Twelve bets I actually hold on where the money lands as AI spreads. The curve shows where each one sits right now: early, running hot, or slowing down.
every dot is one bet, and its spot answers one question: is this early, or is it slowing down? position comes from two things. the evidence, meaning what the daily graded news and numbers say about the idea. and the attention, meaning how hard the market is already chasing the names, read for now from how the shares run against their own slice of the market. strong evidence with little attention sits early on the curve, that is the interesting zone. lots of attention with thin evidence lands in crowded, the zone where late money shows up. the dot's color is what today's evidence says, green for it, red against it. when a bet changes stage, its dot slides along the curve that morning.
the chip is today's evidence. the number is each bet's one stock pick against its own slice of the market.
it will be wrong sometimes. that is what the receipts are for.
each row is one thesis: a bet, written down, about where the money lands as AI spreads. a thesis has three parts. the idea itself. what would prove it right. what would prove it wrong. all three sit on the thesis page, written before I put money behind it.
the rows are ordered by how warm today's evidence is, so the top of the list is where the world is currently agreeing with the bet.
the chip reads today's evidence only. a bet can be down in money while the evidence behind it reads green, and the other way around.
each thesis names one stock as the best way to own the idea right now. the number beside it compares that stock with the slice of market it belongs to, so beating the market is the bar, not simply going up with it. +2.0% means it is two points ahead of that slice, −2.0% means two points behind. just named means the stock was picked this morning and has not traded a full day yet.
the small squares are the last two weeks of daily reads, oldest on the left. the arrow shows whether the bet moved up or down this list since yesterday.
a short is a bet that a stock FALLS. bad news for that company shows up green here, and a falling price counts as the bet working.
One email a week: which bets heated up, which cooled off, which one drifted toward the line I said would kill it, and which idea is brand new. All 12 of them, published unedited.
the bets run on the same handful of forces. green wires reinforce, red wires pull against each other, bright wires fired today.
a green reinforce wire means two bets ride one force from different angles. a red tension wire means the same force taken from opposite sides. $TSLA, for example, sits on both a bet that the shares rise and a bet that they fall. a node's color is today's evidence and its size is how strong that read is. the ring around it is the last ten mornings, so a long green arc means the evidence kept going the bet's way for days. bright, fast wires fired today, meaning both ends moved at once. hover anything for the numbers.
every card is one bet: the companies, their numbers, the kill line, the evidence trail. on a short a falling price counts as green.
Intelligence becomes free; value flows to applications with proprietary data and workflows, not model labs
VLA models turn robots into software platforms; public markets underprice it because leaders are private
Enterprise AI is hard; value flows to integrators and AI-literate services while models commoditize
AI bottleneck shifted from compute to data movement; CPO/photonics wins whether the buildout continues or rationalizes
Deficits get inflated away; scarce assets reprice. Position built during the stress test, not the party
The Musk empire trades on narrative, not cash. Tesla's auto cash cow is cracking — Q2'26 operating profit fell 57%, free cash flow went negative, and regulatory credits (once ~2pts of margin) collapsed as the EV credit expired — yet it's priced at ~145x earnings; the private crown jewels are only reachable at steep premiums to NAV. The bet: the premium compresses toward the cash the cows actually produce.
Companies generating outsized revenue and profit relative to their market cap — high margins, real free cash flow, low multiples — are mispriced. The market treats durable profit machines like they're stagnant; as the cash keeps compounding, they re-rate upward.
The AI buildout's binding constraint is shifting from chips to electricity. Datacenters are signing decade-long power deals faster than anyone can build generation; firms that own or equip gigawatts capture pricing power for years
AI turned memory from a cyclical commodity into the binding constraint — HBM is sold out years ahead, conventional DRAM and NAND supply is being cannibalized to feed it, and pricing power has shifted to a three-player oligopoly for the first time in decades
Per-seat software loses its unit of account when agents do the work. Mid-market SaaS with seat pricing and thin data moats gets repriced first: seat counts stall, net revenue retention slips, and the multiple compresses. The bet: these names underperform the software index
Regulated stablecoins move real payments onto new dollar plumbing. The companies that issue the coins and run the rails collect a growing toll, while the card networks' cut gets squeezed
Europe's rearmament money is shifting from a few expensive platforms to huge numbers of cheap autonomous systems. The factories and software that produce that mass turn decade-long budgets into recurring orders
the next dates these companies open their books. that is when a bet stops being an argument.
one line per bet per weekday, published exactly as written.
Every weekday: share prices, and how each one did against its own slice of the market (the tech index, the chip index, the broad one). The company numbers behind them, meaning how fast sales grow, how much of a sale is left as profit, and how much spare cash the business throws off. Plus fresh headlines, searched one signal at a time, for the exact things each bet depends on.
Every weekday an AI reads that evidence and scores each signal from −2 (strongly against the bet) to +2 (strongly for it), with one line saying why. I publish the score and the line exactly as written, including the mornings it disagrees with me. That is the tape.
Before the money moved, I wrote down what would make me admit each bet is wrong, and what would make me lean in harder. Both lines sit on every thesis page in the words I first used. When the evidence runs against a bet for long enough, the machine flags it for a kill review on its own. No quiet moving of the goalposts.
A one-person experiment in showing your work. Nothing here is investment advice or a recommendation to buy or sell anything. Do your own thinking. That is the whole reason the data is sitting out here.