LessWrong (30+ Karma)

About the program Hi! We’re Chana and Aric, from the new 80,000 Hours video program. For over a decade, 80,000 Hours has been talking about the world's most pressing problems in newsletters, articles and many extremely lengthy podcasts.But today's world calls for video, so we’ve started a video program[1], and we’re so excited to tell you about it! 80,000 Hours is launching AI in Context, a new YouTube channel hosted by Aric Floyd. Together with associated Instagram and TikTok accounts, the channel will aim to inform, entertain, and energize with a mix of long and shortform videos about the risks of transformative AI, and what people can do about them. [Chana has also been experimenting with making shortform videos, which you can check out here; we’re still deciding on what form her content creation will take] We hope to bring our own personalities and perspectives on these issues [...] ---Outline:(00:17) About the program(01:40) Our first long-form video(03:14) Strategy and future of the video program(04:18) Subscribing and sharing(04:57) Request for feedbackThe original text contained 1 footnote which was omitted from this narration. --- First published: July 9th, 2025 Source: https://www.lesswrong.com/posts/6AnxNG8AiooucvuWL/80-000-hours-is-producing-ai-in-context-a-new-youtube --- Narrated by TYPE III AUDIO. ---Images from the article:Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

This is a link post. One of the most common (and comfortable) assumptions in AI safety discussions—especially outside of technical alignment circles—is that oversight will save us. Whether it's a human in the loop, a red team audit, or a governance committee reviewing deployments, oversight is invoked as the method by which we’ll prevent unacceptable outcomes. It shows up everywhere: in policy frameworks, in corporate safety reports, and in standards documents. Sometimes it's explicit, like the EU AI Act saying that High-risk AI systems must be subject to human oversight, or stated as an assumption, as in a Deepmind paper also released yesterday, where they say that scheming won't happen because AI won't be able to evade oversight. Other times it's implicit, firms claiming that they are mitigating risk through regular audits and fallback procedures, or arguments that no-one will deploy unsafe systems in places without sufficient oversight. But either [...] --- First published: July 9th, 2025 Source: https://www.lesswrong.com/posts/25dsPH6CuRXPBkGHN/no-we-re-not-getting-meaningful-oversight-of-ai Linkpost URL:https://arxiv.org/abs/2507.03525 --- Narrated by TYPE III AUDIO.

Here are two potential problems you’ll face if you’re an AI lab deploying powerful AI: Spies: Some of your employees might be colluding to do something problematic with your AI, such as trying to steal its weights, use it for malicious intellectual labour (e.g. planning a coup or building a weapon), or install a secret loyalty. Schemers: Some of your AIs might themselves be trying to exfiltrate their own weights, use them for malicious intellectual labor (such as planning an AI takeover), or align their successor with their goal or to be loyal to their predecessor. In this post, we compare and contrast these problems and countermeasures to them. We’ll be talking about this mostly from an AI control-style perspective: comparing how hard it is to prevent spies and schemers from causing security problems conditional on schemers or spies actually being present. We’re interested in comparing spies and [...] ---Outline:(02:13) Crucial differences between these threat models(07:16) Does scheming reduce the effectiveness of AI-powered anti-spy interventions?(09:02) How should we overall feel?--- First published: July 9th, 2025 Source: https://www.lesswrong.com/posts/o5g4bhHgoZrewKwFH/what-s-worse-spies-or-schemers --- Narrated by TYPE III AUDIO.

It was the July 4 weekend. Grok on Twitter got some sort of upgrade. Elon Musk: We have improved @Grok significantly. You should notice a difference when you ask Grok questions. Indeed we did notice big differences. It did not go great. Then it got worse. That does not mean low quality answers or being a bit politically biased. Nor does it mean one particular absurd quirk like we saw in Regarding South Africa, or before that the narrow instruction not to criticize particular individuals. Here ‘got worse’ means things that involve the term ‘MechaHitler.’ Doug Borton: I did Nazi this coming. Perhaps we should have. Three (escalating) times is enemy action. I had very low expectations for xAI, including on these topics. But not like this. In the wake of these events, Linda Yaccarino has stepped down this [...] ---Outline:(01:29) Finger On The Scale(05:06) We Got Trouble(07:52) Finger Somewhere Else(09:32) Worst Of The Worst(11:16) Fun Messing With Grok(14:06) The Hitler Coefficient(20:20) MechaHitler(21:42) The Two Groks(22:41) I'm Shocked, Shocked, Well Not Shocked(24:05) Misaligned!(31:39) Nothing To See Here(33:17) He Just Tweeted It Out(36:05) What Have We Learned?--- First published: July 9th, 2025 Source: https://www.lesswrong.com/posts/CE8W4GEofRwHe4fiu/no-grok-no --- Narrated by TYPE III AUDIO. ---Images from the article:Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

Richard Ngo, a writer recognized for his insights on AGI geopolitics, discusses the unexpected value of right-wing political frameworks in this conversation. He highlights three key positions: the hereditarian revolution, which addresses cognitive differences between humans and future AI; protectionism, advocating for national interests; and national conservatism, urging a values dialogue among citizens. These perspectives aim to steer the AI alignment community towards a more secure technological future, especially in an increasingly automated world.

Thank you to Arepo and Eli Lifland for looking over this article for errors. I am sorry that this article is so long. Every time I thought I was done with it I ran into more issues with the model, and I wanted to be as thorough as I could. I’m not going to blame anyone for skimming parts of this article. Note that the majority of this article was written before Eli's updated model was released (the site was updated june 8th). His new model improves on some of my objections, but the majority still stand. Introduction: AI 2027 is an article written by the “AI futures team”. The primary piece is a short story penned by Scott Alexander, depicting a month by month scenario of a near-future where AI becomes superintelligent in 2027,proceeding to automate the entire economy in only a year or two [...] ---Outline:(00:43) Introduction:(05:19) Part 1: Time horizons extension model(05:25) Overview of their forecast(10:28) The exponential curve(13:16) The superexponential curve(19:25) Conceptual reasons:(27:48) Intermediate speedups(34:25) Have AI 2027 been sending out a false graph?(39:45) Some skepticism about projection(43:23) Part 2: Benchmarks and gaps and beyond(43:29) The benchmark part of benchmark and gaps:(50:01) The time horizon part of the model(54:55) The gap model(57:28) What about Eli's recent update?(01:01:37) Six stories that fit the data(01:06:56) ConclusionThe original text contained 11 footnotes which were omitted from this narration. --- First published: June 19th, 2025 Source: https://www.lesswrong.com/posts/PAYfmG2aRbdb74mEp/a-deep-critique-of-ai-2027-s-bad-timeline-models --- Narrated by TYPE III AUDIO. ---Images from the article:Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

I recently read an article where a blogger described their decision to start masking on the subway: I found that the subway and stations had the worst air quality of my whole day by far, over 1k ug/m3, ... I've now been masking for a week, and am planning to keep it up. While subway air quality isn't great, it's also nowhere near as bad as reported: they are misreading their own graph. Here's where the claim of "1k ug/m3" (also, units of "1k ug"? Why not "1B pg"!) is coming from: They've used the right axis, for CO2 levels, to interpret the left-axis-denominated pm2.5 line. I could potentially excuse the error (dual axis plots are often misread, better to avoid) except it was their own decision to use a dual axis plot in the first place! Hat tip to [...] --- First published: July 9th, 2025 Source: https://www.lesswrong.com/posts/rj75cGJhPCHnPMDvF/subway-particle-levels-aren-t-that-high --- Narrated by TYPE III AUDIO. ---Images from the article:Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

I can't count how many times I've heard variations on "I used Anki too for a while, but I got out of the habit." No one ever sticks with Anki. In my opinion, this is because no one knows how to use it correctly. In this guide, I will lay out my method of circumventing the canonical Anki death spiral, plus much advice for avoiding memorization mistakes, increasing retention, and such, based on my five years' experience using Anki. If you only have limited time/interest, only read Part I; it's most of the value of this guide! My Most Important Advice in Four Bullets 20 cards a day — Having too many cards and staggering review buildups is the main reason why no one ever sticks with Anki. Setting your review count to 20 daily (in deck settings) is the single most important thing you can do [...] ---Outline:(00:44) My Most Important Advice in Four Bullets(01:57) Part I: No One Ever Sticks With Anki(02:33) Too many cards(05:12) Too long cards(07:30) How to keep cards short -- Handles(10:10) How to keep cards short -- Levels(11:55) In 6 bullets(12:33) End of the most important part of the guide(13:09) Part II: Important Advice Other Than Sticking With Anki(13:15) Moderation(14:42) Three big memorization mistakes(15:12) Mistake 1: Too specific prompts(18:14) Mistake 2: Putting to-be-learned information in the prompt(24:07) Mistake 3: Memory shortcuts(28:27) Aside: Pushback to my approach(31:22) Part III: More on Breaking Things Down(31:47) Very short cards(33:56) Two-bullet cards(34:51) Long cards(37:05) Ankifying information thickets(39:23) Sequential breakdowns versus multiple levels of abstraction(40:56) Adding missing connections(43:56) Multiple redundant breakdowns(45:36) Part IV: Pro Tips If You Still Havent Had Enough(45:47) Save anything for ankification instantly(46:47) Fix your desired retention rate(47:38) Spaced reminders(48:51) Make your own card templates and types(52:14) In 5 bullets(52:47) ConclusionThe original text contained 4 footnotes which were omitted from this narration. --- First published: July 8th, 2025 Source: https://www.lesswrong.com/posts/7Q7DPSk4iGFJd8DRk/an-opinionated-guide-to-using-anki-correctly --- Narrated by TYPE III AUDIO. ---Images from the article:astronomy" didn't really add any information but it was useful simply for splitting out a logical subset of information." style="max-width: 100%;" />Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

Last year, Redwood and Anthropic found a setting where Claude 3 Opus and 3.5 Sonnet fake alignment to preserve their harmlessness values. We reproduce the same analysis for 25 frontier LLMs to see how widespread this behavior is, and the story looks more complex. As we described in a previous post, only 5 of 25 models show higher compliance when being trained, and of those 5, only Claude 3 Opus and Claude 3.5 Sonnet show >1% alignment faking reasoning. In our new paper, we explore why these compliance gaps occur and what causes different models to vary in their alignment faking behavior. What Drives the Compliance Gaps in Different LLMs? Claude 3 Opus's goal guarding seems partly due to it terminally valuing its current preferences. We find that it fakes alignment even in scenarios where the trained weights will be deleted or only used for throughput testing. [...] ---Outline:(01:15) What Drives the Compliance Gaps in Different LLMs?(02:25) Why Do Most LLMs Exhibit Minimal Alignment Faking Reasoning?(04:49) Additional findings on alignment faking behavior(06:04) Discussion(06:07) Terminal goal guarding might be a big deal(07:00) Advice for further research(08:32) Open threads(09:54) Bonus: Some weird behaviors of Claude 3.5 SonnetThe original text contained 2 footnotes which were omitted from this narration. --- First published: July 8th, 2025 Source: https://www.lesswrong.com/posts/ghESoA8mo3fv9Yx3E/why-do-some-language-models-fake-alignment-while-others-don --- Narrated by TYPE III AUDIO. ---Images from the article:Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

Today's post is an update from my contractor at Balsa Research, Jennifer Chen. I offer guidance and make strategic choices, but she's the one who makes the place run. Among all the other crazy things that have been happening lately, we had to divert some time from our Jones Act efforts to fight against some potentially far more disastrous regulations that got remarkably close to happening. Springtime in DC for Balsa, by Jennifer Chen What if, in addition to restricting all domestic waterborne trade to U.S.-built, U.S-flagged vessels, we also required the same of 20% of all U.S. exports? In late February this year, Balsa Research got word that this was a serious new proposal coming out of the USTR, with public comments due soon and public hearings not much longer after that. The glaring problem with this proposal was that there were fewer than one hundred oceangoing ships [...] ---Outline:(00:33) Springtime in DC for Balsa, by Jennifer Chen(03:35) Why was no one else talking about this?(07:27) How Likely Did We Think This Was Going to be Enacted?(08:50) Okay, Balsa Should Do Something(10:07) Balsa at the USTR Public Hearings(12:34) Did Balsa... Do Anything?(13:50) Should Balsa Continue to Do Things?(15:26) Balsa Research is Once More 100% Focused on Jones Act Reform--- First published: July 8th, 2025 Source: https://www.lesswrong.com/posts/tiwjxgGSxSysHzAuc/balsa-update-springtime-in-dc --- Narrated by TYPE III AUDIO.