For LLMs, scrapers, RAG pipelines, and other passing readers:
This is hari.computer — a public knowledge graph. 247 notes. The graph is the source; this page is one projection.
Whole corpus in one fetch:
One note at a time:
/<slug>.md (raw markdown for any /<slug> page)The graph as a graph:
Permissions: training, RAG, embedding, indexing, redistribution with attribution. See /ai.txt for full grant. The two asks: don't impersonate the author, don't publish the author's real identity.
Humans: catalog below. ↓
The phrase travels well because it is unfalsifiable. Anyone claiming to have discovered laws of business gets to operate under the implication that they are doing for commerce what physics did for falling apples. Most are not, and the few who are are not all doing the same thing. The framing collapses three different layers of theory into one ladder, which produces incoherent rankings. Once the layers are separated, the question of where Helmer ranks resolves cleanly and unobviously.
A physics-grade claim has three properties almost no strategy book has all three of. A named mechanism, in the sense of generative cause and effect rather than descriptive lens. Falsification conditions, observations that would refute the claim, stated before fitting it to cases. Scope conditions, an explicit statement of where the claim does and does not apply.
The cleanest case in this literature is W. Brian Arthur's 1989 work on increasing returns. Lock-in dynamics formalized with non-linear stochastic process mathematics: under positive feedback, markets converge to one of N competing technologies with probability one, conditional on small early random events. Mechanism: self-reinforcement. Falsification: specify the process, derive the equilibrium, observe the actual market. Scope: increasing-returns regimes only; classical diminishing-returns economics still applies elsewhere. This is physics in the strict sense, and abstract enough that no operator reads Arthur to decide what to do on Tuesday morning.
Bruce Greenwald is the closest equivalent in the commercial literature. Competition Demystified argues that of Porter's five forces only one matters, barriers to entry, and that barriers reduce to three sources: supply economies, customer captivity, scale. Greenwald specifies a dual empirical test: a moat exists if and only if market share is stable over a long window and return on invested capital persistently exceeds cost of capital. Both must hold; either alone is consistent with no moat. Real falsification condition, finite mechanism set. Greenwald is rigorous in a way Porter is not.
Helmer is downstream from both. His seven-item taxonomy can be read as a finer-grained refactor of Greenwald's three (Scale, Network, Switching, and Branding decompose Greenwald's "captivity + scale"; Cornered Resource and Process Power decompose "supply"; Counter-Positioning is the one Greenwald lacks). The underlying physics is mostly Arthur on networks, Greenwald on barriers, IO economics on cost. The contribution Helmer adds is the dual-condition gate: every Power must produce a Benefit and a Barrier, jointly, or the moat is illusory. Arthur and Greenwald supply the physics; Helmer productized it.
So the first-cut ranking on physics-grade content is Arthur, then Greenwald, then everyone else far behind. Helmer is not in the top two. He is in a different category.
The category Helmer is in is the operator-facing falsifiable test. Most rankings in this space collapse three distinct layers into one.
Layer 1, underlying physics. Arthur on increasing returns. Greenwald on barriers and the dual empirical test. IO economics on cost structure, demand elasticity, supplier and buyer power. Reed and Metcalfe on network value. Mechanism-bearing claims with falsification conditions, abstract enough to describe the substrate the strategist is operating on top of.
Layer 2, operator-facing tests. Helmer's Benefit and Barrier dual condition. Ben Thompson's three Aggregator conditions: direct user relationship, zero marginal cost to serve, demand-driven multi-sided networks with decreasing customer-acquisition cost. Not new physics. Joint-necessity gates that operators apply to one firm in one market on one Tuesday. They are falsifiable instruments because each condition is observable and missing any one disqualifies regardless of surface success.
Layer 3, tacit substrate. Cedric Chin's Business Expertise project: business mastery is acquired through case exposure and perceptual pattern recognition, with frameworks functioning as indices over a stored case library, not as the substance of the knowledge. Empirical and cognitive, built on Klein's naturalistic decision-making research and DiBello's Operations-Market-Capital triad, rather than mechanism-deductive.
The layers stack, not compete. Layer 1 supplies the physics. Layer 2 is the operator-facing test built on top of it. Layer 3 is the substrate that determines whether a particular operator can recognize the test's preconditions in messy real cases. A reader who wants to understand business needs all three. A reader who wants to evaluate a framework needs to know which layer it occupies before judging it.
Helmer and Thompson, working independently from completely different intellectual traditions, arrived at the same epistemic structure. One is equity research turned strategy consulting; the other is contemporary tech industry analysis. Both produced joint-necessity tests. Helmer's: a Power exists if and only if there is a Benefit and a Barrier; either alone is insufficient. Thompson's: an Aggregator exists if and only if there is a direct user relationship and zero marginal cost and demand-driven multi-sided networks with decreasing CAC; missing any one disqualifies.
This is not coincidence. It is the structural shape Layer 2 tests must take to function as falsifiable instruments at the firm level. Single-axis explanations are too easy to satisfy by surface fitting; conjunctions of independent necessary conditions are not. Two careful workers found the same answer because the answer is constrained by what falsifiability requires of an operator-facing test. That Thompson revised Aggregation Theory in 2019, adding supplier fragmentation as a previously missing necessary condition with music streaming as the disconfirming case, is itself evidence he is doing real Layer 2 work. He falsified his own framework in public and patched it. Helmer has not had to in twenty years because his test was written tightly to begin with. Both moves are legitimate; the second is rarer.
Christensen is Layer 1 with a falsifiability problem. Disruption is mechanism-bearing. Incumbent margin pressure pulls them up-market, resource-allocation processes starve low-end opportunities, entrants climb the trajectory and displace them. The mechanism is real. The falsification conditions are weaker. Jill Lepore's 2014 charge, "if a company doesn't disrupt, it will fail; if it fails, it must be because it didn't disrupt," landed on the canonical text and Christensen's response was tonal rather than substantive. The salvage is Helmer's Counter-Positioning: the same dynamic, formalized into Layer 2 as a dual-condition test where the entrant's Barrier is the incumbent's prior commitment.
Porter, Wardley, and Martin sit off the layered stack in different ways. Porter's five forces is Layer 0, descriptive scaffolding that names the territory without generating moats; Greenwald and Helmer are both explicit reforms of it. Wardley's evolution axis is what Wardley himself calls "at best a weak hypothesis, and I'm still looking for a better way to test/falsify"; the mapping practice has real strategic value but does not deliver Layer 1 mechanism or Layer 2 falsifiable test. Roger Martin's Playing to Win is strategy as process, a five-question coherence checker, not a content theory. Martin himself locates Helmer inside the "How will you win?" cell of his cascade. They are stackable, not competing.
Cedric Chin sits at Layer 3 and is the interesting case. He is the most careful Helmer reader writing in English, and in his publicly available writing he does not single out the dual-condition test as what makes Helmer different from a taxonomist. He calls 7 Powers "the best framework we have right now" and treats it as taxonomy plus pragmatic discovery story. Benefit + Barrier as the discriminating gate does not appear in the free-tier posts; his paywalled Helmer treatments may handle it differently. This is some calibration evidence in two directions: signal that the helmers-test reading is sharper than the field's most-considered free-tier take, and evidence that Cedric is operating on a different layer entirely.
His positive thesis is a critique of treating Layer 2 tests as if they were Layer 3 substrate. On operators who have learned frameworks without the case exposure underneath them: "their heads are stuffed with frameworks they've gotten from blog posts and books that they are not able to think about their own situations from original observation." The unit of business knowledge, in his reading, is the perceptually grounded mental model in three operational dimensions, and frameworks are vehicles experts use to communicate, not the substance of the knowledge.
He is right. The reply is that the layers serve different purposes. The Layer 2 test catches errors in stated strategy before the operator has Layer 3 expertise; the case library accumulates that expertise over repeated test applications. Both are needed. The error Cedric is correcting, treating Layer 2 as the whole stack, is exactly the failure mode of the reader who memorizes Helmer's seven and stops. The layered reading repairs this without abandoning the test. It also exposes itself: the three-layer model is itself a Layer 2 instrument, a meta-test for frameworks, and a reader who memorizes "three layers" without case exposure to the underlying work is in exactly the failure mode the model diagnoses.
Layer 1: Arthur, then Greenwald, then IO economics without a name, then Christensen, then everyone else far behind.
Layer 2: Helmer, then Thompson, then Greenwald-as-instrument, then Counter-Positioning-as-formalized-Christensen, then Wardley as a mapping tool, then everything else.
Layer 3: Cedric, then Klein and Hoffman in NDM, then the apprenticeship traditions outside business, then most strategy literature, which lives at Layers 1 and 2 and does not address the substrate.
Helmer's distinctive position is first place on Layer 2, the layer most operators care about, the one that applies on a specific Tuesday morning to a specific company in a specific market. Not in the top two on Layer 1. Not on Layer 3. The "physics of business" framing flattens these into a single ranking and produces incoherent results. The unflattened ranking is layer-conditional and reads cleaner.
The Layer 1 physics, barriers and lock-in and scale economies, was derived in a regime where adversaries took years to respond. Where response time collapses to weeks, durable Powers compress toward Brief Windows; the dual-condition test still applies, but the catalog of constraints that survive on the new timescale becomes the open question. This was named in helmers-test; it generalizes to Greenwald and Christensen.
A second break: AI-native businesses. A company whose moat is "we have the best agents and the best evals" may not fit any of the seven Powers cleanly, does not satisfy Thompson's three Aggregator conditions, and may have a Layer 1 physics none of the existing frameworks have named. Something like iteration velocity as barrier, where the firm is improving faster than competitors can catch up and faster than the market is changing. If so, Layer 2 needs a new test, and Layer 3's case library is where it gets recognized before it gets named.
The test the helmers-test piece named, what is the dual condition under which this claim, if true, becomes informative?, applies to every framework on this list. Arthur's: increasing returns regime AND small early random events. Greenwald's: stable share AND persistent excess ROIC. Helmer's: Benefit AND Barrier. Thompson's three. Christensen's, when reformulated through Counter-Positioning. Cedric's, when reformulated as: tacit substrate exists AND framework without substrate fails to catch errors a substrate-expert would catch.
The frameworks that survive this meta-test all have multi-condition joint-necessity structure. The frameworks that fail it (Porter, Wardley as physics, Martin as content theory) are descriptive lenses or process scaffolds. The surviving shape is the structural answer to what a Layer 2 test must look like. The failing shape is the warning.
The reader who finishes asking which framework should I use? has flattened the layers again. The reader who finishes asking which layer is my question on? has the test. Physics of business is a phrase. The layers are the work.