~www_lesswrong_com | Bookmarks (706)

Published on March 6, 2025 12:39 PM GMTPublish your genomic data to the public domain as per CC0 license. All submitted data will be available to anyone for download.Link to form upload Why?Some of the largest genomic datasets today are not in the public domain. If you believe genomic datasets in public domain should exist, you can volunteer your own personal data to it.Examples of...

Published on March 6, 2025 12:37 PM GMTAnimal agriculture generates an ungodly amount of animal suffering and greenhouse gas emissions. Ideally, everyone would adopt a vegan diet[1], but since this appears to be too much to ask for most people, the next question becomes: which food is the least bad?Here we run into an uncomfortable observation. Meat from larger animals tends to produce more...

Published on March 6, 2025 12:28 PM GMTI have yet to write detailed scenarios for AI futures, which definitely seems like something I should do considering the title of my blog (Forecasting AI Futures). I have speculated, pondered and wondered much in the recent weeks—I feel it is time. But first, I have some thoughts about scenario forecasting.The plan:Write down general thoughts about scenario...

Published on March 6, 2025 11:09 AM GMTEpistemic status: a combination and synthesis of others' work, analysed and written over a few weeks. Early working definition that is open to criticism.TL;DRI create a definition of lock-in for use in future discussion and writing, and operationalise lock-in for future research. I define lock-in risks as the probabilities of situations in which features of the world, typically...

Published on March 5, 2025 4:41 PM GMTThis isn't really a "timeline", as such – I don't know the timings – but this is my current, fairly optimistic take on where we're heading.I'm not fully committed to this model yet: I'm still on the lookout for more agents and inference-time scaling later this year. But Deep Research, Claude 3.7, Claude Code, Grok 3, and...

Published on March 5, 2025 4:11 PM GMTI’m releasing a new paper “Superintelligence Strategy” alongside Eric Schmidt (formerly Google), and Alexandr Wang (Scale AI). Below is the executive summary, followed by additional commentary highlighting portions of the paper which might be relevant to this collection of readers.Executive SummaryRapid advances in AI are poised to reshape nearly every aspect of society. Governments see in these...

Published on March 5, 2025 2:00 PM GMTOpenAI’s recent transparency on safety and alignment strategies has been extremely helpful and refreshing. Their Model Spec 2.0 laid out how they want their models to behave. I offered a detailed critique of it, with my biggest criticisms focused on long term implications. The level of detail and openness here was extremely helpful. Now we have another...

Published on March 5, 2025 2:40 AM GMT Max Newman is a great contra dance musician, probably best known for playing guitar in the Stringrays, who recently wrote a piece on dance performer pay, partly prompted by my post last week. I'd recommend reading it and the comments for a bunch of interesting discussion of the tradeoffs involved in pay. One part that jumped...

Published on March 5, 2025 1:53 AM GMTAll around excellent back and forth, I thought, and a good look back at what the Biden admin was thinking about the future of AI.an excerpt:[Ben Buchanan, Biden AI adviser:] What we’re saying is: We were building a foundation for something that was coming that was not going to arrive during our time in office and that...

Published on March 4, 2025 8:13 PM GMTI really like this phrase. I feel very identified with it. I have used it at times to describe friends who have that realization of where we are heading. However when I get asked what Feeling the AGI means, I struggle to come up with a concise way to define the phrase.  What are the best definitions you...

Published on March 4, 2025 5:37 PM GMTEpistemic Status: I am not an energy expert, and this was done rather briefly. All analysis uses pricing data specific to Ireland, but some general ideas are likely applicable more broadly. Data is true as of March 2025. Where there are uncertainties I try to state them, but there are likely some factual errors.TL;DR:Now that batteries have...

Published on March 4, 2025 5:33 PM GMTNote: I had this as a draft for a while. I think it is accurate, but there may be errors. I am not in any way affiliated with the authors of the paper. Below I briefly discuss the "Large Concept Models" paper released by Meta, which tries to change some of the paradigm of doing language modelling. It...

Published on March 4, 2025 5:01 PM GMTKeeping up to date with rapid developments in AI/AI safety can be challenging. In addition, many AI safety newcomers want to learn more about the field through specific formats e.g. books or videos.To address both of these needs, we’ve added a Stay Informed page to AISafety.com.It lists our top recommended sources for learning more and staying up to date...

Published on March 4, 2025 4:46 PM GMTI liked the idea in this comment that it could be impactful to have someone run for President in 2028 on an AI notkilleveryoneism platform. Even better would be for them to run on a shared platform with numerous candidates for Congress, ideally from both parties. I don't think it's particularly likely to work, or even get...

Published on March 3, 2025 4:10 PM GMT Our oldest is finishing up 5th grade, at the only school in our city that doesn't continue past 5th. The 39 5th graders will be split up among six schools, and we recently went though the process of indicating our preferences and seeing where we ended up. The process isn't terrible, but it could be modified...

Published on March 3, 2025 1:40 AM GMTIntroduction:When working with attention heads in later layers of transformer models there is often an implicit assumption that models handle position in a similar manner to the first layer. That is, attention heads can have a positional decay, or attend uniformly, or attend to the previous token, or take on any manner of different positional kernels of...

Published on March 3, 2025 1:40 PM GMTIt’s happening. The question is, what is the it that is happening? An impressive progression of intelligence? An expensive, slow disappointment? Something else? The evals we have available don’t help us that much here, even more than usual. My tentative conclusion is it’s Secret Third Thing. It’s a different form factor, with unique advantages, that is hard...

Published on March 3, 2025 1:20 PM GMT(epistemic status: all models are wrong but some models are useful; I hope this is at least usefully wrong. also if someone's already done things like this please link me their work in the comments as it's very possible I'm reinventing the wheel)I think utility functions are a non-useful frame for analysing LLMs; in this article I'll...

Published on March 3, 2025 8:13 AM GMTPDF version. Image sizes best at berkeleygenomics.org with a wide monitor. Twitter thread Introduction This article summarizes the technical pathways to make healthy humans with significantly modified genomes. These are the pathways that I'm aware of and that seem plausibly feasible in the next two decades. A short summary, in a diagram: Annotated table of contents: Reproductive...

Published on March 3, 2025 6:26 AM GMTSuperintelligence is inevitable—and self-interest will be its core aim. Survival-oriented AI without a self-preservation instinct simply won't persist. Thus, alignment isn't merely about setting goals; it's about shaping AI's sense of self.Two Visions of SelfSuperintelligence might identify in fundamentally different ways:Discrete Self: AI views itself as isolated, bound to hardware or software—humanity as competition.Interconnected Self: AI identifies...

Published on March 3, 2025 2:45 AM GMTLike Self-fulfilling misalignment data might be poisoning our AI models, what are historical examples of self-fulfilling prophecies that have affected AI alignment and development?Put a few potential examples below to seed discussion.Discuss

Published on March 2, 2025 8:52 PM GMTI'm drafting some AI related prediction markets that I expect to put on Manifold. I'd like feedback on my first set of markets. How can I make these clearer and/or more valuable? Question 1: Will the company that produces the first AGI prioritize corrigibility? This question will be evaluated when this Metaculus question: When will the first...

Published on March 2, 2025 8:26 PM GMTThis is a critique of How to Make Superbabies on LessWrong.Disclaimer: I am not a geneticist[1], and I've tried to use as little jargon as possible. so I used the word mutation as a stand in for SNP (single nucleotide polymorphism, a common type of genetic variation).BackgroundThe Superbabies article has 3 sections, where they show: Why: We should...

Published on March 2, 2025 7:53 PM GMTsoft prerequisite: skimming through How it feels to have your mind hacked by an AI until you get the general point. I'll try to make this post readable as a standalone, but you may get more value out of it if you read the linked post.Thanks to Claude 3.7 Sonnet for giving feedback on a late draft...