Brian Crabtree

Definitions

Catuṣkoṭi is not a four-option answer machine. It is a closure-destruction operator. It takes a proposition P and exhausts the four ways the mind tries to stabilize it: P, not-P, both P and not-P, neither P nor not-P. In Madhyamaka, this is used to show that the proposition lacks independent self-being, not to install a fifth metaphysical answer. Catuṣkoṭi discovers by exhausting closure. It shows that what looked like a fact was a functional relation mistaken for an essence.

1/Strike a match and you get a flame. The most ordinary thing in the world. Around the 2nd century Nāgārjuna stared at that tiny event and built an argument that cause and effect cannot really work the way every one of us assumes. It is far harder to wriggle out of than it looks. Let me take you through the whole thing slowly. Thread🧵🪷☸️ 1/42

@CharlesMullins2 Direct electromagnetic recovery from plasma motion doesn't magically solve the underlying fusion gain problem

🚨 FUSION COMPANIES ARE NOW TRYING TO SKIP THE TURBINE ENTIRELY. Most power plants even nuclear ones still rely on 19th-century technology: heat water → make steam → spin a giant turbine. But companies like Helion and TAE are attempting something radically different. They want to pull electricity directly from the fusion plasma itself using electromagnetic induction. No boiling water. No steam turbines. Just magnetic fields converting the motion of the plasma straight into electrical current. Why this matters: Direct energy conversion could make fusion: • Smaller and cheaper to build • Significantly more efficient • Faster to scale globally Instead of massive power stations, future reactors could behave more like advanced electromagnetic engines. The deeper implication is staggering: Humanity may be approaching the moment where we stop “burning” anything for energy. We would directly manipulate plasma and magnetic fields as programmable energy systems. Not fire. Not steam. Controlled stars turned straight into electricity. The line between reactor, battery, and electromagnetic machine could start to disappear. What happens when fusion power becomes this direct and scalable? Follow for more frontier physics and future technology.

In practice, the ledger acts as the persistent memory of the propagation process. It makes discovery reproducible and auditable: every discrete residue extracted via Ambrose–Singer spanning, every deformation functor application, and every bifurcation is explicitly recorded with its dependency subgraph. This completes the administrative backbone required for systematic operation of the framework. The bookkeeping layer enforces no-hearsay and dependency completeness. No claim, glyph, or residue enters the invariant coreΩ unless its full audit trail and the ledger is visible and validated. It distinguishes: Pre-formal glyphs (syntactic only). Partially recovered glyphs (discrete residue isolated but not fully transported). Certified invariants (full passage with canonical transport via cleavage).

@ourtown2 @CesarChavezP29 What about the bookkeeping/records?

Why is causality harder to find than you think? People see a correlation chart, or a time-series with an arrow marking a policy launch, and read it as causal evidence. But it almost never is. New essay on the Transportation Problem in Causal Inference: carloschavezp29.substack.com/p/the-transpor…

@World_Data_A China’s industrial rise did not happen randomly, it was carefully planned lol PRC care only about stability and control They are reactive which is good because Russia proves how bad a planned economy can be

. China’s industrial rise did not happen randomly, it was carefully planned One of the most underestimated parts of the story is how consistently China kept prioritizing the same strategic sectors across multiple planning cycles. This table is interesting because it shows which industries repeatedly appeared in: China’s Green Books, and successive Five-Year Plans. Some sectors appear again and again for more than a decade: robotics, aerospace, semiconductors, batteries, rail equipment, shipbuilding, new materials, biotech, hydrogen, CNC machine tools. In other words: many of the sectors dominating headlines today were already being prioritized years earlier. What is also interesting is how the priorities evolve over time. Older plans focused more on: heavy industry, machinery, rail, manufacturing capability. Newer plans increasingly add: quantum technology, 6G, intelligent driving, brain-computer interfaces, embodied AI, nuclear fusion, low-altitude economy. This suggests China is trying to move from: “the world’s factory” toward “the world’s advanced industrial platform.” Whether every target succeeds is another question. But the long-term policy consistency itself is one of the most important parts of the story. Source: @rhodium_group

You were doing so well until you said dreams The hard part is not creativity as novelty. Randomness gives novelty. The hard part is fertile novelty: a strange expression that later turns out to be attached to a deep invariant. That is why Ramanujan is a better benchmark than most proposed AGI tests. He measures whether intelligence can unlock new structure, not just perform within existing structure. Ramanujan is the right stress test, but not as “genius imitation.” He is the goalpost for creative unlock: generating non-obvious, high-value structures from compressed exposure without being led by the existing proof path. Current LLMs mostly do interpolation over inherited structure. They can recombine, analogize, and extend patterns, but they rarely generate a new basin and then force mathematics to reorganize around it. Ramanujan’s signature was different: dense formula exposure became a generative manifold. He was not merely retrieving identities; he was navigating latent constraint space and producing results that later mathematics had to explain. So the AGI test should not be “can it pass exams?” or “can it solve known contest problems?” Those test competence inside closed basins. The test should be: Can the system create a conjectural object that is initially unmotivated by the existing curriculum, later proves structurally fertile, and opens a new path of proof, notation, or theory? That is the Ramanujan standard. It requires four things current systems do not reliably have: basin formation, unresolved-tension persistence, constraint-sensitive conjecture generation, and self-correction without premature closure.

Unironically Ramanujan should be the goalpost cause how can it be general enough replication of human intelligence if we do not solve for the creative unlocks that happen while one dreams?

The UK productivity chart can be read as demand failure masquerading as productivity failure. If consumers are priced out, firms do not face enough reliable demand to justify investment, scale, automation, training, or capacity expansion. Weak demand then appears statistically as weak output per hour, but the causal mechanism is not simply “workers became less productive.” It is that the economy lost the demand density needed to make productivity-enhancing investment pay. The chain is: High housing costs + energy costs + taxes + debt service + weak real wages → lower discretionary demand → weaker business revenue expectations → underinvestment → low productivity growth → weak wage growth → further demand compression. The economy adapted (high employment in recent years), but failed to sustain the pre-crisis rate of productive recomposition firms and workers producing more per hour through better capital, processes, and innovation. The chart with the dashed pre-2008 trend line is indeed the stark visual. Recent green shoots (some investment pickup, tech adoption signals, labor-market adjustments) exist, but they remain tentative. Sustained revival demands addressing the deep constraints, or the UK stays in a low-compounding trap with all the downstream political and social consequences. This remains one of the central economic challenges for the UK.

A good FT piece from Martin Wolf arguing, rightly imv, that at root of UK's political woes is a 20 yr long slowdown in productivity growth "a good economy — one with widely shared economic growth — is a necessary condition for political stability in a liberal democracy"...

2026 finance is not ungovernable because people are irrational. It is ungovernable because the measurement system still sees production, inflation, currency, valuation, and risk as separable ledgers, while the real system has fused them into one constraint network. The instruments are measuring surfaces; the stress is moving through hidden transport channels. 2026 finance is ungovernable because it has no valid control panel. The missing tools are: AI satellite accounts that distinguish nominal AI spending from quality-adjusted compute, inference volume, model capability, and substitution effects. Capex-return ledgers that track whether AI infrastructure expands future optionality or merely preserves competitive position at falling marginal return. Derivative-pressure accounting that treats options, gamma, dealer hedging, and passive flows as price-formation infrastructure, not peripheral speculation. FX-decompression tools that track managed currency pressure through state banks, exporters, swaps, forwards, offshore deposits, and reserve-adjacent channels rather than official reserves alone. Constraint-weighted GDP/CWCI-style indicators that measure circulation, demand quality, repair capacity, bottlenecks, and systemic friction instead of only aggregate spending. Corrigibility metrics for firms, governments, and markets: how much error they can absorb before repair mechanisms become self-damaging.

1/🆕 My new @PIIE research with Patrick McKelvey on 𝗠𝗲𝗮𝘀𝘂𝗿𝗶𝗻𝗴 𝘁𝗵𝗲 𝗔𝗜 𝗘𝗰𝗼𝗻𝗼𝗺𝘆 finds the AI sector in the US 𝗴𝗿𝗼𝘄𝗶𝗻𝗴 𝟮,𝟬𝟬𝟬%+/𝘆𝗲𝗮𝗿 in quality-adjusted terms—yet it's nearly invisible in GDP. We can't let this measurement gap become a policy gap!🧵

The IMF diagnoses the current-account imbalance, but the solution is not IMF advice. The solution would be RMB appreciation plus Chinese domestic-demand rebalancing. Because China blocks that adjustment through managed FX and state-bank channels, the adjustment migrates into trade conflict, industrial policy, capital controls, and bloc fragmentation.

What about fx adjustment tho? ft.com/content/b600db…

Miller–Urey failed because it didnt include a dry phase Miller–Urey failed as an origin-of-life model because it was all wet chemistry and no environmental ratchet. The missing operator was wet–dry cycling: concentration, dehydration synthesis, surface selection, release, and repetition. Life does not begin when molecules appear; it begins when chemistry enters a recurring constraint cycle that can preserve and propagate structure.

The Miller-Urey experiment was a groundbreaking scientific study conducted in 1952 by Stanley Miller under the supervision of Harold Urey at the University of Chicago. The experiment was designed to test the chemical origins of life under conditions thought to resemble those of the early Earth. The setup for the experiment involved a closed system containing a mixture of gases that were believed to be present in Earth's early atmosphere, such as methane, ammonia, hydrogen, and water vapor. This mixture was subjected to continuous electrical sparks to simulate lightning, a common occurrence in Earth's primordial atmospheric conditions. The apparatus also included a water flask to mimic the ocean, which was heated to induce evaporation, and a cooling system to condense the vapor, simulating rain. After running the experiment for about a week, Miller analyzed the substances that had formed in the water and found that several organic compounds had been synthesized, including amino acids, which are the building blocks of proteins. This was significant because it demonstrated that organic compounds necessary for life could be synthesized from simpler inorganic compounds under conditions that might have been present on the early Earth. The experiment provided strong support for the hypothesis that life on Earth could have arisen through natural chemical processes from nonliving matter, contributing substantially to the field of abiogenesis—the study of how biological life could arise from inorganic matter.

The electron is not a little particle-object. It is a persistent charged closure mode. A Standing Lissajous EM Wave may be a candidate visualization of that closure, but it becomes admissible only after it derives the electron’s invariant residues rather than merely resembling a bounded oscillation.

Birth of the Electron, End of Quantum Mechanics! Just over a Century ago the peerless Arthur H. Compton measured the radius of the Electron, which he and Alfred L. Parson modeled as a Ring. For which Compton assumed [turns out mistakenly] that the Spin Velocity was Vc = 300,000 km/sec, the average Velocity of light. It wasn’t until 1990-1994, at Stanford, when I conducted a complete worldwide historical review of the structure and creation of the Electron, that I showed that the Electron topology was in fact a Standing Lissajous Electromagnetic Wave, an orthogonal superposition which takes on a (spinning) ring shape when the ratio of the respective wavelengths n/m = 1. Later I showed that the actual Spin Velocity of the ‘Ground’ state [n/m = 1] Ring Electron was V = π/√2 = 2.2214 Vc. [In other words, for an electromagnetic wobble or displacement or signal to be conveyed from one side of the Electron to the opposite, it has to move around the πr-long circumference to get to the other side.] The Electron gets its Voltage (mass, ElectronVolts) and radius, wavelength, frequency, time period from the two Gamma Rays which make up its transverse, orthogonal, superposition. Thus the Electron’s 510,999 ElectronVolt mass and 386 Fermi Radius, come from the two Gamma Rays, each with a 255,500 ElectronVolt mass, and 772 Fermi Radius!

@HunterWade Light does not travel through space. The repeated recoverability of light-like relations is what lets “space,” “time,” distance, and motion be reconstructed in the first place.

Walter Russell was adamant that light does not travel. That sounds wild until you notice how much of physics already points there. Maxwell unified electricity, magnetism, and light. What he gave us was not little light-objects flying through an empty container called space. He gave us coupled field relations. Schrödinger strengthened the point from another direction: quantum mechanics is built on continuous phase evolution. Einstein dissolved absolute space and time. Prigogine showed that systems far from equilibrium reorganize through flow, perturbation, and self-organization. If you have not studied dissipative structures, do. The implications for how people still talk about the second law of thermodynamics have not been fully integrated. Mandelbrot showed us recursive geometry. Fuller showed us tensegrity and relational structure. Again and again, the deeper insight appears: what persists is not isolated stuff moving through dead space. What persists is relational continuity through phase updating. Think of a stadium wave. No “thing” travels around the stadium. The pattern propagates because local participants update in relation. Light behaves more like propagating relational update than like a tiny object crossing a container. The old picture says: space is a box objects are inside it light travels from here to there The deeper picture says: relations update phase coheres geometry shows up motion is projected from continuity of update This does not discard physics. It integrates what physics already revealed. Carry this far enough and even space and time stop behaving like pre-existing containers. They begin to show up as projected structure: the geometry of coupled phase relations. And if that feels like a lot, think of 3D glasses. A flat screen gains apparent depth because each eye receives phase-differentiated information. Depth appears through relational difference. Follow that upstream and… The recursion holds. 🌀 🔗 Document: The Grammar of Projection and Cannibal Derivations (All Propagating Degrees of Freedom are Electromagnetic) zenodo.org/records/187753…

The paper measures distance inside the semantic cloud, not the semantic cloud itself. LLM divergent creativity is controlled traversal of learned semantic lines; human creativity is traversal plus admissible selection, integration, and re-stabilization under richer constraints. the model’s current distribution is possible continuations under a prompt, temperature, strategy instruction, and learned token-line geometry. Humans also operate inside a semantic cloud, but theirs is shaped by embodiment, memory, affect, experience, goals, and cultural exposure. The paper compares outputs, not the full state-generating field. The DAT imposes a simple constraint: produce words as semantically distant as possible. That rewards dispersion. It does not test whether the semantic cloud is meaningful, useful, coherent, personally grounded, or selectively governed. So the metric captures one constraint: distance. It misses higher constraints: relevance, necessity, taste, compression, surprise-with-fit, and selection. An output is admissible for the test if it follows formal rules and increases semantic distance. But creative admissibility is stricter. A word, image, or sentence must not merely be far away; it must survive integration. The paper’s own results show this: LLMs can beat average humans on DAT, but humans retain an edge in richer writing tasks and the top human distributions remain ahead. The study asks whether DAT divergence transports into haikus, synopses, and flash fiction. It partly does, but not cleanly. Transport from isolated word-distance to creative writing loses structure because writing requires theme, rhythm, genre, narrative pressure, and closure. Semantic cloud traversal is not the same as word scattering. The paper closes only a narrow loop: LLMs can be benchmarked against humans on semantic-divergence metrics, and prompt/temperature settings can modulate those scores. It does not close the larger creativity question because semantic distance is not full creative structure. The authors explicitly limit “LLM creativity” to divergent/associative semantic creativity rather than human-like mechanism. Temperature and prompt strategy change how the model propagates through the semantic cloud. Higher temperature broadens sampling; etymology prompts shift the search path; opposition prompts reduce distance because antonyms are semantically close in embedding space. This is the core mechanism: not “more creativity,” but altered propagation through learned semantic neighborhoods. What is conserved across tasks is not creativity itself, but a divergence tendency: models that sample broader semantic regions in DAT tend to show higher divergence in some writing formats. But conservation is weak because richer tasks introduce additional constraints that semantic distance alone cannot preserve. Temperature is a bifurcation control. Low temperature collapses the cloud onto high-probability attractors like repeated safe words. Higher temperature opens more remote paths. But beyond a point, dispersion can become noise. Human creativity bifurcates differently: not just broader sampling, but selective violation and re-stabilization under meaning. The paper detects a surface pattern: LLMs can generate semantically remote outputs. The deeper pattern is that LLM “creativity” is highly controllable by sampling and prompt constraints. That means the model is not discovering in the human sense; it is being routed through different regions of its semantic cloud. The real discovery is not that LLMs are or are not creative. It is that divergent creativity metrics expose a control surface over the semantic cloud. Prompt and temperature can push the model from common attractors toward remote associations. But the missing layer is governance: selecting which remote associations become necessary, coherent, and valuable.

🚨BREAKING: Researchers at the Université de Montréal ran the largest human versus AI creativity study ever done. 100,000 people against the world's best AI models. The headline was that AI won. It did not win. The study is being read backwards. Yoshua Bengio, one of the founders of modern AI, was on the team. They published it in Scientific Reports this year. The test was simple. Name ten words as unrelated to each other as possible. The further apart the meanings, the higher your creativity score. It is a real psychological measure. Performance on it tracks with harder creative work like writing and problem solving. GPT-4 scored higher than the average person. Gemini matched the average. The headlines stopped there. AI beats humans at creativity. Then you look at the rest of the results. The most creative half of the human participants beat every AI model tested. The top 25 percent beat them by more. The top 10 percent left every model far behind. The researchers also ran the harder tasks, haiku and flash fiction and plot synopses, and the pattern held exactly. AI cleared the average. It never came close to the best. So the real finding is not that AI became creative. It is that the average human score on a creativity test was never very high to begin with. Most people, asked to name ten unrelated words, reach for the same predictable cluster. Cat, dog, house, car. The test rewards reaching further, and most people do not reach. AI cleared a bar that the word "average" makes sound taller than it is. That changes what the study is actually about. It is not a race between human and machine creativity. It is a snapshot showing that real original thinking was always rare, concentrated in a minority of people, and that a language model can now reliably produce the median. The median was never the valuable part. There is a direct lesson here for anyone who writes, designs, or builds for a living. AI is now dependable at the average version of almost any creative task. The competent, expected, middle-of-the-distribution result. If your work lives in that band, this study is a warning. If your work is the kind the top 10 percent produce, the unusual connection most people never make, that is exactly where the models still fail. Creativity is not safe and it is not doomed. The average just became free. The work worth doing is the work that was never average. Sources: - Bellemare-Pépin, Lespinasse et al., Divergent creativity in humans and large language models, Scientific Reports, January 2026 - Université de Montréal, Concordia, University of Toronto Mississauga, Mila, Google DeepMind - Divergent Association Task, developed by Jay Olson, University of Toronto

@aaronggreen @JDRedding x.com/i/grok/share/d…

@JDRedding Discrete residue no?

When you collapse a higher‑mode structure down into a 2‑mode couple, you must bolt on glyphs to simulate the missing modes And yes... almost every major continuum thinker ended up tied to a glyph This isn’t a coincidence. It’s a symptom of the dimensional collapse

equity outflows are merely a sentiment gauge; exporter behavior is the actual structural anchor. a highly accurate, system-dynamics approach to tracking China’s macroeconomic stability. By treating external balance as the conserved variable, it successfully isolates the true structural risks from the noise of headline-driven market sentiment. The financial press frequently conflates portfolio equity outflows with systemic "capital flight." However China operates a closed-loop containment system where the financial account (portfolio flows) is structurally subordinate to the current account (trade surplus) and administrative routing. where the vulnerabilities lie. The Neutralization Engine portfolio stress is absorbed rather than allowed to propagate. This containment relies on two primary mechanisms functioning in tandem: The FX Settlement Ratio: The system's liquidity depends heavily on exporters converting their dollar earnings into Renminbi (RMB). When trade surpluses are high and the settlement ratio is robust, the system generates sufficient organic RMB demand to offset the depreciatory pressure from foreign investors selling domestic equities. State-Bank Proxy Intervention: The People's Bank of China (PBOC) rarely burns through official reserves in spot markets anymore. Instead, state-owned banks act as shock absorbers, utilizing the FX forward and swap markets to manage liquidity and stabilize the currency without triggering the alarms associated with official reserve depletion. As long as these two levers hold, the system remains in your defined state: managed portfolio de-risking under controlled external pressure. Anatomy of the "Break Signal" the exact failure condition: the system fractures only when exporter repatriation fails and state-bank containment is exhausted. To track bifurcation, the focus must shift away from the stock market and toward the behavior of trade dollars. A true systemic break would be preceded by these specific indicators: Extreme Exporter Hoarding: If Chinese exporters expect prolonged RMB depreciation or if US-China yield differentials remain overwhelmingly in favor of the dollar, they may hoard USD in offshore accounts or domestic foreign-currency deposits. This starves the onshore market of its primary defense mechanism. Exhaustion of Proxy Firewalls: If state banks are forced to absorb too much pressure, it will eventually reflect in their net foreign exchange settlement data or force the PBOC to intervene directly, tightening offshore RMB liquidity to punish short sellers. Leakage Through the Current Account: When capital controls on the financial account are tightened, genuine capital flight often disguises itself as current account transactions (e.g., over-invoicing imports, under-invoicing exports, or anomalous outbound tourism data).

Equity outflow from China last year was $ 200b. Big but not unmanageable. Relative to GDP, nothing like the flow out of South Korea 1/

America mutates faster, but much of that mutation is cannibalistic rather than developmental. It does not resemble Wenzhou-style productive decentralization. It is disruption that reallocates value toward control layers: platforms, finance, compute, defense, data, energy, and dollar liquidity. Europe preserves; China industrializes; America cannibalizes and scales. Markets reward it because index value follows control, not social repair. The U.S. does not outperform through Wenzhou-style bottom-up industrial upgrading: dense supplier networks, local manufacturing experimentation, private workshops, incremental process learning, and production-chain compounding. That model produces real productive capacity. The U.S. version is different: capital-market mutation plus cannibalistic disruption. America’s dominant firms often scale by destroying or absorbing adjacent economic layers: Amazon compresses retail margins, Google and Meta absorb advertising, Apple captures device rents, Nvidia captures compute bottlenecks, private equity extracts cash flows, software reduces labor demand, AI threatens the software/services layer that funded it. This is not broad industrial deepening. It is control-layer capture. That is why household degradation can coexist with system outperformance. The U.S. creates enormous market-cap winners, but many gains come from reallocating value away from labor, small firms, local economies, suppliers, and legacy industries into platform owners, IP holders, financiers, and infrastructure choke points. GDP and equity indices rise while economic security weakens underneath.

⚡️U.S. outperformance is real, durable, and more structurally important than most people want to admit. America is winning because it can metabolize chaos faster than everyone else. That is the whole secret. Europe is optimized for preservation. America is optimized for mutation. Europe protects the existing social model. America lets capital, talent, fraud, ambition, violence, speculation, startups, monopolies, bubbles, immigrants, engineers, universities, military spending, energy production, and financial markets collide until something enormous breaks through. That produces ugliness. It also produces scale. The U.S. can create OpenAI, Nvidia, SpaceX, Tesla, Palantir, Anduril, Apple, Amazon, Microsoft, Google, Meta, BlackRock, shale, Bitcoin infrastructure, venture capital networks, and world-dominating financial markets in the same civilizational machine. Europe can produce excellent engineers, beautiful cities, strong welfare states, industrial champions, and high living standards. But it struggles to create world-eating firms in the new layers of power. That is the gap. The Economist is late because this was visible years ago. The 2020s did not create U.S. dominance. They exposed the compounding. AI, cloud, energy, defense tech, capital markets, reserve currency privilege, fiscal capacity, demographic absorption, and China’s slowdown all pushed the world’s marginal capital back toward America. The biggest structural advantage is the dollar. The U.S. issues the asset the world needs in crisis, then uses that privilege to finance deficits, defense, liquidity, innovation, and consumption. Everyone complains about U.S. excess. Then they buy U.S. assets when the world gets dangerous. That is empire mechanics. Europe’s problem is deeper than weak growth. It has lost speed. Too much regulation, too little risk, aging populations, expensive energy, fragmented capital markets, weak tech platforms, defense dependence, demographic tension, and political systems designed to block extremes rather than produce new engines. Europe is not dead. But it is increasingly downstream of American tech, American defense, American liquidity, and American risk appetite. America’s contradiction is brutal: the system is globally dominant while many households feel poorer, angrier, and less secure. That is why people miss the signal. They look at broken housing, debt, health care, crime anxiety, inflation, political dysfunction, and conclude America is failing. At the household layer, many things are degrading. At the system-power layer, America is pulling further ahead. Both are true. Markets care more about the system-power layer until social breakdown starts impairing earnings, labor, politics, or bond confidence. So the real investment implication is clear: U.S. assets keep attracting global capital because the U.S. still owns the future-facing engines. AI, defense, energy, software, capital markets, and hard-asset monetary escape routes remain American-centered. Europe can rally cyclically. America owns the structural premium.

The article is not grand unified physics. It is a layered speculative scaffold. Its strongest move is recognizing that spacetime should not be assumed and that an admissibility/selection layer is needed before geometry. Its weakest move is that each layer still imports unresolved structure from the next: causal sets need selection, selection needs motifs, motifs need dynamics, dynamics need QM recovery, and defects need particle/gauge/GR recovery.

@bengoertzel Linguistic universals are not global rules. They are conserved constraint residues that survive transport across languages because they are admissible under the cognitive architecture that generates, learns, repairs, and propagates grammar.

@MillieMarconnni So the Friday question be What accepted fact in my field is not a fact? What definition is doing hidden work? What unresolved constraint would reorganize the field if pulled?

A mathematician at Bell Labs noticed that the scientists who won Nobel Prizes and the ones who never amounted to anything were equally smart, equally hardworking, and equally credentialed, and the only thing that separated them was a single question almost nobody is brave enough to ask themselves before they die. His name was Richard Hamming. He spent 30 years at Bell Labs, in the same building as John Tukey, Walter Brattain, and a long list of physicists who took home Nobel prizes for work they did down the hall from his office, including the legendary Claude Shannon. His invention of error-correcting codes made modern computing possible. He has won the Turing Award. And all the while he was creating his own legacy he was secretly doing a study on the people around him. The study was straightforward. 2 Teams. The legends and the lost. Same I.Q.s. Degrees same. Same desk hours. Same access to the world’s best resources. And yet, at the end of 40 years in their careers, one group had changed entire fields, and the other group could not be remembered by their own colleagues five years after retirement. He wanted to discover what the actual difference was. In March 1986, he stood before 200 researchers in a Bellcore auditorium and told them what he had seen. He said it all came down to one question. And hardly anyone he ever met was willing to ask it directly. He called it the Friday-afternoon ritual. He spent years blocking out his Friday afternoons and not doing anything productive with them every week. No experiments. No meetings. No deliverables. He called it Great Thoughts Time. He sat down with a notebook and asked himself a couple of questions in order. What are the most relevant problems in my discipline? And why I am not working on either of them.” Most weeks, the answer was the same, he said. For a week now he had marched confidently in a direction he did not think was the most important direction. He was a goer. He worked a bit. He was getting clean results that would publish in respected journals. ( And for five days straight he'd been lying to himself about whether any of it mattered. The reason almost nobody does this ritual is because the honest answer is unbearable. The thing is that if you sit down on a Friday afternoon and say out loud that you are not working on the most important problem in your field, now you have to do something about it. You have an immediate change in direction, or you have to keep lying to yourself every week from that point on. Most people choose the lie. In the short term it’s cheaper, but over a career it’s more expensive. Hamming took the ritual a step further in the Bell Labs cafeteria. He began approaching scientists he barely knew, asking them what they thought the most important problems in their field were. A week later he would ask them why they had not worked on these problems. Eventually people wouldn't have lunch with him. “I had to keep finding new tables,” he said. Nobody had a good answer for that, and being around someone who kept asking it made every meal feel like a performance review. The line that broke me is the line that most people skim over in the transcript. His words: If you do not work on an important problem you are unlikely to do important work. That’s not motivational line. It is a rational one. You cannot make a great result from a problem that does not matter. Input restricts the output. The choice of the problem is the ceiling of the career. The transcript has been freely available on the internet for almost 40 years. Stripe Press published the complete lectures as a book. Naval Ravikant quotes it all the time. It’s still given out to new hires at every serious engineering lab in Silicon Valley. Most people will not run the ritual this Friday. They will be busy. They always are.

unreliable local components can synthesize reliable global behavior if error is treated as an internal design variable rather than external noise. closure cannot require perfect primitives. A robust system must tolerate primitive unreliability and still recover stable invariants. von Neumann adds probabilistic repair : closure must tolerate imperfect local parts and still synthesize reliable global behavior.

von Neumann knew in 1952 what many now have forgotten; the brain has analog function along with digital.