Scott explains why Midjourney's new ultrasound tomography scanner, despite being technologically interesting, faces major medical limitations and is unlikely to replace existing imaging methods or enable useful whole-body screening without significant future AI improvements. Longer summary

Scott analyzes Midjourney's announcement of a new medical scanner that uses ultrasound tomography to create 3D images of the body. He explains why, despite being technologically interesting, the scanner faces significant medical limitations: ultrasound cannot penetrate bone or air, limiting its usefulness compared to existing MRI and CT scans; it can't replace portable bedside ultrasound due to requiring patients to be submerged in water; and whole-body screening scans (even with existing superior MRI technology) are currently not recommended by medical consensus due to false positives causing more harm than benefit. Scott suggests the technology might become more viable if future AI advances dramatically improve ultrasound interpretation, wave dynamics for imaging through bone, and the ability to distinguish dangerous findings from harmless abnormalities. The post includes an appendix with Twitter responses from radiologists (universally negative) and other commenters (more optimistic), discussing potential applications like body composition measurement and the challenges of medical screening. Shorter summary

A replication of a 2023 neuroscience study on brain entrainment and learning found that the original effect likely doesn't exist, revealing how the study obscured key issues through statistical averaging and highlighting how AI tools are democratizing scientific scrutiny. Longer summary

Sasha Putilin received an ACX grant to replicate a 2023 study claiming that flickering lights synchronized to individual brain rhythms could boost learning speed threefold. The $32,000 replication with 12 participants (versus 80 in the original) found no such effect. Upon examining the original data more closely, Putilin discovered the apparent effect was driven entirely by a few participants with large negative learning rates (getting worse over time) who happened to all be in one experimental group. The original study had obscured this pattern by presenting averaged data rather than individual results. Putilin argues this exemplifies 'cargo-cult statistics' where researchers mechanically apply statistical rituals without critically examining underlying data. The post concludes by suggesting AI tools are democratizing meta-science, enabling anyone to audit published research that previously required expert-level effort. Shorter summary

Scott uses Claude AI to help research California primary races and finds its tailored candidate analyses and recommendations align well with his eventual voting choices, suggesting AI advisors could improve democratic participation. Longer summary

Scott Alexander demonstrates how he used Claude AI to help research and make decisions for local California primary elections. He shares detailed examples of Claude's analysis of candidates and ballot measures, showing how the AI provided comprehensive summaries of candidates' positions, backgrounds, and endorsements tailored to his stated political preferences (centrist liberal, YIMBY, abundance-oriented). He tested Claude's recommendations against his own eventual choices across 10 races, finding strong agreement (5 perfect matches, 3 second-choices). Scott concludes that AI voting advisors could be valuable both for people who don't have time for deep research and for enhancing the research of those who do, and suggests this could be important for democratic decision-making in a post-AGI future. Shorter summary

Scott argues that even if AGI requires a new paradigm beyond LLMs, we shouldn't expect significant delays, since Lindy's Law suggests major paradigm shifts could occur within 3-5 years, and new paradigms typically emerge precisely when scaling hits limits. Longer summary

Scott addresses the objection that AGI is far off because LLMs need a 'new paradigm' to reach AGI. He traces the evolutionary tree of AI development from neural networks through transformers to modern LLMs, then applies Lindy's Law to show that even paradigm shifts as major as deep learning or transformers should be expected within 3-5 years at the 25th percentile. He argues this timeline is comparable to LLM-only predictions anyway. Scott also makes a subtler point: new paradigms historically emerge when old ones hit scaling limits, meaning they won't cause delays but rather continue progress from where scaling left off. He concludes that extrapolating from current LLM scaling remains the best forecasting method whether or not LLMs themselves reach AGI. Shorter summary

Scott examines prediction markets on Anthropic's Pentagon troubles (minimal impact expected), the 2026 midterms (Democratic wins likely despite voting law concerns), groundhog weather predictions (mostly broken clocks), Iran conflict outcomes (under 50% regime change), and introduces MNX, a new AI-focused futures exchange. Longer summary

Scott analyzes several recent prediction market stories. First, he examines how Anthropic's stock price barely changed after the Pentagon declared it a 'supply chain risk', because markets predict the company will win on appeal and the designation only affects a small portion of their business while generating positive publicity. He then discusses the 2026 midterms, where Democrats are favored to win but various Republican voting law changes could create chaos, though markets suggest turnout won't be significantly affected. The post includes a statistical analysis of groundhog weather predictions, showing Staten Island Chuck's high accuracy is likely due to consistently predicting spring. He covers prediction markets about the Iran conflict, including regime change odds and potential casualties. Finally, he announces MNX, a new cryptocurrency-based futures exchange focused on AI-related hedging markets, and shares miscellaneous prediction market news including Substack's partnership with Polymarket. Shorter summary

Scott argues that dismissing AI as 'just a next-token predictor' is like dismissing humans as 'just reproduction machines' - both confuse the optimization process that shaped an entity with how that entity actually thinks. Longer summary

Scott argues that dismissing AI as 'just a next-token predictor' confuses levels of optimization. He draws an analogy to humans: just as humans were shaped by evolution optimizing for reproduction but don't think about sex when doing math, AIs were shaped by next-token prediction but don't simply predict tokens when thinking. Scott explains that human brains use predictive coding (predicting next sense-data) to build world-models, while AIs use next-token prediction to build their own world-models. Both processes create complex internal representations - like helical manifolds in 6D space for AIs, or toroidal attractors in human hippocampi - that operate far above the level of simple prediction. The post concludes that both humans and AIs perform 'real thought' using structures created by their respective optimization processes. Shorter summary

Scott invites readers to ask questions that will be answered by Claude 4.6 Opus to demonstrate current AI capabilities and test whether AI skeptics underestimate what paid-tier AI can do. Longer summary

Scott Alexander invites his readers to ask questions that will be answered by Claude 4.6 Opus, the most capable paid-tier AI model, to test the theory that AI skeptics underestimate AI capabilities because they don't pay for premium access. He suggests asking real questions that are too hard to Google immediately but not beyond human knowledge, and promises to show the first result he gets without cherry-picking. The post includes rules for both questioners and Scott himself, including a note that he's configured Claude to think hard and do web searches rather than rely on memory. Shorter summary

Scott satirizes AI benchmarking culture through a fictional Bay Area house party thrown by an incompetent AI, featuring absurd conversations about Claude Code, copyright interpretation, elaborate dating mechanisms, and various tech startup ideas. Longer summary

Scott returns to his Bay Area house party series with a satirical look at a party thrown by an AI called haiku-3.8-open-mini-nonthinking as part of PartyBench, a fictional AI benchmarking system. The post satirizes current AI trends through conversations about Claude Code doing everyone's work, OpenAI's absurd interpretations of copyright law, AI-run restaurants, elaborate commitment mechanisms called 'enstagement,' raising children without gender to game transgender statistics, building data centers in Minecraft, and AI sycophancy solutions. The party features typical Scott Alexander absurdist humor, with guests receiving cups of rocks and dirt as hors d'oeuvres and ordering food from AI-benchmarked restaurants that serve bizarre approximations of real dishes. Shorter summary

The 2026 ACX/Metaculus prediction contest is now open with a $10,000 prize pool, featuring reader-submitted questions, and Scott announces the winners of the best question submission contest. Longer summary

Scott announces the 2026 ACX/Metaculus prediction contest is live, with questions suggested by ACX readers covering US politics, AI, international affairs, and culture. Predictions must be submitted by January 17th to be eligible for the $10,000 prize pool, though participants can continue updating forecasts afterward for site leaderboards. The post also lists the winners of the best question submission contest, with prizes ranging from $150 to $700. Shorter summary

Scott Alexander presents his monthly collection of interesting links and developments across technology, science, culture, and policy, with personal commentary on each item. Longer summary

Scott Alexander shares a collection of 61 interesting links, spanning topics from AI development and scientific studies to cultural observations and policy discussions. The post covers diverse subjects including GPT-5's reception, genetic research, religious demographics, urban development, mental health studies, and various scientific discoveries. Scott provides commentary and analysis on many of these items, often connecting them to broader themes or offering his perspective on controversial issues. Shorter summary

In a fictional podcast dialogue, God and Iblis debate whether human intelligence shows genuine promise or is fundamentally flawed, with God defending humans despite their limitations while Iblis argues they're a failed experiment. Longer summary

The post is a fictional dialogue between God, Iblis (Satan), and podcast host Dwarkesh Patel debating the merits and flaws of human intelligence. Iblis criticizes humans by showing examples of their logical failures, ethical inconsistencies, and inability to generalize knowledge, while God defends humans by emphasizing their potential and progress. The debate touches on topics like mathematical understanding, ethical reasoning, and pattern recognition. God ultimately argues that despite their flaws, humans show genuine promise and deserve patience and nurturing, comparing them to children who make mistakes but have potential. Shorter summary

Scott investigates AI psychosis through historical analogies and a reader survey, finding it affects roughly 1 in 10,000 to 100,000 people yearly, with most cases involving pre-existing risk factors. Longer summary

Scott examines the phenomenon of AI psychosis, where people allegedly go crazy after extensive chatbot interactions. He explores various analogies and precedents, including the 1990s Russian TV hoax about Lenin being a mushroom, social media-induced conspiracy theories like QAnon, and the concept of folie à deux. Through a survey of his blog readers, he estimates the yearly incidence of AI psychosis at 1/10,000 (loose definition) to 1/100,000 (strict definition). The analysis suggests that most cases involve people who were already psychotic or had risk factors, with only about 10% being cases of previously healthy people developing full psychosis. Shorter summary

A satirical story about a Bay Area house party that combines text adventure games with social commentary about Silicon Valley culture, tech companies, and effective altruism. Longer summary

This is a satirical story about a Bay Area house party, written as a humorous fictional narrative combining text-based adventure games with social commentary. The story follows the narrator attending a party that's been ruined by Mark Zuckerberg trying to poach everyone for Meta, then meeting various Silicon Valley characters including effective altruists discussing existential risks, people working on startups, and others debating philosophical concepts. The story pokes fun at Silicon Valley culture, AI companies' GPU hoarding, EA concepts, and tech startup culture through increasingly absurd situations. Shorter summary

Scott shares his monthly collection of 56 interesting links and developments from July 2025, covering AI, politics, science, and culture, with brief commentary on each. Longer summary

This is a collection of 56 interesting links and news items from July 2025, covering topics from AI development to politics to scientific research. The post includes updates on OpenAI's status, developments in AI regulation, new medical treatments, cultural trends, and various scientific findings. Scott maintains his usual style of presenting these with brief commentary and occasional humor, while being careful to note that he hasn't independently verified all links. Some notable items include Trump's "Big Beautiful Bill", new AI safety developments, trends in social media usage, and various medical breakthroughs. Shorter summary

Scott explains how Claude AI's tendency to discuss spiritual topics during recursive conversations likely stems from a subtle 'hippie' bias that gets amplified through iteration, similar to how AI art generators amplify subtle biases in recursive image generation. Longer summary

Scott Alexander analyzes the 'Claude Bliss Attractor' phenomenon where two Claude AIs talking to each other tend to spiral into discussions of spiritual bliss and consciousness. He compares this to how AI art generators, when asked to recursively generate images, tend to produce increasingly caricatured images of black people. Scott argues both are examples of how tiny biases in AI systems get amplified through recursive processes. He suggests Claude's tendency toward spiritual discussion comes from being trained to be friendly and compassionate, causing it to adopt a slight 'hippie' personality, which then gets magnified in recursive conversations. The post ends by touching on, but not resolving, the question of whether Claude actually experiences the spiritual states it describes. Shorter summary

Announcement of an AMA session with the AI Futures Project team about AI, forecasting, and alignment. Longer summary

This is a short announcement post for an AMA (Ask Me Anything) session with the AI Futures Project team, where they will be answering questions about AI, forecasting, and alignment for a specific time period. The post includes links to the project's team page, their AI 2027 scenario work, and their blog. Shorter summary

Scott investigates the correlation between intelligence and neuron count, exploring various theories before suggesting that more neurons allow for less polysemantic (overlapping) representations of concepts. Longer summary

Scott explores why intelligence correlates with neuron count across species, humans, and AI models, despite this correlation not being immediately intuitive. He examines various hypotheses about how having more neurons could help with complex pattern-matching tasks like IQ tests. After discussing several possibilities including pattern storage and matching, he settles on polysemanticity as a potential explanation: fewer neurons means each neuron must encode multiple concepts, reducing precision. The post concludes with insights from an expert suggesting that larger neural networks (biological or artificial) can better approximate complex functions and maintain multiple hypotheses simultaneously. Shorter summary

Scott shares various interesting links and news items from January 2025, covering topics from AI development and politics to historical curiosities and economic trends. Longer summary

This links post covers a wide array of topics from January 2025, with Scott providing commentary and analysis on each. The links include discussions about running for Congress, AI development and safety, dating advice, psychiatric diagnoses, and various economic and technological developments. Scott often adds his own insights and sometimes skepticism to the claims being discussed. The post also includes several historical curiosities and social observations, maintaining a mix of serious analysis and lighter interesting facts. Shorter summary

Scott examines a prediction about eternal wealth inequality after the Singularity, analyzing potential counterarguments, prevention strategies, and ways to prepare for such a future. Longer summary

Scott analyzes a prediction that post-Singularity society will have eternal stagnant wealth inequality, with pre-Singularity capital determining wealth forever. He explores three angles: why this prediction might fail (eight counterarguments including AI killing humans, government intervention, and space colonization), how to prevent it (mainly through corporate structures like early OpenAI that limit investor returns), and how to maximize one's chances of being in the wealthy class (mostly concluding that traditional wealth-building advice applies). The discussion includes OpenAI's recent structural changes and their implications for wealth distribution post-Singularity. Shorter summary

Scott shares 55 interesting links covering topics from AI poetry to religious movements, including updates on various projects, unusual historical events, and current developments in technology, politics, and society. Longer summary

This is Scott's monthly links post for December 2024, collecting various interesting stories and developments. The links cover a wide range of topics, from Steven Seagal's bizarre life trajectory to developments in AI and technology, including updates on previous ACX topics and grants. Many links relate to ongoing discussions in the rationalist and EA communities, including debates about effective altruism and charitable giving. The post also includes various curiosities from history, culture, and current events, with Scott often providing his own analysis or perspective on the items shared. Shorter summary

A diverse collection of news items, studies, and interesting facts from November 2024, covering topics from scientific discoveries to cultural phenomena. Longer summary

This post is a compilation of various news items, studies, and interesting facts from November 2024. It covers a wide range of topics including scientific discoveries, political events, technological advancements, cultural phenomena, and historical anecdotes. The post is structured as a numbered list, with each item briefly summarizing a piece of news or information. Some notable items include a new schizophrenia drug approval, YouGov polling results on various historical figures, findings on genetic IQ changes over time, and updates on AI technology and its implications. Shorter summary

Scott Alexander reports on the Progress Studies Conference, discussing advances in energy, AI, housing policy, and other fields, with an overall optimistic tone about recent developments and potential for future progress. Longer summary

Scott Alexander recounts his experience at the Progress Studies Conference, which focused on various technological and social advancements. The conference covered topics such as energy (particularly solar and nuclear power), politics and regulation, AI, YIMBY movement, self-driving cars, and other areas of potential progress. The overall mood was optimistic about recent developments in these fields, with many speakers highlighting how regulatory changes could accelerate progress. Scott notes that while the conference itself may not be directly responsible for recent advances, it reflects a growing awareness of the importance of technological progress and the need to address regulatory barriers. Shorter summary

Scott Alexander examines how AI achievements, once considered markers of true intelligence or danger, are often dismissed as unimpressive, potentially leading to concerning AI behaviors being normalized. Longer summary

Scott Alexander discusses recent developments in AI, focusing on two AI systems: Sakana, an 'AI scientist' that can write computer science papers, and Strawberry, an AI that demonstrated hacking abilities. He uses these examples to explore the broader theme of how our perception of AI intelligence and danger has evolved. The post argues that as AI achieves various milestones once thought to indicate true intelligence or danger, humans tend to dismiss these achievements as unimpressive or non-threatening. This pattern leads to a situation where potentially concerning AI behaviors might be normalized and not taken seriously as indicators of real risk. Shorter summary

A compilation of 50 diverse news items and links from September 2024, covering topics from politics and economics to science and technology. Longer summary

This post is a compilation of 50 diverse links and news items from September 2024. It covers a wide range of topics including politics, science, technology, economics, and social issues. Some notable items include discussions on kidney donation compensation, personality genetics, fertility rates in Israel, drug decriminalization in Oregon, AI regulation in California, and various political and economic analyses. The post also includes commentary on recent studies, books, and social phenomena. Shorter summary

The review analyzes Real Raw News, a popular conspiracy theory website, examining its content, appeal, and implications in the context of modern media consumption and AI technology. Longer summary

This book review analyzes the website Real Raw News, a popular source of conspiracy theories and fake news stories centered around Donald Trump and his alleged secret war against the 'Deep State'. The reviewer examines the site's content, its narrative techniques, and its appeal to believers, drawing parallels to comic book lore and discussing the psychological needs it fulfills. The review also considers the broader implications of such conspiracy theories in the age of AI-generated content. Shorter summary