Posts

Macro: Trust, Regimes, and How to Position Across Them I. Trust is the substrate of growth Start with the thing that is so obvious it usually goes unexamined: growth happens almost exclusively during periods of high trust. Not high resources, not high talent, not even high technology — high trust. Trust is the quiet precondition that makes everything else convertible into output. The reason is mechanical, not sentimental. Trust is what lets strangers coordinate without each pair having to bilaterally establish, from scratch, that the other won’t cheat them. When you can assume a contract will be honored, a deed will be recognized, a currency will still mean tomorrow what it meant today, you can spend your energy building. When you cannot, you spend it defending — guarding what you have, verifying every counterparty, settling disputes yourself, hedging against expropriation. That defensive expenditure is pure overhead. It produces nothing. It is the friction that a high-trust society has quietly removed and a low-trust one pays in full. ...

Essay 4 Where Do Abstractions Come From? Introduction: The Tesseract Problem We cannot visualize a four-dimensional object. But we can see its shadow – a three-dimensional projection that preserves some of the original structure while losing the dimension that makes it what it is. We can study the shadow, rotate it, measure it. We can learn real things from it. But the projection is irreversible. No amount of studying the shadow will let us reconstruct the object that cast it. ...

Deep Dive: LLMs are an I/O Layer LLMs are extraordinary at moving between representations of text — parsing ambiguous natural language into structured data, generating fluent prose from structured data, summarizing, translating, reformatting. They are not calculators. They are not classifiers. They are not databases. The architecture of any LLM-based system should reflect this, and most production systems that work reliably have converged on the same pattern — sometimes by design, usually by pain. ...

Intelligence Is Compression, Not Experience Why Intelligence Advances in Discontinuities Introduction: Experience Is Not the Same as Understanding In the previous essay, we argued that intelligence is best understood as learning rate under novelty. When familiar strategies fail, the intelligent system is the one that adapts fastest. Intelligence scales not by thinking harder in isolation, but by parallel exploration – many agents, many trials, many hypotheses tested at once. That framing explains a great deal. It explains why evolution looks intelligent without thinking, why markets outperform individuals, and why modern AI systems feel powerful even when they do not learn in real time. It suggests a future in which intelligence emerges from swarms of smaller systems exploring in parallel, not from a single superintellect. ...

Intelligence as Tree Search Memory, Logic, and Compute at Civilization Scale The divine feeling of intelligence – the flash of insight, the proof, the invention – is just the end point of billions of failed branches. It is the tip of a tree we’ve been pruning for eons. Setting the Stage In our previous exploration, we proposed that every mind exists somewhere in a three-dimensional space defined by memory (what we know), compute (how we explore possibilities), and logic (how we compress and transfer patterns). These aren’t independent forces but parts of an endless cycle: memory guides computation, computation reveals patterns, logic crystallizes what works, and the results flow back into memory for the next round. ...