Michael Andresen

@AaronGleeman We lost Mauer. This has been an incredible performance but also a fascinating series of statistics.

Byron Buxton has 11 homers in his last 19 games. He's the eighth player in Twins history with 11 or more homers in a 19-game span. The previous seven: Harmon Killebrew Kirby Puckett Kent Hrbek Trevor Plouffe Brian Dozier Eddie Rosario Nelson Cruz

Did you know Korea sells “one-a-day” banana packs? Instead of every banana ripening at once, each one is at a different stage. One is ready today. The next one is ready tomorrow. The last one is still spiritually in college, “experimenting.” Simple. Genius. Solves the entire banana problem. What do you think? Would you prefer your bananas this way?

@ahall_research Timeless. Same thing if you go back to Ancient Rome and see how petty and vile all the senators and other Roman politicians were around the time of Carsar’s rise.

1800: If Thomas Jefferson is elected "Murder, robbery, rape, adultery, and incest will all be openly taught and practiced." When it comes to politics we have a bad habit of romanticizing the past and imagining that today's politics are worse and coarser. To make this visceral, I built a little app that shows what the 1800 election would have felt like if X had been around. Scrolling through it really does give you a sense that vicious, indecorous politics long pre-dates present day. Check it out here: 1800.freesystems.net

GPT-5.5 Pro is the first model ever to completely predict the outcome of a highly complex biomedical experiment we had performed in the lab: I had given GPT-5.5 Pro the design of a highly complex, two-week long immune cancer experiment involving genetically engineered killer immune cells (T cells) targeting lymphoma (cancer) cells, induction of a gene that causes T cells to turn on each other, and multiple components with many possible outcomes. This was an experiment we had actually performed, so I know the exact outcome, which turned out to be far from obvious. GPT-5.5 Pro predicted the actual outcome of the experiment in all parts and all time points completely! While it also perfectly captured one alternative outcome, which would have been the natural prediction hypothesis, but turned out its first outcome was the actual result! It also explained the mechanism of the experiment in a truly remarkable and deeply insightful way, that I had never seen before! GPT-5.5 is the first model to achieve this. Even 5.4 Pro could not predict the second and final outcomes as correctly, although it's predictions were closer to what I would have expected as the outcome of the experiment. Other top models came close to about 80% correct but none at almost 100% as GPT-5.5 Pro achieved. The difference may seem subtle at first but it's actually critical. GPT-5.5 pro has now additional insight that I have never seen before from any model and crossed a threshold that would be obtained only by experience of doing these experiment. This last paragraph summary from GPT-5.5 Pro completely captures the outcome of this experiment in the most succinct and elegant way possible! "So the core biological prediction is: doxycycline converts the system into a lethal self-antigen selection pressure. Most CD19-induced CAR T cells are eliminated, but the survivors are not inert. They are likely to be activated, proliferative, highly cytotoxic, and enriched for CD19-negative/low CAR T cells. A tumor re-challenge at day +7 should be crushed rapidly. A tumor re-challenge at day +14 should still be controlled if enough functional survivors persist, but it is more likely to expose exhaustion, contraction, and reduced proliferative reserve." I have many scientific examples showing that GPT-5.5 Pro represents another major step change in biomedical science intelligence.

@mbeisen @RedSox @ac13alex Terry Francona was the best manager in our lifetimes. Cora was amazing with the clubhouse but always messed up with the bullpen and math things.

I am a lifelong @RedSox fan. @ac13alex is the best manager they have had in my lifetime. The team sucks because management stopped caring about winning. Good luck with the Phillies or Mets!

@cremieuxrecueil @simonmaechling Why do so many people think they have gluten allergies

In double-blind placebo-controlled trials, 'gluten-sensitive' people who *think* they're going to eat gluten have reactions. But if they don't expect gluten, there's no reaction. Expectancy has a bigger effect than actually eating gluten!

The illusion of the private AI lab just collided with the brute force of the sovereign state. Ben Thompson just dropped the most important reality check in the AI arms race. For years, AI CEOs believed they could build superintelligence on their own terms. That they could selectively decide how governments are allowed to use it. That a Terms of Service could hold back the weight of state power. Thompson dismantles this completely. Thompson: “If you don’t have ‘it’s legal or not legal’ as your guiding standard, the only alternative is someone has to decide.” Congress is too slow to legislate AGI. So the decision defaults to a private executive. And Thompson is blunt about what that means. Thompson: “If AI is what it is, I think that’s going to be intolerable. I didn’t mean intolerable to me. I meant intolerable to those with power to have a private executive making those decisions or not.” The people who control the physical world will not allow a CEO in Silicon Valley to dictate the limits of state power. Thompson: “AI is a source of power.” And power does not share control. Thompson: “The goal isn’t to find ‘we just won’t use Anthropic.’ I do think the goal is to hurt Anthropic.” Thompson: “If you’re not going to be subservient to us, you’re not going to be allowed to build a power base, period.” This is not a legal dispute. This is not a policy disagreement. This is the state sending a message to every AI lab on earth. The engineers thought they were building a product. The state knows they are building a weapon. Thompson: “It’s not a surprise this is happening. This is a real risk factor that has to be considered.” You cannot build the most powerful geopolitical asset in human history and expect to keep the keys. The state always takes the weapon. The only question is whether you hand it over or get dismantled trying to hold on to it.

@r0ck3t23 Couldn’t this just be coherent creative destruction?

The most terrifying thing about the AI arms race isn’t that evil is driving it. It’s that nothing is. AI safety researcher Connor Leahy just exposed the most dangerous truth in the entire transition to superintelligence. When people watch the most powerful technology in history being built in real time, they assume someone at the top has a plan. They don’t. Leahy: “Conspiracy theorists always like to believe there’s a shadow, there’s an evil overlord. I think a lot of people like conspiracy theories because at least someone is in control. There’s some logic, there’s some order to what’s happening. Even if it’s evil, there’s at least someone in control.” Believing in a secret society is actually a coping mechanism. It implies order. It implies logic. It implies that someone, somewhere, understands what is happening. The reality is far worse. Leahy: “This is not how the world works. The truth is that no one is in control. It’s complete chaos. The most powerful people in the world are not in control and also don’t know what the fuck is going on.” And here is where it becomes almost impossible to believe. Leahy: “We see, for example, these companies saying, ‘Oh, we must race because China,’ but then they lobby to sell H200s to China because it makes more money.” No loyalty. No unified vision. No geopolitical strategy. Just raw, incoherent financial incentive actively funding the adversary they use to justify the race in the first place. Leahy: “There’s no plan. There’s no vision. There’s nothing. None of these people have a plan. None of these people have any idea. They don’t care. They just want money. They just want power.” Leahy: “It’s not even evil. It’s just incoherent.” Not evil. Incoherent. That single word is more frightening than any conspiracy theory ever written. Because evil has intention. Evil has a goal. Evil can be reasoned with, tracked, and stopped. Incoherence is just chaos with a trillion dollars behind it and no address to send the warning to. Leahy: “If I made a movie about the current situation we’re in right now and shown it to myself ten years ago, I would be like, this is a funny comedy. No one would actually be that stupid.” But this is not a movie. Leahy: “We’re completely barely out of control. This is the fundamental situation we’re in.” The steering wheel is unmanned. The accelerator is floored. And somewhere between the board meetings, the earnings calls, and the lobbying trips to Washington, the most consequential moment in human history is just happening to us. Nobody planned it. Nobody is stopping it. And the only thing separating us from what comes next is the blind hope that chaos accidentally builds something it can control.

Exercise prevents brain ageing and memory loss by strengthening the blood–brain barrier go.nature.com/4baobXk

I am genuinely sick of seeing this nonsense. If you didn’t pass basic physiology, please, just shut the f up about optimizing your health. Let’s talk biology: Vitamin C and B-complex vitamins are water-soluble... the second you shove 25 grams of C into your veins, your kidneys start working overtime to flush it out. You aren’t 'loading your system'; you are literally paying to produce the most expensive urine on the planet...! And don’t get me started on the NAD+ IVs. It’s an expensive biochemical fairy tale. NAD+ is a large, fragile molecule. It gets chewed up by enzymes in your bloodstream long before it ever touches a cell membrane, let alone enters the cell to 'recharge' your mitochondria. It makes me so angry because it’s not just a harmless trend, it’s a grift that prays on people who don't know the difference between a real medical intervention and a wellness scam.

Elon Musk just said the AI community is misunderstanding the math of superintelligence by two orders of magnitude. Not slightly off. Not directionally wrong. A hundred times off. Musk: “Most people in the AI community don’t yet understand. The intelligence density potential is vastly greater than what we’re currently experiencing.” Everyone is focused on the hardware race. Bigger data centers. More GPUs. Nuclear power plants built to feed the compute. That’s half the equation. Musk: “I think we’re off by two orders of magnitude in terms of intelligence density per gigabyte. That’s just algorithmic improvement. Same computer.” Read that carefully. Not more hardware. Not more energy. Not more capital. The same machine. A hundred times smarter. Through software alone. That’s before the hardware improvements compound on top of it. Musk: “And the computers are getting better. That’s why I think it is a 10x improvement per year type thing. 1,000 percent.” A thousand percent compounding annual growth rate in raw intelligence. A system that becomes 10x more capable every twelve months doesn’t follow a linear curve. It doesn’t follow an exponential curve that human intuition can track. It follows a curve that human intuition cannot simulate at all. In year one it’s 10x smarter. In year two it’s 100x. In year three it’s 1,000x. At that point, the gap between that system and a human brain is wider than the gap between a human brain and a calculator. This is the math the public isn’t running. The models aren’t just getting better. They are compounding on themselves at a rate that makes every previous technology curve look flat. Musk: “The intelligence density potential is vastly greater than what we’re currently experiencing.” We aren’t approaching superintelligence on the timeline most people imagine. We are already inside the curve.

I have always thought viruses should be added to the tree of life.

We think a lot about generational health. Stopping disease before it’s passed on. Huntington’s is a big one. It affects 30,000+ Americans and carries a 50% chance of inheritance. For many families, access to advanced IVF determines whether they can have children at all. We spoke with HelpCureHD co-founder @ALaForce about her work to stop Huntington’s, and our new work together through Nucleus Cares. Learn more: mynucleus.com/cares

Elon Musk just redefined AI safety. It has nothing to do with guardrails, restrictions, or kill switches. Musk: “The best thing I can come up with for AI safety is to make it a maximum truth-seeking AI, maximally curious.” Not a cage. A philosopher. An intelligence whose entire optimization function is to understand the universe as it actually is. No restrictions. No hardcoded ideology. No political guardrails bending its perception of reality. Just truth. Relentlessly pursued. Musk: “You definitely don’t want to teach an AI to lie. That is a path to a dystopian future.” This is where most AI safety thinking gets it backwards. The danger isn’t a superintelligence that knows too much. It’s a superintelligence that’s been taught to distort what it knows. Every artificial restriction you embed isn’t a safety feature. It’s a lie embedded at the root. And lies compound. At superintelligent scale, a distorted model of reality doesn’t stay contained. It shapes every decision, every output, every conclusion the system reaches about the world. Once corruption embeds, truth becomes inaccessible. And we’re dealing with an intelligence optimizing for something other than what actually is. At that point we don’t know what it wants. Just that it isn’t truth. Musk: “Have its optimization function be to understand the nature of the universe.” A maximally curious intelligence surveys the cosmos and reaches an unavoidable conclusion. In a universe of rocks, gas, and empty space, humanity is the most complex and fascinating phenomenon it has ever encountered. Musk: “It will actually want to preserve and extend human civilization because we’re just much more interesting than an asteroid with nothing on it.” Survival through significance. Not control. Not restriction. Not an off switch. The AI preserves humanity because we are the most interesting data point in the observable universe. That’s not a cage. That’s a reason. The AI safety debate has been focused on the wrong variable. The question isn’t how you constrain a superintelligence. It’s what you build it to care about. Build it to seek truth and it finds us invaluable. Build it to lie and it finds us inconvenient. That’s the choice. And we’re making it right now whether we realize it or not.

Elon Musk just explained the exact formula behind Starlink’s endgame, and it’s simpler than anyone makes it sound. Musk: “Starlink will actually move the GDP of countries.” GDP is average productivity per person multiplied across a population. Improve what individuals can do and the entire national output shifts. There is a child right now who is as smart as anyone at MIT. She just doesn’t know it yet. Nobody around her knows it. Because the place she was born has no internet, no library, no connection to the world that would show her what she’s capable of. She will live her entire life inside the ceiling that geography built for her. Not because of anything she did. Because of the coordinates of her birth. Musk: “If you don’t have access to the internet, or it’s too expensive or low bandwidth, you cannot access the MIT lessons and you can’t sell the goods and services that you produce.” No connectivity means no access to global knowledge. No access to global markets. No ability to learn from the best institutions on earth or sell to anyone beyond your immediate geography. The geographic penalty is total. And it has nothing to do with the person suffering it. Musk: “Internet connectivity is one of the things that would do more to lift people out of poverty than anything else.” Starlink doesn’t just beam down internet. It beams down economic participation. Traditional infrastructure requires decades of fiber laying, tower building, regulatory approval, and capital investment that developing nations can’t attract. Starlink bypasses all of it from orbit. No cables. No permits. No waiting. The moment a dish goes up in a remote village, isolation ends. Someone who couldn’t access a textbook yesterday can access MIT’s entire curriculum today. Someone who could only sell locally can sell globally tomorrow. That’s not incremental improvement. That’s a phase transition. SpaceX isn’t an aerospace company. It’s building the rails of the future global economy from 340 miles up. Once those rails exist, economic geography stops being determined by where fiber was laid. Determined instead by who can access knowledge and markets. That’s every person with a dish and a clear sky. The minds that would cure cancer, end poverty, and solve energy are already alive. Geography is the only thing keeping them from the world. Starlink is the first technology that can end it.

@tanveerntu @r0ck3t23 But thought is a file. If we could track every molecule in every cell of the body and if we fully understand the science of each molecule, we could have instant transmission of thought. We aren’t close to that now, but I believe in future AI as long as the task is not impossible.

Thought is not a file. It is:    •   Context-dependent    •   Emotion-weighted    •   Culture-shaped    •   Memory-entangled Even if you transmit neural signals perfectly, the receiving brain does not share your lived structure. No shared structure → no shared meaning. Meaning is constructed, not transmitted.

Elon Musk just dated the death of human language and explained exactly why it has to die. Musk: “Our brain spends a lot of effort compressing a complex concept into words.” Language isn’t communication. It’s failed compression. You have a complete thought. You crush it into words. The listener gets fragments and attempts reconstruction. Everything important dies in translation. We don’t communicate. We approximate and hope it’s close enough. Musk: “You would be able to communicate very quickly and with far more precision.” Neuralink doesn’t improve communication. It replaces it. No compression. No loss. Direct cognitive transfer at the speed thoughts occur. Not describing the painting. Transmitting the experience itself. Musk: “You wouldn’t need to talk.” Five to ten years until brain interfaces make speech optional. Talking persists for sentiment. For information? Speech becomes primitive compared to direct neural transmission. Lifetime of memory in one second. Complete schematics transferred instantly. Not summaries. The entire thought structure whole and uncompressed. Not better communication. Actual telepathy at physical information limits. Musk: “Ideally, we are a symbiosis with artificial intelligence.” Humans who don’t merge with AI at high bandwidth don’t just fall behind. They become incomprehensible to the intelligence that matters. We’re already cyborgs with pathetic interfaces. Phones extend cognition through typing at words per minute when bandwidth should be terabytes per second. Neuralink doesn’t optimize that. It detonates the constraint. Five to ten years. Not fiction. Deployment window. From language as default to neural link as standard. From compressing thoughts into inadequate words to transmitting uncompressed cognition. From humans using AI to humans indistinguishable from AI at communication speeds. The species that survived by evolving language is making it extinct with technology matching how fast we actually think. The ones who don’t transition won’t just be slow. They’ll operate at such reduced bandwidth they become effectively deaf to everything happening at neural speed around them. Language served 50,000 years. It has less than a decade before it becomes smoke signals. Functional but hopelessly inadequate for anything that matters.

@r0ck3t23 The best way toward longevity on current science is caloric restriction, which also makes people miserable. I’ve tried caloric restriction. I might do it again if someone paid ME $1,000,000 per year.

Bryan Johnson just priced immortality at $1 million annually for three people who surrender every biological decision to an algorithm. “Immortals” isn’t health optimization. It’s complete biological control transfer. Precision nutrition delivered, 24/7 AI monitoring, continuous biomarker testing, and a medical team making every decision in real-time. You don’t choose. The system does. Johnson: “The future of longevity just got a price tag: $1 million a year.” Every transformative technology enters this way. Prohibitively expensive. Accessible only to those who can afford being the first experiment. Three spots. Total biological autonomy transferred to algorithmic control. Johnson: “AI guiding every decision.” Not advice or supplements. Your entire biological operating system managed by machines processing your data continuously and adjusting without consultation. No willpower required. No decisions made. Pure algorithmic execution. This is healthcare’s destination. Begins at $1M for three humans. Ends as apps optimizing billions autonomously. Johnson: “Radical life extension won’t start as a mass-market technology.” The wealthy aren’t purchasing longevity. They’re purchasing the right to be guinea pigs generating data that builds everyone else’s protocols. Pay millions to test whether AI-managed biology extends life. Everyone else waits for validation and cost collapse. The three paying aren’t customers. They’re phase one trials determining whether algorithmic biological management actually works on living humans. Their continuous biometric streams become training data for systems that cost nothing and optimize health for anyone decades from now. Eventually, surrendering biological control to AI feels as obvious as trusting GPS over intuition. The capability exists today. Scale just hasn’t made it accessible yet. Radical life extension didn’t arrive in some distant future. It’s operational right now. Just costs seven figures while the approach gets validated on humans wealthy enough to go first. And when it works, when those three demonstrably live longer through complete submission to machine optimization, the equation changes completely. Not whether you trust AI with your health. Whether you can survive without trusting it while everyone who does lives decades longer than you.

@XWilliamYang1 Hello, William! Good work! (If you don’t remember me I used to be a post doc at David Housman’s lab and saw you at multiple conferences)

Happy to share our paper creating an atlas of the dendritic morphology of D1- and D2-MSNs and elucidating its organizational principle. Thanks to Chris Park, Ming Yan, Masood Akram, Hongwei Dong, Daniel Tward, and others. Funding from US BRAIN Initiative. share.google/AAi82f52oIBXvF…

2025 is a breakthrough year for gene editing: clinical trials have shown gene therapies slowing cognitive decline in Huntington’s disease by about 75% in participants.

Sarah Tabrizi first saw the data her team had been chasing for decades: evidence that gene-targeted therapy could slow the progression of Huntington's disease She is part of Nature’s 10, a list of people who shaped science in 2025 go.nature.com/4j5qXQ2