steve hsu
@hsu_steve • 46,461 subscribers
Physicist, AI Founder, Manifold Podcast
Videos
Clip from excellent DW documentary on drone warfare. "War against Russia is one thing. War against China is quite another. China make all our electronic components. You know, I use electric components to make circuit boards, to make timers for the bombs we use all the time. Can I find components that are made outside of China? No. So, what happens if we go to war with China and we suddenly feel and realize that we haven't got factories that make a lot of stuff that becomes essential..." Note, Dutch guy is wrong about DJI being a state-owned company. Worth watching the whole thing:
steve hsu266,220 次观看 • 27 天前
RSI section from the AI documentary Machine God The next threshold is Recursive Self-Improvement: the moment when AI can improve itself without human assistance. For decades this sounded like science fiction. Intelligence explosion scenarios imagined a system rewriting its own code, becoming smarter, then using that new intelligence to make still better versions of itself. But the idea looks less remote now that AI contributes directly to frontier science. In mathematics, recent systems have moved beyond solving contest problems to producing serious new arguments on long-standing open problems. AI is used to build physics world models and propose candidate theories or computational methods. These are early signs that machine cognition is entering the creative loop of science itself. The crucial transition comes when that loop turns inward. AI research is, after all, a technical discipline made of code, mathematics, models of information flow. These are exactly the domains in which frontier models are improving fastest. A model that can solve hard mathematical problems, write production-quality code, design experiments, read the literature, and evaluate benchmark results is already participating in the work of building its successor. At first this will look prosaic. AI systems will write kernel optimizations, improve training infrastructure, discover better data filters, tune reinforcement-learning pipelines, design new benchmarks, and suggest architectural modifications. Human researchers will remain in the loop, approving changes and interpreting results. But the important point is that the search process accelerates. The model becomes not just the product of the lab, but part of the lab’s research machinery. The system being optimized helps optimize the next system. This is the core RSI feedback loop: better models make AI research faster; faster AI research produces still better models; those models, in turn, become better researchers. The danger is that once this loop becomes sufficiently autonomous, it may stop resembling ordinary technological progress. Human institutions are slow because humans are slow: we read papers, attend meetings, debug code, sleep, argue, and wait for funding cycles. Machines do not have to operate on that timescale. An AI research collective can run continuously across millions of processors. This is the runaway possibility. Not that an AI instantly wakes up and recursively rewrites itself into a god, but that the entire AI ecosystem becomes an autocatalytic process. Capital buys compute; compute trains models; models improve models; better models attract more capital. At some point the dominant input into AI progress may no longer be human insight, but machine-generated insight, machine-written code, and machine-run experiments. Then the Butler-Land analogy becomes sharper. Humanity is no longer merely building machines. We are building machines that help build better machines. Once intelligence itself becomes part of the production function, the old categories — tool, worker, inventor, firm, market — begin to blur. The question is whether recursive self-improvement remains a managed industrial process, or whether it becomes the first technological process in history whose natural endpoint lies beyond human comprehension.
steve hsu61,671 次观看 • 14 天前
Key fact claims for anyone watching the AI boom/bubble! They may contradict the Bridgewater AI slides I posted recently. Ed Zitron According to Dylan, both OpenAI and Anthropic have shifted from being venture-funded loss-making companies to generating massive positive gross margins (01:39 - 02:37). Key points regarding their profitability per gigawatt include: Revenue vs. Cost: While the base cost of compute is roughly $10 million to $15 million per megawatt, these labs have significantly increased their ability to monetize that capacity (02:29 - 03:00). Margin Improvement: OpenAI’s newer models (like 5.6) and Anthropic’s models (such as Opus 5 or Mythos) have pushed revenue generation well beyond the incremental cost of compute (02:54 - 03:02). Monetization Scaling: In the case of Anthropic, revenue has reached as high as $50 million per megawatt, allowing them to take the profit from that inference capacity and reinvest it heavily into training (03:07 - 03:23). Listen further for incredible projections: OAI + Ant will control most of the computing power on Earth within next few years?
steve hsu14,230 次观看 • 26 天前
Letter from Beijing 2: Tsinghua University – Manifold #113 This special episode was recorded at Tsinghua University in Beijing, generally regarded as the top university in China. Our guests are 3 Americans studying and working at Tsinghua: Gabriel (undergrad), Justin (PhD student in AI), and Alex (Professor in AI research). Topics discussed include: Tsinghua University and elite human capital, AI in China, US-China competition, and the flow of human capital between the US and China Han Feizi, columnist at Asia Times and the guest from the previous "Letter from Beijing" episode, is also in the room. Chapter Markers: (00:00) - Welcome to Tsinghua University (02:47) - Gabriel’s Undergrad Journey (12:35) - Justin’s PhD (25:10) - Professor Alex on AI and Rankings (42:51) - Second Chances and Status Signals (46:48) - China’s Exam Ladder Explained (50:20) - Infrastructure and Tech Competition (01:17:18) - Semiconductors, EUV, and Wrap Up
steve hsu39,015 次观看 • 3 个月前
This is a small Vision-Language model running on an inexpensive phone chipset. Executes reasoning task while controlling UAV (drone) in sim world. The task is to find the human doctor near the medic truck. If desired, the drone could be instructed to hit the doctor or the truck. SuperfocusAI
steve hsu22,608 次观看 • 2 个月前
What happens if we create a machine more intelligent than humanity—and then discover we can no longer control it? Machine God has been submitted to both the Sundance and SXSW film festivals! We can provide clips upon request, and arrange limited private showings. Short Summary: Artificial intelligence is advancing so quickly that some of the people building it believe we may soon create intelligence greater than our own. The film follows leading AI researchers, entrepreneurs, accelerationists, and safety advocates to explore this question. Dominic Cummings Dwarkesh Patel Sam Altman Beff (e/acc) Scott Alexander
steve hsu11,348 次观看 • 28 天前
Industrial Maximalism and Its Discontents: Dan Wang on US-China Competition – Manifold #104 Great conversation with Dan Wang at Hoover. 0:00 - Introduction and Welcome 02:14 - Breakneck - Dan's huge book 05:00 - China's Technological and Political Landscape 21:07 - Industrial Maximalism and its Discontents 47:59 - Chinese Researchers in Silicon Valley and Tsinghua 51:09 - Excerpts from Dan's 2025 annual letter 52:56 - China's Market Competition and Innovation 56:34 - AI, Automation, and Future Risks
steve hsu51,991 次观看 • 7 个月前
Embryo Selection and Frontier Genomics with Dr. Alex Young – Manifold #111 Dr. Alex Young, a statistical geneticist and assistant professor in the Human Genetics department at UCLA, joins Steve Hsu to discuss the cutting edge of genomic prediction. They cover his research on polygenic embryo screening in IVF (including the ImputePGTA method), family-based DNA analysis, missing heritability, and the implications of polygenic scores for traits like education and disease. Alex also discusses his recent battles with cancer. Alex Strudwick Young Chapter Markers: (00:00) - Alex Young Bio (06:36) - Biobank Era Genetics (10:49) - Missing Heritability Debate (27:18) - Embryo Selection Controversy (50:32) - Embryo Selection Backlash (53:42) - Mexico City Admixture Study (01:00:13) - Censorship Via Data Access Control (01:05:02) - Battle With Cancer and Circulating Tumor DNA (ctDNA)
steve hsu19,551 次观看 • 4 个月前