Two ideas under a quiet, Apple-shaped tape, each bigger than its headline: AI has quietly become an inflation engine for the physical economy, and it is the one price pressure the Fed cannot see coming; and Google just proved you cannot capitalize a moat that walks out of the building, paying $2.7 billion to rent a researcher who was gone within a year. The news is the evidence. The ideas are the point.
Overnight, Asia turned the breadth problem global: Korea's KOSPI tripped its circuit breaker and closed down 5.8 percent in an AI-cost selloff, the Nikkei fell 4.2 percent, and US futures followed lower. It lands on a market already thin, a flat S&P hiding a 6 percent Apple-shaped hole with no breadth to spare. Bitcoin near 59,300 with DeFi deposits rising for the first time in three weeks has stopped selling but found nothing to buy. Gold's round trip above 4,000 in under 48 hours says the break was forced sellers, not a vanished bid. A 10-year pinned at 4.40 through the hottest core PCE since 2023 trades the Fed's response, not the print.
Three of today's stories are one mechanism wearing three faces. The Fed's preferred inflation gauge reaccelerated, core PCE hitting 3.4 percent, the highest since October 2023, and the rate market now puts a September hike at 73 percent, up from 29 percent a week ago. Underneath that print, a second inflation channel opened that the data will not capture for months: memory chip prices are surging, consumer DRAM up roughly 50 percent in a quarter and NAND flash about 90 percent. And the first company to pass it through did so today, Apple raising prices on every Mac and iPad it sells and taking its worst session in over a year for the admission. The bigger idea is that these are not three stories. They are one. AI's appetite for high-bandwidth memory, where margins run three to five times higher, is crowding consumer-grade chips out of the same factory queue, and that scarcity is now arriving at the checkout counter. Consensus holds that AI is deflationary, that cheap cognition pushes prices down everywhere. The first hard print says the opposite: AI is bidding away the manufacturing capacity that used to make physical goods cheaper every year, and the bill lands on a laptop. The forward read is that this channel is invisible to the gauge that sets policy, because PCE reads finished durable goods months after the component cost moves, so the Fed is hiking against the inflation it can measure while a second, AI-driven one builds underneath unseen. What would change my mind is whether this is structural or just the memory cycle doing what it always does. DRAM and NAND spike and crash on their own clock, and if fabs add capacity and prices roll over within two quarters, Apple walks the hikes back and this was an ordinary upcycle in an AI costume. The tell that it is real: consumer memory stays bid and the pass-through widens past Apple, more device makers raising prices on the same cause, even as overall demand softens. Structural scarcity raises prices into weak demand. A cycle does not.
Four senior researchers walked out of Google DeepMind in a single week, and three went to the same place. John Jumper, the Nobel laureate who led AlphaFold, took two colleagues to Anthropic; Noam Shazeer, a co-author of the Transformer whom Google bought back for 2.7 billion dollars less than a year ago through the Character.AI deal, left for OpenAI. Alphabet fell about 6 percent. The number to sit with is that 2.7 billion: Google paid it to bring one man home, and he was gone within months. The bigger idea is that you cannot capitalize a moat that has legs. When the breakthrough lives in a handful of heads, owning the lab is renting the talent, and the acquihire is the most expensive lease ever written, because the asset can quit. The dated precedent is Xerox PARC, which invented the graphical interface, the mouse, and Ethernet, then watched the people and the ideas walk to Apple and Microsoft while its own position eroded for two decades. The institution that assembles a breakthrough rarely captures its value, because what it owns is mobile and the building is not. That forces a binary onto Alphabet the market now has to price. If its moat is model capability, that moat just walked out the door three times in one week. If its moat is distribution, the Search box and the Cloud that put Gemini in front of billions regardless of who builds it, the departures are noise and the 2.7 billion was always a vanity overpay. Those are two very different companies trading under one ticker. What would change my mind is which one the next model cycle reveals. It is signal if Gemini visibly slips against the rivals that just hired its people and Alphabet's AI premium compresses toward a distribution multiple. It is noise if the next flagship holds the frontier anyway, proof the moat was the data and the infrastructure all along, and the people, however brilliant, were replaceable.
For five days the Iran story was a headline that kept almost dying. Today it turned real: Treasury issued the 60-day waiver and Iran was taking bids from Asian refiners within hours, the tankers staged and waiting. This is a supply story with a delivery date now, not diplomacy, and the 2015 template says a premium that leaves rarely comes home. The risk that snaps it back: a Senate blocking vote or an inspection breakdown before the late-August expiry.
“Take care of your own soul and another person's body, not of your own body and another person's soul.”
You spend more energy than you realize managing other people's attitudes. The colleague who needs to "get over" a disappointment. The partner whose priorities you want to rearrange. The friend who would be happier if they could just see what you see. You put your insight into their emotional architecture, certain the intervention is generous, and meanwhile your own inner landscape goes untended. The reversal is just as true: you are meticulous about your own comfort, your routine, the precise calibration of your physical environment, and remarkably vague about whether the people around you have what they need to get through the day.
The Kotzker's instruction is not gentle: your jurisdiction over another person's inner life is zero. Not small. Zero. But their hunger, their exhaustion, their practical need for a hand with something unglamorous, those are yours to notice and yours to act on. And the reverse: your own comfort is not your project. Your own soul, the part of you that keeps getting postponed because the body's requests are louder and more specific, that is the project. You know the conversation you keep putting off, the one about what you actually want from this year. You know the question you stopped asking because the answer might require you to change something comfortable.
Today's practice: The next time you catch yourself about to give someone unsolicited advice about their feelings, stop, and instead do one concrete thing for their day or their body: make the meal, send the ride, take the tedious task off their hands. Then turn the same attention on yourself: say out loud, to one person, the question about your own life you have been postponing, and put a thirty-minute block on the calendar this week to actually answer it. Tend their body and your own soul; leave their soul and your own comfort alone.
In 1954 the fishermen of Alanya, on Turkey's coast, raced each other to the best spots until catches collapsed. They rejected both textbook fixes, privatize the water or call in the government, and built their own: each September every fisherman drew lots for a starting position, then shifted one spot east each day, so everyone fished every spot and cheating was visible to the boats on either side. Elinor Ostrom found the same structure in commons that lasted centuries, from Swiss meadows to Maine lobster fisheries: clear boundaries, locally-set rules, graduated sanctions, and above all monitoring. The reason commons fail is not the selfishness the textbook predicts. They fail through enforcement, and enforcement collapses the moment the first defector goes uncaught and everyone quietly adjusts to the rules being optional.