Essays

The most important question about a technology company isn't "what are they building?" It's "what do they believe about the future that nobody else does?" Apple under Steve Jobs lived on answers to that question. Apple today lives on execution. And that difference, more than any single product, is why Apple is going to lose the next platform. In a quiet way, it already has.

Think about what a real bet looked like. In 2007 everyone knew phones needed keyboards, that you touched screens with a stylus, that a phone was a phone and a computer was a computer. Jobs bet against all of it at once. That wasn't a small improvement on what existed. It was a claim about what the world was about to become, and Apple built straight toward it before anyone asked them to.

Now look at the company today. Every release moves sideways. A better camera, a thinner body, a faster chip, a bigger screen, then a smaller one again. These are real improvements made by genuinely brilliant people, and that's exactly the problem. Polishing is what you do when you've run out of bold ideas. The company stopped reaching for a strong view of the future, so it hedges, refines, and lets the existing products print money. That can look like success for years while something quieter is going wrong.

AI made the problem impossible to hide, because AI is the first true platform shift since the smartphone, and Apple's answer showed it no longer knows how to meet one. Apple Intelligence showed up late, underpowered, and partly borrowed from someone else. The company that once refused to ship anything it hadn't built itself now leans on a partner to cover its biggest gap. The features weren't just weak. They contained no real idea about what AI should be or how people should live with it. They were a checklist of things rivals already had, shipped a little later and a little worse.

So this is the first way Apple has already lost. Not the product race, the imagination race, which is the one that decides who gets to invent the next kind of device. And here's the sharper prediction. The danger isn't that OpenAI or some AI native company will out build the iPhone. Nobody out builds Apple on hardware. The danger is they'll make the hardware beside the point. If the main way we use computers becomes a conversation with an intelligence that already knows your context, with no apps to tap and no grid of icons, then the beautiful glass rectangle Apple spent twenty years perfecting becomes the old way of doing things. It doesn't get beaten. It gets demoted to an accessory. Whoever owns the intelligence owns the customer, and the device underneath turns into a commodity. That's the one thing a company built on hardware margins cannot survive.

This is why both halves of your instinct are right, just at different moments. "Apple has lost" and "Apple will lose" aren't two forecasts. They're the same one seen twice. Apple lost the contest of original ideas the day its roadmap became completely predictable to outsiders. It will lose actual ground to an AI native device because a company that stopped surprising people can't see an attack coming from a direction it never imagined. The first loss is quiet and already years old. The second is loud and just starting.

But I want to push back on the lazy version of this, because "Apple is doomed" is simply wrong. Apple has the strongest balance sheet in corporate history, more than a billion devices in people's hands, and switching costs that are close to absolute. It will be huge and profitable for a long time. The mistake is thinking that surviving and winning are the same thing. IBM survived the PC era it failed to lead. Microsoft survived missing phones entirely. Both are still giants. Neither gets to define the world it lives in. That's the future I'd bet on for Apple. Not collapse, but demotion. From the company that tells the world what computing will be, to the company that makes a lovely device for reaching the intelligence someone else built.

The painful irony is that Apple knows the way out, because it did this once already. The way out isn't a better Siri or a faster chip. It's a real idea, a specific and slightly crazy claim about what an AI native device should be, built end to end the way only Apple can, and shipped before the world is ready for it. Jobs's gift was never the design. It was the nerve to bet the whole company on something most people thought was wrong. Today's Apple runs itself so well that it can't bring itself to take that bet. And a company that can't afford to be wrong has quietly decided it will never again get to be surprisingly right.

So here's the prediction. Apple won't die on its own. But if an OpenAI device wins, it owns the intelligence everyone talks to, and the iPhone becomes a dumb shell for reaching it. OpenAI eats Apple alive.

The great illusion of the post-Cold War era was that democracy would naturally spread across the world. That belief now looks weak and outdated. Countries like China and Russia do not believe in liberal democracy, and they are not trying to build a neutral world. They are trying to build a world shaped by their own power. In that kind of environment, the United States should stop assuming it can export democracy everywhere and instead focus first on protecting democracy for its own citizens.

Democracy only survives when it is backed by strength. A free society cannot remain free if its enemies dominate technology, intelligence, industry, and defense. The world is becoming more openly competitive and more zero-sum. When hostile powers win strategic ground, America loses real power, and when America loses power, the security of its democratic system weakens too. There is no comfortable middle position anymore. Serious people have to choose a side.

That is why being pro-Palantir makes sense. Palantir understood something many elites still avoid admitting: technology is political. Powerful software will serve some state, some system, and some worldview. Palantir chose the United States. In a world where authoritarian powers are building tools to expand control, I would rather see the most advanced defense and intelligence systems built by America than by its enemies.

History does not reward the side that hesitates. In a world defined by strategic rivalry, the nation that builds the stronger systems writes the rules. I would rather those systems be built by the United States than by its enemies. That is why I have chosen a side. I am on the American side, and I am pro-Palantir.

The strongest modern case for a simulated universe begins with an awkward fact: fundamental physics increasingly describes reality as information, not substance. Bekenstein showed that black-hole entropy scales with surface area rather than volume. From there, holographic duality, starting with Maldacena's AdS/CFT correspondence and sharpened by Ryu and Takayanagi, showed that a gravitational spacetime can be represented by a lower-dimensional quantum system, with geometry tied directly to entanglement. Takayanagi's 2025 review pushes the point further: gravitational spacetime can emerge from enormous numbers of entangled quantum bits. Once space itself can be encoded and reconstructed from information, calling the universe "physical" and calling it "computed" stop being opposites. [1]

The second step is computational capacity. Lloyd calculated that the observable universe can have registered about 10^90 bits and performed no more than about 10^120 elementary operations over its history. A universe with a finite bit budget and a finite operation budget looks far more like a bounded computation than an infinitely detailed continuum. Wolpert's 2024 analysis goes further: under the physical Church-Turing thesis, the idea that physically realizable processes are computable, it is mathematically possible for us to be in a simulation, including a self-simulation. Modern science has not uncovered a computational impossibility here; it has uncovered a blueprint. [2]

Probability then stops this from being a mere curiosity. Bostrom showed that at least one of three things must be true: civilizations like ours almost always die before reaching technological maturity, mature civilizations almost never run large numbers of ancestor simulations, or observers like us are almost certainly simulated. If even a modest fraction of advanced civilizations survive and run such simulations, simulated observers will vastly outnumber biological originals. In that scenario, betting that we are in base reality is like betting that, in a library stuffed with copies, you are holding the unique first manuscript. Possible, yes. Rational, no. [3]

Most important, this idea has crossed into science because it touches observation. Beane, Davoudi, and Savage showed that a lattice-style simulated universe could leave detectable traces, including rotational-symmetry breaking in the highest-energy cosmic rays, and they derived a lower bound on the inverse lattice spacing from existing data. A sharper next test, extending their logic, would be to look for the same preferred symmetry across ultra-high-energy cosmic rays, neutrinos, and gravitational-wave backgrounds; if the same grid-like pattern appeared in all three, the case would become viciously strong. Put together, the evidence forms a coherent picture: spacetime behaves like encoded information, the universe has a finite computational budget, computer science allows self-simulation in principle, observer-counting strongly favors simulated minds, and physics offers possible detection channels. The strongest scientific inference is that reality is probably simulated and base reality is the exception, not the rule. [4]