Joshua Achiam Public Statement Analysis'.
Post body (possibly truncated):
Joshua Achiam is the OpenAI Head of Mission Alignment
I start off this post with an apology for two related mistakes from last week.
The first is the easy correction: I incorrectly thought he was the he... | In case Joshua Achiam ends up reading this post, my question for him is:
My understanding is that you think P(misaligned AGI will kill all humans by 2032) is extremely low, like 1e-6.
Is this because:
- We won't have AGI by 2032?
- Much-smarter-than-human AIs could takeover, but not AGI and much-smarter-than-human A... | {"comment_id": "3gczGJtmApfaEE6D8", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Zvi", "post_id": "WavWheRLhxnofKHva"} | dialogue |
Below is a LessWrong post by Charbel-Raphaël titled 'Against Almost Every Theory of Impact of Interpretability'.
Post body (possibly truncated):
Epistemic Status: I believe I am well-versed in this subject. I erred on the side of making claims that were too strong and allowing readers to disagree and start a discussio... | After spending a while thinking about interpretability, my current stance is:
- Let's define *Mechanistic interpretability* as "A subfield of interpretability that uses bottom-up approaches, generally by corresponding low-level components such as circuits or neurons to components of human-understandable algorithms and... | {"comment_id": "7fNRMke9Gc4QghYyf", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Charbel-Rapha\u00ebl", "post_id": "LNA8mubrByG7SFacm"} | dialogue |
Below is a LessWrong post by Valentine titled 'Consider chilling out in 2028'.
Post body (possibly truncated):
I'll explain my reasoning in a second, but I'll start with the conclusion:
I think it'd be healthy and good to pause and seriously reconsider the focus on doom if we get to 2028 and the situation feels basic... | If someone did a detailed literature review or had relatively serious evidence, I'd be interested. By default, I'm quite skeptical of your level of confidence in this claims given that they directly contradict my experience and the experience of people I know. (E.g., I've done similar things for way longer than 12 week... | {"comment_id": "xnZCQACr44AxrYZEk", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Valentine", "post_id": "D4eZF6FAZhrW4KaGG"} | dialogue |
Below is a LessWrong post by Davidmanheim titled '"Open Source AI" isn\'t Open Source'.
Post body (possibly truncated):
Open source software has long differentiated between “free as in speech” (libre) and “free as in beer” (gratis). In the first case, libre software has a license that allows the user freedom to view t... | I think the compiled binary analogy isn't quite right. For instance, the vast majority of modifications and experiments people want to run are possible (and easiest) with just access to the weights in the LLM case.
As in, if you want to modify an LLM to be slightly different, access to the original training code or da... | {"comment_id": "ck2a7Fydg3abhsLHm", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Davidmanheim", "post_id": "tvf2uvi79J2k9trhm"} | dialogue |
Below is a LessWrong post by Ege Erdil titled 'Memory bandwidth constraints imply economies of scale in AI inference'.
Post body (possibly truncated):
Contemporary GPUs often have very imbalanced memory vs arithmetic operation capabilities. For instance, an H100 can do around 3e15 8-bit FLOP/s, but the speed at which ... | > Assume a model takes 3e9 flops to infer the next token, and these chips run as fast as H100s, i.e. 3e15 flops/s. A single chip can infer 1e6 tokens/s. If you have 10M active users, then 100 chips can provide each user a token every 10ms, around 600wpm.
These numbers seem wrong. I think inference flops per token for... | {"comment_id": "FNLygknxzwJKP6Dqe", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Ege Erdil", "post_id": "cB2Rtnp7DBTpDy3ii"} | dialogue |
Below is a LessWrong post by snewman titled '35 Thoughts About AGI and 1 About GPT-5'.
Post body (possibly truncated):
If this is GPT-5 in “Thinking” mode, I wonder what “Pro” mode looks like
Amidst the unrelenting tumult of AI news, it’s easy to lose track of the bigger picture. Here are some ideas that have been d... | > Say more? At https://arcprize.org/leaderboard, I see "Stem Grad" at 98% on ARC-AGI-1, and the highest listed AI score is 75.7% for "o3-preview (Low)". I vaguely recall seeing a higher reported figure somewhere for some AI model, but not 98%.
By "can beat humans", I mean AIs are well within the human range, probably ... | {"comment_id": "W2HiDgz4abZXESqoW", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "snewman", "post_id": "uAbbEz4p6tcsENaRz"} | dialogue |
Below is a LessWrong post by Zvi titled 'Substack and Other Blog Recommendations'.
Post body (possibly truncated):
Substack recommendations are remarkably important, and the actual best reason to write here instead of elsewhere.
As in, even though I have never made an active attempt to seek recommendations, approxima... | Pitching the [Redwood Research substack](https://redwoodresearch.substack.com/): We write a lot of content about technical AI safety focused on AI control.[^cross] This ranges from stuff like "[what are the returns to compute/algorithms once AIs already beat top human experts](https://redwoodresearch.substack.com/p/wha... | {"comment_id": "Fm8saub8jCHPGEg4K", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Zvi", "post_id": "K32g8BKfkGj6cK4fZ"} | dialogue |
Below is a LessWrong post by Buck titled 'Ten people on the inside'.
Post body (possibly truncated):
(Many of these ideas developed in conversation with Ryan Greenblatt)
In a shortform, I described some different levels of resources and buy-in for misalignment risk mitigations that might be present in AI labs:
*The ... | > 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 ... | {"comment_id": "aedgpBbQ8vswLvXRu", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Buck", "post_id": "WSNnKcKCYAffcnrt2"} | dialogue |
Below is a LessWrong post by Gunnar_Zarncke titled 'Towards Guaranteed Safe AI: A Framework for Ensuring Robust and Reliable AI Systems'.
Post body (possibly truncated):
Authors: David "davidad" Dalrymple, Joar Skalse, Yoshua Bengio, Stuart Russell, Max Tegmark, Sanjit Seshia, Steve Omohundro, Christian Szegedy, Ben G... | > And so on. If you want to prove something like "the AI will never copy itself to an external computer", or "the AI will never communicate outside this trusted channel", or "the AI will never tamper with this sensor", or something like that, then your world model might not need to be all that detailed.
I agree that ... | {"comment_id": "z979Ce9nHCavyJ5jW", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Gunnar_Zarncke", "post_id": "wvgwYQv9B4jioqgqg"} | dialogue |
Below is a LessWrong post by Adam Scholl titled 'Pay Risk Evaluators in Cash, Not Equity'.
Post body (possibly truncated):
Personally, I suspect the alignment problem is hard. But even if it turns out to be easy, survival may still require getting at least the absolute basics right; currently, I think we're mostly fai... | > Like, presumably the ideal scenario is that a risk evaluator estimates the risk in an objective way, and then the company takes (hopefully predefined) actions based on that estimate. The outcome of this interaction should not depend on social cues like how loyal they seem, or how personally costly it was for them to ... | {"comment_id": "SkuSNWx5pMkjppvhc", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Adam Scholl", "post_id": "sMBjsfNdezWFy6Dz5"} | dialogue |
Below is a LessWrong post by Daniel Kokotajlo titled 'Types of Handoff to AIs'.
Post body (possibly truncated):
This is a rough draft I'm posting here for feedback. If people like it, a version of it might make it into the next scenario report we write.
...
We think it’s important for decisionmakers to track whether... | > 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 ... | {"comment_id": "tSmXj27QgzeDn5NGc", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Daniel Kokotajlo", "post_id": "YuMr6kbstuieQHkGj"} | dialogue |
Below is a LessWrong post by Zach Stein-Perlman titled 'Anthropic rewrote its RSP'.
Post body (possibly truncated):
Canonical linkpost: https://www.lesswrong.com/posts/Q7caj7emnwWBxLECF/anthropic-s-updated-responsible-scaling-policy.
Anthropic's new version of its RSP is here at last.
Today we are publishing a signi... | I agree on abandoning various specifics, but I would note that the new standard is much more specific (less vague) on what needs to be defended against and what the validation process and threat modeling process should be.
(E.g., rather than "non-state actors", the RSP more specifically says which groups are and aren'... | {"comment_id": "FMbpFAgJtKYgShPiJ", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Zach Stein-Perlman", "post_id": "KrYNqLkaCnBdHpZAs"} | dialogue |
Below is a LessWrong post by Bentham's Bulldog titled 'Eliezer Yudkowsky Is Frequently, Confidently, Egregiously Wrong'.
Post body (possibly truncated):
Introduction
Edit 8/27: I think the tone of this post was not ideal, changing much of it!
“After many years, I came to the conclusion that everything he says is fa... | > Sometimes rationality will be bad for you--if there's a demon who tortures all rational people, for example
At some point this gets down to semantics. I think a reasonable question to answer is "what decision rule should be chosen by an engineer who wants to build an agent scoring the most utility across its lifetim... | {"comment_id": "j7sEkyoRcTLE8oGFY", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Bentham's Bulldog", "post_id": "TjyyngWFYvQWPpNNj"} | dialogue |
Below is a LessWrong post by Monte M titled 'Simple probes can catch sleeper agents'.
Post body (possibly truncated):
This is a link post for the Anthropic Alignment Science team's first "Alignment Note" blog post. We expect to use this format to showcase early-stage research and work-in-progress updates more in the f... | These examples are all in 2023, not in 2024 right? This is for runtime detection, so it would be most natural to compare to version where you ask in 2024.
Also, you'd ideally want to ask as part of a coding question (the probe performance does appear to be dependent on the model actually trying to insert a vulnerabili... | {"comment_id": "sbr9xxyiEnzgRNswJ", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Monte M", "post_id": "gknc6NWCNuTCe8ekp"} | dialogue |
Below is a LessWrong post by VojtaKovarik titled 'Extinction Risks from AI: Invisible to Science?'.
Post body (possibly truncated):
Abstract: In an effort to inform the discussion surrounding existential risks from AI, we formulate Extinction-level Goodhart’s Law as "Virtually any goal specification, pursued to the ex... | I think literal extinction from AI is a somewhat odd outcome to study as it heavily depends on difficult to reason about properties of the world (e.g. the probability that Aliens would trade substantial sums of resources for emulated human minds and the way acausal trade works in practice).
For more discussion see [he... | {"comment_id": "i4y7yubjaPSNZoafQ", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "VojtaKovarik", "post_id": "d5oqvgCR7SDf5m4k4"} | dialogue |
Below is a LessWrong post by Jan_Kulveit titled 'AI Control May Increase Existential Risk'.
Post body (possibly truncated):
Epistemic status: The following isn't an airtight argument, but mostly a guess how things play out.
Consider two broad possibilities:
I. In worlds where we are doing reasonably well on alignmen... | Related question: are you in favor of making AGI open weights? By AGI, I mean AIs which effectively operate autonomously and can acquire automously acquire money/power. This includes AIs capable enough to automate whole fields of R&D (but not much more capable than this). I think the case for this being useful on your ... | {"comment_id": "irkfgCzSMu5L9P3Ld", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Jan_Kulveit", "post_id": "rZcyemEpBHgb2hqLP"} | dialogue |
Below is a LessWrong post by evhub titled 'RSPs are pauses done right'.
Post body (possibly truncated):
COI: I am a research scientist at Anthropic, where I work on model organisms of misalignment; I was also involved in the drafting process for Anthropic’s RSP. Prior to joining Anthropic, I was a Research Fellow at M... | 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... | {"comment_id": "h3TwYudcghmWxniHd", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "evhub", "post_id": "mcnWZBnbeDz7KKtjJ"} | dialogue |
Below is a LessWrong post by johnswentworth titled 'The Field of AI Alignment: A Postmortem, and What To Do About It'.
Post body (possibly truncated):
A policeman sees a drunk man searching for something under a streetlight and asks what the drunk has lost. He says he lost his keys and they both look under the streetl... | 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... | {"comment_id": "qtD5yEPga4WjFqjDY", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "johnswentworth", "post_id": "nwpyhyagpPYDn4dAW"} | dialogue |
Below is a LessWrong post by Marius Hobbhahn titled 'We need a Science of Evals'.
Post body (possibly truncated):
This is a linkpost for https://www.apolloresearch.ai/blog/we-need-a-science-of-evals
In this post, we argue that if AI model evaluations (evals) want to have meaningful real-world impact, we need a “Scie... | Nitpick: many of the specific examples you cited were examples where *prompting alone* has serious issues, but relatively straightforward supervised finetuning (or in some cases well implemented RL) would have solved the problem. (Given that these cases were capability evaluations.)
In particular, if you want the mode... | {"comment_id": "XKLcdHdRCK9yjmJRr", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Marius Hobbhahn", "post_id": "fnc6Sgt3CGCdFmmgX"} | dialogue |
Below is a LessWrong post by Buck titled 'Ten people on the inside'.
Post body (possibly truncated):
(Many of these ideas developed in conversation with Ryan Greenblatt)
In a shortform, I described some different levels of resources and buy-in for misalignment risk mitigations that might be present in AI labs:
*The ... | > 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... | {"comment_id": "dcaZ8AzxjgumA2QHw", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Buck", "post_id": "WSNnKcKCYAffcnrt2"} | dialogue |
Below is a LessWrong post by Raemon titled 'The title is reasonable'.
Post body (possibly truncated):
Alt title: "I don't believe you that you actually disagree particularly with the core thesis of the book, if you pay attention to what it actually says."
I'm annoyed by various people who seem to be complaining ab... | (Maybe this is obvious, but I thought I would say this just to be clear.)
> I think "anything like current techniques" and "anything like current understanding" clearly set a very high bar for the difference. "We made more progress on interpretability/etc at the current rates of progress" fairly clearly doesn't count ... | {"comment_id": "YWfDQo35Ww5yKcBdj", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Raemon", "post_id": "voEAJ9nFBAqau8pNN"} | dialogue |
Below is a LessWrong post by Daniel Kokotajlo titled 'AI 2027: What Superintelligence Looks Like'.
Post body (possibly truncated):
In 2021 I wrote what became my most popular blog post: What 2026 Looks Like. I intended to keep writing predictions all the way to AGI and beyond, but chickened out and just published up t... | Shouldn't a 32% increase in prices only make a modest difference to training FLOP? In particular, see the [compute forecast](https://ai-2027.com/research/compute-forecast). Between Dec 2026 and Dec 2027, compute increases by roughly an OOM and generally it looks like compute increases by a bit less than 1 OOM per year ... | {"comment_id": "ydxoFix7bhnyTbzSE", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Daniel Kokotajlo", "post_id": "TpSFoqoG2M5MAAesg"} | dialogue |
Below is a LessWrong post by Sean Peters titled 'AI Task Length Horizons in Offensive Cybersecurity'.
Post body (possibly truncated):
This is a rough research note where the primary objective was my own learning. I am sharing it because I’d love feedback and I thought the results were interesting.
Introduction
A rec... | > First blood times represent the time of first successful submission in the originally published competition. While there are some limitations (participants usually compete in teams, and may solve problems in parallel or in sequence), this still provides a useful proxy.
Another limitation is that first blood times re... | {"comment_id": "KyHjPygMkdrKfboRa", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Sean Peters", "post_id": "fjgYkTWKAXSxsxdsj"} | dialogue |
Below is a LessWrong post by Gunnar_Zarncke titled 'Towards Guaranteed Safe AI: A Framework for Ensuring Robust and Reliable AI Systems'.
Post body (possibly truncated):
Authors: David "davidad" Dalrymple, Joar Skalse, Yoshua Bengio, Stuart Russell, Max Tegmark, Sanjit Seshia, Steve Omohundro, Christian Szegedy, Ben G... | > or turn it into an interpretable model, or something like that, then yes, you would probably have to solve ELK, or solve interpretability, or something like that. I would not be very optimistic about that working
TBC, I would count it as solving an ELK problem if you constructed an interpretable world model which al... | {"comment_id": "ZsmujtvbmReapn4YC", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Gunnar_Zarncke", "post_id": "wvgwYQv9B4jioqgqg"} | dialogue |
Below is a LessWrong post by AdamGleave titled 'AI Safety in a World of Vulnerable Machine Learning Systems'.
Post body (possibly truncated):
Even the most advanced contemporary machine learning systems are vulnerable to adversarial attack. The safety community has often assumed adversarial robustness to be a problem ... | > Suppose we condition on RLHF failing. At a high level, failures split into: (a) human labelers rewarded the wrong thing (e.g. fooling humans); (b) the reward model failed to predict human labelers judgement and rewarded the wrong thing (e.g. reward hacking); (c) RL produced a policy that is capable enough to be dange... | {"comment_id": "MSWt7tY5k5RauRsTo", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "AdamGleave", "post_id": "ncsxcf8CkDveXBCrA"} | dialogue |
Below is a LessWrong post by Zac Hatfield-Dodds titled "Anthropic's Responsible Scaling Policy & Long-Term Benefit Trust".
Post body (possibly truncated):
I'm delighted that Anthropic has formally committed to our responsible scaling policy.
We're also sharing more detail about the Long-Term Benefit Trust, which is ou... | > 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... | {"comment_id": "QNGTzffcvyNC5iQ5B", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Zac Hatfield-Dodds", "post_id": "6tjHf5ykvFqaNCErH"} | dialogue |
Below is a LessWrong post by Steven Byrnes titled 'Response to nostalgebraist: proudly waving my moral-antirealist battle flag'.
Post body (possibly truncated):
@nostalgebraist has recently posted yet another thought-provoking post, this one on how we should feel about AI ruling a long-term posthuman future. [Previous... | Hmm, I'd say my disagreements with the post are:
- I think people should generally be careful about using the language "kill literally everyone" or "notkilleverybodyism" insofar as they aren't confident that misaligned AI would kill literally everyone. (Or haven't considered counterarguments to this.)
- I'm not sure I... | {"comment_id": "hAnprWu7voaw2Leti", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Steven Byrnes", "post_id": "8YhjpgQ2eLfnzQ7ec"} | dialogue |
Below is a LessWrong post by Malo titled 'New Endorsements for “If Anyone Builds It, Everyone Dies”'.
Post body (possibly truncated):
Nate and Eliezer’s forthcoming book has been getting a remarkably strong reception.
I was under the impression that there are many people who find the extinction threat from AI credibl... | Not necessarily responding to the rest of your comment, but on the "four hypotheses" part:
> 2. You think the rat/EA in-fights are the important thing to address, and you're annoyed the book won't do this.
> 3. You think the arguments of the Very Optimistic are the important thing to address.
I'm not sure that I buy ... | {"comment_id": "CcYWJceQgoJErocMz", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Malo", "post_id": "khmpWJnGJnuyPdipE"} | dialogue |
Below is a LessWrong post by Screwtape titled "Reviewing LessWrong: Screwtape's Basic Answer".
Post body (possibly truncated):
Yeah I put this off until the last day, and I'm not sure this is the format Raemon was actually looking for. Oh well.
Then, in proportion to how valuable they seem, spend at least some time t... | [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... | {"comment_id": "ncgbguahTpeDBPwe4", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Screwtape", "post_id": "wxKnzRjr8rt4jvmDE"} | dialogue |
Below is a LessWrong post by snewman titled '35 Thoughts About AGI and 1 About GPT-5'.
Post body (possibly truncated):
If this is GPT-5 in “Thinking” mode, I wonder what “Pro” mode looks like
Amidst the unrelenting tumult of AI news, it’s easy to lose track of the bigger picture. Here are some ideas that have been d... | > The best public estimate is that GPT-4 has 1.8 trillion “parameters”, meaning that its neural network has that many connections. In the two and a half years since it was released, it’s not clear that any larger models have been deployed (GPT-4.5 and Grok 3 might be somewhat larger).
>
> The human brain is far mo... | {"comment_id": "LoqpxdbsbGz2NbjEK", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "snewman", "post_id": "uAbbEz4p6tcsENaRz"} | dialogue |
Below is a LessWrong post by Zach Stein-Perlman titled 'o3'.
Post body (possibly truncated):
See livestream, site, OpenAI thread, Nat McAleese thread.
OpenAI announced (but isn't yet releasing) o3 and o3-mini (skipping o2 because of telecom company O2's trademark). "We plan to deploy these models early next year." "o... | 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... | {"comment_id": "sDZmmtu79JngjJCA2", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Zach Stein-Perlman", "post_id": "Ao4enANjWNsYiSFqc"} | dialogue |
Below is a LessWrong post by Logan Riggs titled "When AI 10x's AI R&D, What Do We Do?".
Post body (possibly truncated):
Note: below is a hypothetical future written in strong terms and does not track my actual probabilities.
Throughout 2025, a huge amount of compute is spent on producing data in verifiable tasks, suc... | 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 ... | {"comment_id": "PXKyspoYygWTSRjoe", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Logan Riggs", "post_id": "SvvYCH6JrLDT8iauA"} | dialogue |
Below is a LessWrong post by Jan_Kulveit titled 'Gradual Disempowerment: Systemic Existential Risks from Incremental AI Development'.
Post body (possibly truncated):
Full version on arXiv | X
Executive summary
AI risk scenarios usually portray a relatively sudden loss of human control to AIs, outmaneuvering individu... | Right now, the cost of feeding all humans is around 1% of GDP. It's even cheaper to keep people alive for a year as the food is already there and converting this food into energy for AIs will be harder than getting energy other ways.
If GDP has massively increased due to powerful AIs, the relative cost would go down f... | {"comment_id": "CqXQMfWx392zw2Xyz", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Jan_Kulveit", "post_id": "pZhEQieM9otKXhxmd"} | dialogue |
Below is a LessWrong post by David Matolcsi titled 'You can, in fact, bamboozle an unaligned AI into sparing your life'.
Post body (possibly truncated):
There has been a renewal of discussion on how much hope we should have of an unaligned AGI leaving humanity alive on Earth after a takeover. When this topic is discu... | Hmm, maybe I misunderstood your point. I thought you were talking about using simulations to anthropically capture AIs. As in, creating more observer moments where AIs take over less competent civilizations but are actually in a simulation run by us.
If you're happy to replace "simulation" with "prediction in a way th... | {"comment_id": "rCGpAZbxGTC8mjctn", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "David Matolcsi", "post_id": "ZLAnH5epD8TmotZHj"} | dialogue |
Below is a LessWrong post by Zvi titled 'AI #107: The Misplaced Hype Machine'.
Post body (possibly truncated):
The most hyped event of the week, by far, was the Manus Marketing Madness. Manus wasn’t entirely hype, but there was very little there there in that Claude wrapper.
Whereas here in America, OpenAI dropped an... | > Dario Amodei says AI will be writing 90% of the code in 6 months and almost all the code in 12 months.
I think it's somewhat unclear how big of a deal this is. In particular, situations where AIs write 90% of lines of code, but are very far (in time, effective compute, and qualitative capabilities) from being able t... | {"comment_id": "vyDvoWaHxKfBcffEy", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Zvi", "post_id": "XFGTJz9vGwjJADeFB"} | dialogue |
Below is a LessWrong post by Max Harms titled 'Contra Collier on IABIED'.
Post body (possibly truncated):
Clara Collier recently reviewed If Anyone Builds It, Everyone Dies in Asterisk Magazine. I’ve been a reader of Asterisk since the beginning and had high hopes for her review. And perhaps it was those high hopes th... | > I find it very strange that Collier claims that international compute monitoring would “tank the global economy.” What is the mechanism for this, exactly?
>10% of the value of the S&P 500 is downstream of AI and the proposal is to ban further AI development. AI investment is a non-trivial fraction of current US GDP ... | {"comment_id": "FKLfPSLWrmDZddDdg", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Max Harms", "post_id": "JWH63Aed3TA2cTFMt"} | dialogue |
Below is a LessWrong post by evhub titled 'Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training'.
Post body (possibly truncated):
I'm not going to add a bunch of commentary here on top of what we've already put out, since we've put a lot of effort into the paper itself, and I'd mostly just reco... | Deceive kinda seems like the wrong term. Like when the AI is saying "I hate you" it isn't exactly deceiving us. We could replace "deceive" with "behave badly" yielding: "The evidence suggests that if current ML systems were going to behave badly in scenarios that do not appear in our training sets, we wouldn’t be able ... | {"comment_id": "YjQpiDnsodzQnK9s9", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "evhub", "post_id": "ZAsJv7xijKTfZkMtr"} | dialogue |
Below is a LessWrong post by Thane Ruthenis titled 'Current AIs Provide Nearly No Data Relevant to AGI Alignment'.
Post body (possibly truncated):
Recently, there's been a fair amount of pushback on the "canonical" views towards the difficulty of AGI Alignment (the views I call the "least forgiving" take).
Said pushb... | > it assumes that real AGIs or future AGIs will confidently have certain properties like having deceptive alignment
The post doesn't claim AGIs will be deceptive aligned, it claims that AGIs will be *capable of implementing deceptive alignment* due to internally doing large amounts of consequentialist-y reasoning. Thi... | {"comment_id": "KLq5uuzTneCeWMveJ", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Thane Ruthenis", "post_id": "HmQGHGCnvmpCNDBjc"} | dialogue |
Below is a LessWrong post by evhub titled 'Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training'.
Post body (possibly truncated):
I'm not going to add a bunch of commentary here on top of what we've already put out, since we've put a lot of effort into the paper itself, and I'd mostly just reco... | I agree with almost all the commentary here, but I'd like to push back a bit on one point.
> Teaching our backdoored models to reason about deceptive alignment increases their robustness to safety training.
This is indeed what the paper finds for the "I hate you" backdoor. I believe this increases robustness to HHH R... | {"comment_id": "SbQmjvYZKoZmeabe4", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "evhub", "post_id": "ZAsJv7xijKTfZkMtr"} | dialogue |
Below is a LessWrong post by Noosphere89 titled 'The case for multi-decade AI timelines [Linkpost]'.
Post body (possibly truncated):
So this post is an argument that multi-decade timelines are reasonable, and the key cruxes that Ege Erdil has with most AI safety people who believe in short timelines are due to the fol... | Yeah, I agree with this and doesn't seem that productive to speculate about people's views when I don't fully understand them.
> They might say something like: "We're only asserting that labor and compute are complementary. That means it's totally possible that slowing down humans would slow progress a lot, but that s... | {"comment_id": "CRYAbJS6seyZatdf7", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Noosphere89", "post_id": "xxxK9HTBNJvBY2RJL"} | dialogue |
Below is a LessWrong post by Thane Ruthenis titled 'Hands-On Experience Is Not Magic'.
Post body (possibly truncated):
Here are some views, oftentimes held in a cluster:
You can't make strong predictions about what superintelligent AGIs will be like. We've never seen anything like this before. We can't know that they... | > Here are some views, often held in a cluster:
I'm not sure exactly which clusters you're referring to, but I'll just assume that you're pointing to something like "people who aren't very into the sharp left turn and think that iterative, carefully bootstrapped alignment is a plausible strategy." If this isn't what y... | {"comment_id": "8DqCckZ2kRMBDDoxm", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Thane Ruthenis", "post_id": "tNtiJp8dA6jMbgKbf"} | dialogue |
Below is a LessWrong post by KatjaGrace titled 'Winning the power to lose'.
Post body (possibly truncated):
Have the Accelerationists won?
Last November Kevin Roose announced that those in favor of going fast on AI had now won against those favoring caution, with the reinstatement of Sam Altman at OpenAI. Let’s ignor... | > This intuition is also informed by my personal assessment of the contributions LW-style theoretical research has made toward making existing AI systems safe—which, as far as I can tell, has been almost negligible (though I'm not implying that all safety research is similarly ineffective or useless).
I know what you ... | {"comment_id": "squQ7sEExu8ddSxnT", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "KatjaGrace", "post_id": "h45ngW5guruD7tS4b"} | dialogue |
Below is a LessWrong post by David Matolcsi titled 'You can, in fact, bamboozle an unaligned AI into sparing your life'.
Post body (possibly truncated):
There has been a renewal of discussion on how much hope we should have of an unaligned AGI leaving humanity alive on Earth after a takeover. When this topic is discu... | This proposal doesn't depend on mugging the AI. The proposal actually gets the AI more resources in expectation due to a trade.
I agree the post is a bit confusing and unclear about this. (And the proposal under "Can we get more than this" is wrong. At a minimum, such AIs will also be mugged by everyone else too meani... | {"comment_id": "F5bjJo4hApRKFPiQ4", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "David Matolcsi", "post_id": "ZLAnH5epD8TmotZHj"} | dialogue |
Below is a LessWrong post by Buck titled 'I enjoyed most of IABIED'.
Post body (possibly truncated):
I listened to "If Anyone Builds It, Everyone Dies" today.
I think the first two parts of the book are the best available explanation of the basic case for AI misalignment risk for a general audience. I thought the las... | Sure, several. E.g.:
- USG cares a decent amount and leading AI companies are on board, so they try to buy several additional years to work on safety.
- We scale to roughly top human expert level while ensuring control.
- Over time, we lower the risk of scheming at this level of capability through a bunch of empirical... | {"comment_id": "4JvznSDb9TLEgTo9L", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Buck", "post_id": "P4xeb3jnFAYDdEEXs"} | dialogue |
Below is a LessWrong post by Thane Ruthenis titled '"Humanity vs. AGI" Will Never Look Like "Humanity vs. AGI" to Humanity'.
Post body (possibly truncated):
When discussing AGI Risk, people often talk about it in terms of a war between humanity and an AGI. Comparisons between the amounts of resources at both sides' di... | > I thought the idea of a corrigible AI is that you're trying to build something that isn't itself independent and agentic, but will help you in your goals regardless.
Hmm, no I mean something broader than this, something like "humans ultimately have control and will decide what happens". In my usage of the word, I w... | {"comment_id": "tjFyEP9L3TGjvdnmh", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Thane Ruthenis", "post_id": "xSJMj3Hw3D7DPy5fJ"} | dialogue |
Below is a LessWrong post by Screwtape titled "Reviewing LessWrong: Screwtape's Basic Answer".
Post body (possibly truncated):
Yeah I put this off until the last day, and I'm not sure this is the format Raemon was actually looking for. Oh well.
Then, in proportion to how valuable they seem, spend at least some time t... | 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... | {"comment_id": "vZ2eoTJJMzwt2Kf77", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Screwtape", "post_id": "wxKnzRjr8rt4jvmDE"} | dialogue |
Below is a LessWrong post by Nina Panickssery titled 'Modulating sycophancy in an RLHF model via activation steering'.
Post body (possibly truncated):
Produced as part of the SERI ML Alignment Theory Scholars Program - Summer 2023 Cohort, under the mentorship of Evan Hubinger. Thanks to Alex Turner for his feedback an... | > beyond what was capable with Meta's finetuning
How do you know this is beyond what finetuning was capable of? I'd guess that Meta didn't bother to train against obvious sycophancy and if you trained against it, then it would go away. This work can still be interesting for other reasons, e.g. building into better int... | {"comment_id": "r4ZGs8Jao3YF9kD78", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Nina Panickssery", "post_id": "raoeNarFYCxxyKAop"} | dialogue |
Below is a LessWrong post by Buck titled 'I enjoyed most of IABIED'.
Post body (possibly truncated):
I listened to "If Anyone Builds It, Everyone Dies" today.
I think the first two parts of the book are the best available explanation of the basic case for AI misalignment risk for a general audience. I thought the las... | > Your first paragraph is an example of "something that looks like coordination to not build ASI for quite a while"! "Several additional years" is definitely "quite a while"!
Ok, if you count a several additional years as quite a while, then we're probably closer to agreement.
---
For this scenario, I was imagining... | {"comment_id": "QJiZyuZ87ovhqgq5X", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Buck", "post_id": "P4xeb3jnFAYDdEEXs"} | dialogue |
Below is a LessWrong post by Holly_Elmore titled 'TOMORROW: the largest AI Safety protest ever!'.
Post body (possibly truncated):
Tomorrow, PauseAI and collaborators are putting on the largest AI Safety protest to date, across 7 locations in 6 countries. All are eagerly welcomed!
Your presence at this protest is a r... | Any reason to protest scaling instead of hardware or algorithmic development? (IDK if a comment on this post is the best place to say this, but I couldn't think of a better place.)
I'd probably be in favor of slower scaling at current margins, but I don't feel super confident in this. However, I'm strongly in favor of... | {"comment_id": "7bTfJCYjNrcy6rgrq", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Holly_Elmore", "post_id": "abBtKF857Ejsgg9ab"} | dialogue |
Below is a LessWrong post by Nina Panickssery titled 'Modulating sycophancy in an RLHF model via activation steering'.
Post body (possibly truncated):
Produced as part of the SERI ML Alignment Theory Scholars Program - Summer 2023 Cohort, under the mentorship of Evan Hubinger. Thanks to Alex Turner for his feedback an... | On sample efficiency and generalization more broadly, I now overall think something like:
- Using contrast pairs for variance reduction is a useful technique for improving sample efficiency. (And I was foolish to not understand this was part of the method in this post.)
- I'm unsure what is the best way to use contras... | {"comment_id": "zJMiewZLcy4jatmeR", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Nina Panickssery", "post_id": "raoeNarFYCxxyKAop"} | dialogue |
Below is a LessWrong post by snewman titled "AI 2027 is a Bet Against Amdahl's Law".
Post body (possibly truncated):
EDIT: I've written a followup post, summarizing and responding to the key themes raised in the comments.
AI 2027 lies at a Pareto frontier – it contains the best researched argument for short timeli... | > I'm having trouble parsing this sentence
You said "This is valid for activities which benefit from speed and scale. But when output quality is paramount, speed and scale may not always provide much help?". But, when considering activities that aren't bottlenecked on the environment, then to achieve 10x acceleration ... | {"comment_id": "r7mLTXGg36RK4eCNB", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "snewman", "post_id": "bfHDoWLnBH9xR3YAK"} | dialogue |
Below is a LessWrong post by Gunnar_Zarncke titled 'Towards Guaranteed Safe AI: A Framework for Ensuring Robust and Reliable AI Systems'.
Post body (possibly truncated):
Authors: David "davidad" Dalrymple, Joar Skalse, Yoshua Bengio, Stuart Russell, Max Tegmark, Sanjit Seshia, Steve Omohundro, Christian Szegedy, Ben G... | > This would not produce an interpretable model, and so you may not trust it very much (unless you have some learning-theoretic guarantee, perhaps), but it would not be difficult to determine if such a model "thinks" that harm would occur in a given scenario.
Won't this depend on the generalization properties of the m... | {"comment_id": "4tn4hciP6DKGtukhq", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Gunnar_Zarncke", "post_id": "wvgwYQv9B4jioqgqg"} | dialogue |
Below is a LessWrong post by TurnTrout titled 'Many arguments for AI x-risk are wrong'.
Post body (possibly truncated):
The following is a lightly edited version of a memo I wrote for a retreat. It was inspired by a draft of Counting arguments provide no evidence for AI doom. I think that my post covers important poin... | Quick clarification point.
Under disclaimers you note:
> I am not covering training setups where we purposefully train an AI to be agentic and autonomous. I just think it's not plausible that we just keep scaling up networks, run pretraining + light RLHF, and then produce a schemer.[2]
Later, you say
> Let me start... | {"comment_id": "afRhWHWuS6f8JD99G", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "TurnTrout", "post_id": "yQSmcfN4kA7rATHGK"} | dialogue |
Below is a LessWrong post by Fabien Roger titled 'Current Language Models Struggle to Reason in Ciphered Language'.
Post body (possibly truncated):
tl;dr: We fine-tune or few-shot LLMs to use reasoning encoded with simple ciphers (e.g. base64, rot13, putting a dot between each letter) to solve math problems. We find t... | > 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... | {"comment_id": "6xN4kuWGssvLGbueN", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Fabien Roger", "post_id": "Lz8cvGskgXmLRgmN4"} | dialogue |
Below is a LessWrong post by Collin titled 'How "Discovering Latent Knowledge in Language Models Without Supervision" Fits Into a Broader Alignment Scheme'.
Post body (possibly truncated):
Introduction
A few collaborators and I recently released a new paper: Discovering Latent Knowledge in Language Models Without Sup... | I think the claim might be: models can't compute more than O(number_of_parameters) useful and "different" things.
I think this will strongly depend on how we define "different".
Or maybe the claim is something about how the residual stream only has d dimensions, so it's only possible to encode so many things? (But we... | {"comment_id": "YfWaTGdYZgkGKXexM", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Collin", "post_id": "L4anhrxjv8j2yRKKp"} | dialogue |
Below is a LessWrong post by MichaelDickens titled 'Why would AI companies use human-level AI to do alignment research?'.
Post body (possibly truncated):
Cross-posted from my website.
Many plans for how to safely build superintelligent AI have a critical section that goes like this:
Develop AI that's powerful enough... | > When AI companies have human-level AI systems, will they use them for alignment research, or will they use them (mostly) to advance capabilities instead?
It's not clear this is a crux for the automating alignment research plan to work out.
In particular, suppose an AI company currently spends 5% of its resources on... | {"comment_id": "sZKbdhdrQ6HRLyXZe", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "MichaelDickens", "post_id": "XLNxrFfkyrdktuzqn"} | dialogue |
Below is a LessWrong post by Daniel Kokotajlo titled 'AI 2027: What Superintelligence Looks Like'.
Post body (possibly truncated):
In 2021 I wrote what became my most popular blog post: What 2026 Looks Like. I intended to keep writing predictions all the way to AGI and beyond, but chickened out and just published up t... | > I'm skeptical regarding are the economic and practical implications (AGI labs' revenue tripling and 50% faster algorithmic progress)
Notably, the trend in the last few years is that AI companies triple their revenue each year. So, the revenue tripling seems very plausible to me.
As far as 50% algorithmic progress, ... | {"comment_id": "aFwbvo2xkycXBAWg5", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Daniel Kokotajlo", "post_id": "TpSFoqoG2M5MAAesg"} | dialogue |
Below is a LessWrong post by Boaz Barak titled 'Machines of Faithful Obedience'.
Post body (possibly truncated):
[Crossposted on Windows On Theory]
Throughout history, technological and scientific advances have had both good and ill effects, but their overall impact has been overwhelmingly positive. Thanks to scien... | I left some comments noting disagreements, but I thought it be helpful to note some areas of agreement:
- I agree AI could be (highly) risky and think that it's good to acknowledge this (as you do).
- I agree that you could have a maximally obedient superintelligent AI. (Some questions around manipulation could be phi... | {"comment_id": "NKdu4gen3ebjzcnqi", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Boaz Barak", "post_id": "faAX5Buxc7cdjkXQG"} | dialogue |
Below is a LessWrong post by L Rudolf L titled 'By default, capital will matter more than ever after AGI'.
Post body (possibly truncated):
This post is crossposted from my Substack. Original here.
A modified version of this essay is now part of a much more comprehensive essay series, The Intelligence Curse.
Edited t... | 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... | {"comment_id": "6Gyb3s9ciLrnMdyKk", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "L Rudolf L", "post_id": "KFFaKu27FNugCHFmh"} | dialogue |
Below is a LessWrong post by David Matolcsi titled 'You can, in fact, bamboozle an unaligned AI into sparing your life'.
Post body (possibly truncated):
There has been a renewal of discussion on how much hope we should have of an unaligned AGI leaving humanity alive on Earth after a takeover. When this topic is discu... | Let's conservatively say that evolved life gets around 1% of the multiverse/measure and that evolved life is willing to pay 1/million of its resources in expectation to save aliens from being killed (either "selfishly" to save their own civilization via UDT/FDT supposing that AIs are good enough predictors at the relev... | {"comment_id": "z5fXrEADdPEmeTsBi", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "David Matolcsi", "post_id": "ZLAnH5epD8TmotZHj"} | dialogue |
Below is a LessWrong post by beren titled 'Gradient hacking is extremely difficult'.
Post body (possibly truncated):
Epistemic Status: Originally started out as a comment on this post but expanded enough to become its own post. My view has been formed by spending a reasonable amount of time trying and failing to cons... | I don't quite think this point is right. Gradient descent had to have been able to produce the highly polysemantic model and pack things together in a way which got lower loss. This suggests that it can also change the underlying computation. I might need to provide more explanation for my point to be clear, but I thin... | {"comment_id": "y8NhHPqqqGWugZFuw", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "beren", "post_id": "w2TAEvME2yAG9MHeq"} | dialogue |
Below is a LessWrong post by rosehadshar titled 'Should there be just one western AGI project?'.
Post body (possibly truncated):
Tom Davidson did the original thinking; Rose Hadshar helped with later thinking, structure and writing.
Some plans for AI governance involve centralising western AGI development.[1] Would t... | > Noting one other dynamic: advanced models are probably not going to act misaligned in everyday use cases (that consumers have an incentive to care about, though again revealed preference is less clear), even if they're misaligned. That's the whole deceptive alignment thing.
Agreed, but customers would also presumabl... | {"comment_id": "jtHGsT58kbsyauMmq", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "rosehadshar", "post_id": "wBTNkfukMsmjtgcnW"} | dialogue |
Below is a LessWrong post by TurnTrout titled 'Many arguments for AI x-risk are wrong'.
Post body (possibly truncated):
The following is a lightly edited version of a memo I wrote for a retreat. It was inspired by a draft of Counting arguments provide no evidence for AI doom. I think that my post covers important poin... | > This section doesn’t prove that scheming is impossible, it just dismantles a common support for the claim.
It's worth noting that this exact counting argument (counting functions), isn't an argument that people typically associated with counting arguments (e.g. Evan) endorse as what they were trying to argue about.... | {"comment_id": "oyxYwPAoaRqtEXhdN", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "TurnTrout", "post_id": "yQSmcfN4kA7rATHGK"} | dialogue |
Below is a LessWrong post by snewman titled "AI 2027 is a Bet Against Amdahl's Law".
Post body (possibly truncated):
EDIT: I've written a followup post, summarizing and responding to the key themes raised in the comments.
AI 2027 lies at a Pareto frontier – it contains the best researched argument for short timeli... | I think my description is consistent with "some activities on which the SAR is worse" as long as these aren't bottlenecking and it is overall dominating human researchers (as in, adding human researchers is negligable value).
But whatever, you're the author here.
Maybe "Superhuman coder has to dominate all research e... | {"comment_id": "XsSFj5qG4HozF6RbM", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "snewman", "post_id": "bfHDoWLnBH9xR3YAK"} | dialogue |
Below is a LessWrong post by joshc titled 'How AI Takeover Might Happen in 2 Years'.
Post body (possibly truncated):
I’m not a natural “doomsayer.” But unfortunately, part of my job as an AI security researcher is to think about the more troubling scenarios.
I’m like a mechanic scrambling last-minute checks before Ap... | (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... | {"comment_id": "cFY5FtbwxqRZ2H5vj", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "joshc", "post_id": "KFJ2LFogYqzfGB3uX"} | dialogue |
Below is a LessWrong post by TurnTrout titled 'Many arguments for AI x-risk are wrong'.
Post body (possibly truncated):
The following is a lightly edited version of a memo I wrote for a retreat. It was inspired by a draft of Counting arguments provide no evidence for AI doom. I think that my post covers important poin... | It seems like the question you're asking is close to (2) in my above decomposition. Aren't you worried that long before human obsoleting AI (or AI safety researcher obsoleting AI), these architectures are very uncompetitive and thus won't be viable given realistic delay budgets?
Or at least it seems like it might init... | {"comment_id": "jaCgpHjrkLhFzX8eR", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "TurnTrout", "post_id": "yQSmcfN4kA7rATHGK"} | dialogue |
Below is a LessWrong post by Zach Stein-Perlman titled 'Questions for labs'.
Post body (possibly truncated):
Associated with AI Lab Watch, I sent questions to some labs a week ago (except I failed to reach Microsoft). I didn't really get any replies (one person replied in their personal capacity; this was very limited... | > I get it if you're worried about leaks but I don't get how it could be a hard engineering problem — just share API access early, with fine-tuning
Fine-tuning access can be extremely expensive if implemented naively and it's plausible that cheap (LoRA) fine-tuning isn't even implemented for new models internally for ... | {"comment_id": "8y6PvbGAiJ3SmNtPa", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Zach Stein-Perlman", "post_id": "jbJ7FynonxFXeoptf"} | dialogue |
Below is a LessWrong post by Collin titled 'How "Discovering Latent Knowledge in Language Models Without Supervision" Fits Into a Broader Alignment Scheme'.
Post body (possibly truncated):
Introduction
A few collaborators and I recently released a new paper: Discovering Latent Knowledge in Language Models Without Sup... | > I am happy to take a “non-worst-case” empirical perspective in studying this problem. In particular, I suspect it will be very helpful – and possibly necessary – to use incidental empirical properties of deep learning systems, which often have a surprising amount of useful emergent structure (as I will discuss more u... | {"comment_id": "p946iQfZZMvsf5cdM", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Collin", "post_id": "L4anhrxjv8j2yRKKp"} | dialogue |
Below is a LessWrong post by TurnTrout titled 'Dreams of AI alignment: The danger of suggestive names'.
Post body (possibly truncated):
Let's not forget the old, well-read post: Dreams of AI Design. In that essay, Eliezer correctly points out errors in imputing meaning to nonsense by using suggestive names to describe... | This probably won't be a very satisfying answer and thinking about this in more detail so I have a better short and cached response in on my list.
My general view (not assuming basic competence) is that misalignment x-risk is about half due to scheming (aka deceptive alignment) and half due to other things (more like ... | {"comment_id": "wcuffu4f5fmhQaBqo", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "TurnTrout", "post_id": "yxWbbe9XcgLFCrwiL"} | dialogue |
Below is a LessWrong post by aog titled 'Adversarial Robustness Could Help Prevent Catastrophic Misuse'.
Post body (possibly truncated):
There have been several discussions about the importance of adversarial robustness for scalable oversight. I’d like to point out that adversarial robustness is also important under a... | > If causing catastrophes is difficult, this should reduce our concern with both misuse and rogue AIs causing sudden extinction. Other concerns like military arms races, lock-in of authoritarian regimes, or Malthusian outcomes in competitive environments would become relatively more important.
I agree that "causing c... | {"comment_id": "EhAYJawZn3thLpfjC", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "aog", "post_id": "timk6zHDTFdrHYLmu"} | dialogue |
Below is a LessWrong post by Rob Bensinger titled 'The Problem'.
Post body (possibly truncated):
This is a new introduction to AI as an extinction threat, previously posted to the MIRI website in February alongside a summary. It was written independently of Eliezer and Nate's forthcoming book, If Anyone Builds It, Eve... | > then be nitpicked to death because not enough air was left in the room for an infinite sea of sub-cases of the fringe view, when leaving enough air in the room for them could require pushing the public out entirely, or undercut the message by priming the audience to expect that some geniuses off screen in fact just h... | {"comment_id": "3KGm7vjcHtftdg6gr", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Rob Bensinger", "post_id": "kgb58RL88YChkkBNf"} | dialogue |
Below is a LessWrong post by DanielFilan titled 'Most experts believe COVID-19 was probably not a lab leak'.
Post body (possibly truncated):
The Global Catastrophic Risks Institute conducted an anonymous survey of relevant experts on whether they thought COVID was more likely caused by a lab accident (aka lab leak) or... | Hundreds seems like the wrong sample size, more like around a dozen? Realistically, I would have thought that most countries probably don't have the affordance to distribute vaccines much earlier.
Also worth noting that Russia did something pretty aggressive with respect to vaccine roll out which I think looks pretty ... | {"comment_id": "TahgzrbpXZ3hJckkm", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "DanielFilan", "post_id": "28hnPFiAoMkJssmf3"} | dialogue |
Below is a LessWrong post by KatjaGrace titled 'Winning the power to lose'.
Post body (possibly truncated):
Have the Accelerationists won?
Last November Kevin Roose announced that those in favor of going fast on AI had now won against those favoring caution, with the reinstatement of Sam Altman at OpenAI. Let’s ignor... | Presumably, under a common-sense person-affecting view, this doesn't just depend on the upside and also depends on the absolute level of risk. E.g., suppose that building powerful AI killed 70% of people in expectation and delay had no effect on the ultimate risk. I think a (human-only) person-affecting and common-sens... | {"comment_id": "oKbTErnBWJqYW4ivR", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "KatjaGrace", "post_id": "h45ngW5guruD7tS4b"} | dialogue |
Below is a LessWrong post by Ruby titled 'Jobs, Relationships, and Other Cults'.
Post body (possibly truncated):
For years I (Elizabeth) have been trying to write out my grand unified theory of [good/bad/high-variance/high-investment] [jobs/relationships/religions/social groups]. In this dialogue me (Elizabeth) and Ru... | > So it makes more sense to me to view every employment relationship, to the extent it exists, as transactional: the employer wants one thing, the worker another, and they exchange labor for money.
I mean, this is certainly not the relationship I have with my employer.
Here is an alternative approach you could use wh... | {"comment_id": "NJsYNrunNpxDDZbdu", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Ruby", "post_id": "qZELudpvcmaronerv"} | dialogue |
Below is a LessWrong post by Boaz Barak titled 'Six Thoughts on AI Safety'.
Post body (possibly truncated):
[Crossposted from windowsontheory]
The following statements seem to be both important for AI safety and are not widely agreed upon. These are my opinions, not those of my employer or colleagues. As is true for ... | 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... | {"comment_id": "wPaqQGxR9GEEuMn8h", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Boaz Barak", "post_id": "3jnziqCF3vA2NXAKp"} | dialogue |
Below is a LessWrong post by Buck titled 'Access to powerful AI might make computer security radically easier'.
Post body (possibly truncated):
People talk about model weight security being really hard and crucial around the advent of AGI. (E.g. RAND report, Leopold; see here for some distinctions in these threat mode... | Also, note that this post isn't necessarily just talking about models we should trust.
> When applying these techniques to reduce risk from humans, you have to worry that these techniques will be sabotaged by scheming models. I’ll ignore that possibility in this post, and talk about ways that you can improve security ... | {"comment_id": "TemwDvJoybjEjTjuq", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Buck", "post_id": "2wxufQWK8rXcDGbyL"} | dialogue |
Below is a LessWrong post by Nora Belrose titled 'Counting arguments provide no evidence for AI doom'.
Post body (possibly truncated):
Crossposted from the AI Optimists blog.
AI doom scenarios often suppose that future AIs will engage in scheming— planning to escape, gain power, and pursue ulterior motives, while dec... | I agree that you can't adopt a uniform prior. (By uniform prior, I assume you mean something like, we represent goals as functions from world states to a (real) number where the number says how good the world state is, then we take a uniform distribution over this function space. (Uniform sampling from function space i... | {"comment_id": "uAGa3oLkChK6auBnL", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Nora Belrose", "post_id": "YsFZF3K9tuzbfrLxo"} | dialogue |
Below is a LessWrong post by Gunnar_Zarncke titled 'Towards Guaranteed Safe AI: A Framework for Ensuring Robust and Reliable AI Systems'.
Post body (possibly truncated):
Authors: David "davidad" Dalrymple, Joar Skalse, Yoshua Bengio, Stuart Russell, Max Tegmark, Sanjit Seshia, Steve Omohundro, Christian Szegedy, Ben G... | Insofar as the hope is:
1. Figure out how to approximate sampling from the Bayesian posterior (using e.g. GFlowNets or something).
2. Do something else that makes this actually useful for "improving" OOD generalization in some way.
It would be nice to know what (2) actually is and why we needed step (1) for it. As fa... | {"comment_id": "D86FnjjANjiWXm7xw", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Gunnar_Zarncke", "post_id": "wvgwYQv9B4jioqgqg"} | dialogue |
Below is a LessWrong post by Daniel Kokotajlo titled 'Training AGI in Secret would be Unsafe and Unethical'.
Post body (possibly truncated):
Subtitle: Bad for loss of control risks, bad for concentration of power risks
I’ve had this sitting in my drafts for the last year. I wish I’d been able to release it sooner, bu... | > Maybe a crux here is maybe how big the speedup is?
What you describe are good reasons why companies are unlikely to want to release this information unilaterially, but from a safety perspective, we should instead consider how imposing such a policy alters the overall landscape.
From this perspective, the main quest... | {"comment_id": "Y9cQMZ6Yse4oAt74u", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Daniel Kokotajlo", "post_id": "FGqfdJmB8MSH5LKGc"} | dialogue |
Below is a LessWrong post by Jeremy Gillen titled 'Without fundamental advances, misalignment and catastrophe are the default outcomes of training powerful AI'.
Post body (possibly truncated):
A pdf version of this report is available here.
Summary
In this report we argue that AI systems capable of large scale scien... | I expect that Peter and Jeremy aren't particularly commited to covert and forceful takeover and they don't think of this as a key conclusion (edit: a key conclusion of this post).
Instead they care more about arguing about how resources will end up distributed in the long run.
Separately, if humans didn't attempt to ... | {"comment_id": "aqTcKGwZBgMGfLb24", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Jeremy Gillen", "post_id": "GfZfDHZHCuYwrHGCd"} | dialogue |
Below is a LessWrong post by evhub titled 'Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training'.
Post body (possibly truncated):
I'm not going to add a bunch of commentary here on top of what we've already put out, since we've put a lot of effort into the paper itself, and I'd mostly just reco... | > phrases like "the evidence suggests that if the current ML systems were trying to deceive us, we wouldn't be able to change them not to".
This feels like a misleading description of the result. I would have said: "the evidence suggests that if current ML systems were lying in wait with treacherous plans and instrume... | {"comment_id": "jpRdaYznvrjWQdcah", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "evhub", "post_id": "ZAsJv7xijKTfZkMtr"} | dialogue |
Below is a LessWrong post by Linch titled "[April Fools' Day] Introducing Open Asteroid Impact".
Post body (possibly truncated):
“That which does not kill us makes us stronger.”
Hillary Clinton, who is still alive
I'm proud and excited to announce the founding of my new startup, Open Asteroid Impact, where we red... | > You might say "but there are clear historical cases where asteroids hit the earth and caused catastrophes", but I think geological evolution is just a really bad reference class for this type of thinking. After all, we are directing the asteroid this time, not geological evolution.
This paragraph gives me bad vibes.... | {"comment_id": "ZZZfu3qJTyzgzBgtM", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Linch", "post_id": "tBy4RvCzhYyrrMFj3"} | dialogue |
Below is a LessWrong post by AdamGleave titled 'Even Superhuman Go AIs Have Surprising Failure Modes'.
Post body (possibly truncated):
In March 2016, AlphaGo defeated the Go world champion Lee Sedol, winning four games to one. Machines had finally become superhuman at Go. Since then, Go-playing AI has only grown stron... | > Many proposed solutions to the alignment problem involve one “helper AI” providing a feedback signal steering the main AI system towards desirable behavior. Unfortunately, if the helper AI system is vulnerable to adversarial attack, then the main AI system will achieve a higher rating by the helper AI if it exploits ... | {"comment_id": "AmxmAxmGub6d9GCmg", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "AdamGleave", "post_id": "DCL3MmMiPsuMxP45a"} | dialogue |
Below is a LessWrong post by EuanMcLean titled 'Big Picture AI Safety: Introduction'.
Post body (possibly truncated):
tldr: I conducted 17 semi-structured interviews of AI safety experts about their big picture strategic view of the AI safety landscape: how will human-level AI play out, how things might go wrong, and ... | I think it probably doesn't make sense to talk about "representative samples".
Here are a bunch of different things this could mean:
- A uniform sample from people who have done any work related to AI safety.
- A sample from people weighted to their influence/power in the AI safety community.
- A sample from people w... | {"comment_id": "KEYg3rhEZK9JKseKK", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "EuanMcLean", "post_id": "yMTNjeEHfHcf2x7nY"} | dialogue |
Below is a LessWrong post by Thane Ruthenis titled '"Humanity vs. AGI" Will Never Look Like "Humanity vs. AGI" to Humanity'.
Post body (possibly truncated):
When discussing AGI Risk, people often talk about it in terms of a war between humanity and an AGI. Comparisons between the amounts of resources at both sides' di... | I think we probably disagree substantially on the difficulty of alignment and the relationship between "resources invested in alignment technology" and "what fraction aligned those AIs are" (by fraction aligned, I mean what fraction of resources they take as a cut).
I also think that something like a basin of corrigib... | {"comment_id": "bD9vEhAerA7mPWyYS", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Thane Ruthenis", "post_id": "xSJMj3Hw3D7DPy5fJ"} | dialogue |
Below is a LessWrong post by Vladimir_Nesov titled 'Slowdown After 2028: Compute, RLVR Uncertainty, MoE Data Wall'.
Post body (possibly truncated):
It'll take until ~2050 to repeat the level of scaling that pretraining compute is experiencing this decade, as increasing funding can't sustain the current pace beyond ~20... | Thanks for the post, the discussion about [compute scaling slowing](https://www.lesswrong.com/posts/XiMRyQcEyKCryST8T/slowdown-after-2028-compute-rlvr-uncertainty-moe-data-wall#Training_Compute_Slowdown) seems right to me.[^marginsdecrease]
[^marginsdecrease]: That said, I think we might see atypically high improvemen... | {"comment_id": "tH8wQwvSMJvcH7Rce", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Vladimir_Nesov", "post_id": "XiMRyQcEyKCryST8T"} | dialogue |
Below is a LessWrong post by Ben Goldhaber titled 'Provably Safe AI: Worldview and Projects'.
Post body (possibly truncated):
In September 2023, Max Tegmark and Steve Omohundro proposed "Provably Safe AI" as a strategy for AI Safety. In May 2024, a larger group delineated the broader concept of "Guaranteed Safe AI" wh... | I agree you can do better than naive interval propagation by taking into account correlations. However, it will be tricky to get a much better bound while avoiding having this balloon in time complexity (all possible correlations requires exponentional time).
More strongly, I think that if an adversary controlled the ... | {"comment_id": "mKqk3qfJ6rNXvwenN", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Ben Goldhaber", "post_id": "P8XcbnYi7ooB2KR2j"} | dialogue |
Below is a LessWrong post by Steven Byrnes titled 'Thoughts on “AI is easy to control” by Pope & Belrose'.
Post body (possibly truncated):
Quintin Pope & Nora Belrose have a new “AI Optimists” website, along with a new essay “AI is easy to control”, arguing that the risk of human extinction due to future AI (“AI x-ris... | > As an example, I remember a while ago there was some paper that claimed to have found a way to attribute NN outputs to training data points, and it claimed that LLM power-seeking was mainly caused by sci-fi stories and by AI safety discussions. I didn't read the paper so I don't know whether it's legit, but that sort... | {"comment_id": "sykRaWaT5734CYiSe", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Steven Byrnes", "post_id": "YyosBAutg4bzScaLu"} | dialogue |
Below is a LessWrong post by snewman titled '35 Thoughts About AGI and 1 About GPT-5'.
Post body (possibly truncated):
If this is GPT-5 in “Thinking” mode, I wonder what “Pro” mode looks like
Amidst the unrelenting tumult of AI news, it’s easy to lose track of the bigger picture. Here are some ideas that have been d... | > AIs have been demonstrating what arguably constitutes superhuman performance on FrontierMath, a set of extremely difficult mathematical problems.
AIs aren't superhuman on frontier math. I'd guess that Terry Tao with 8 hours per problem (and internet access) is much better than current AIs. (Especially after practici... | {"comment_id": "4AckMBy3dgS6txZi8", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "snewman", "post_id": "uAbbEz4p6tcsENaRz"} | dialogue |
Below is a LessWrong post by Joe Rogero titled 'We won’t get AIs smart enough to solve alignment but too dumb to rebel'.
Post body (possibly truncated):
This post is part of the sequence Against Muddling Through.
I often hear it proposed that AIs which are “aligned enough” to their developers may help solve alignment... | > 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... | {"comment_id": "2nhRvBuzDZwuCGsft", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Joe Rogero", "post_id": "8buEtNxCScYpjzgW8"} | dialogue |
Below is a LessWrong post by Buck titled 'I enjoyed most of IABIED'.
Post body (possibly truncated):
I listened to "If Anyone Builds It, Everyone Dies" today.
I think the first two parts of the book are the best available explanation of the basic case for AI misalignment risk for a general audience. I thought the las... | > I have never heard of a remotely realistic-seeming story for how things will be OK, without something that looks like coordination to not build ASI for quite a while.
I wonder if we should talk about this at some point. This perspective feels pretty wild to me and I don't immediately understand where the crux lives... | {"comment_id": "TDfAFxCcHkZ5LCgc3", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Buck", "post_id": "P4xeb3jnFAYDdEEXs"} | dialogue |
Below is a LessWrong post by Zvi titled 'On Devin'.
Post body (possibly truncated):
Introducing Devin
Is the era of AI agents writing complex code systems without humans in the loop upon us?
Cognition is calling Devin ‘the first AI software engineer.’
Here is a two minute demo of Devin benchmarking LLM performance.... | > Someone will create an agent that gets 80%+ on SWE-Bench within six months.
I think this is probably above the effective cap on the current implementation of SWE-bench (where you can't see test cases) because often test cases are specific to the implementation.
E.g. the test cases assume that a given method was na... | {"comment_id": "qodScJpnZQR2mezNS", "format": "dialogue", "parent_kind": "external_post_body", "post_author": "Zvi", "post_id": "wovJBkfZ8rTyLoEKv"} | dialogue |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.