Field notes · No. 01 · September 2026
applequantum is an independent briefing on quantum computing — what the public record supports, which futures would actually matter, and which claims are still costume. No laboratory speaks here. No company roadmap is implied.
The record
01
Universities and companies have published machines that prepare, entangle, and measure quantum states. Noise still decides how long those states last, and how deep a circuit can run before the answer dissolves. A larger count of physical qubits is not, by itself, a more powerful computer.
02
In December 2024, Google Quantum AI reported in Nature that a surface-code experiment ran below the fault-tolerance threshold: as the code grew, the logical error fell. That is a real milestone. It is not yet a calculation anyone would trust with a molecule, or with a cipher.
03
On 13 August 2024, NIST published the first three post-quantum standards: ML-KEM, ML-DSA, and SLH-DSA. Peter Shor showed in 1994 that a large enough fault-tolerant quantum computer would break RSA and elliptic-curve cryptography. The migration is a standards project. It does not require the machine to exist first, because encrypted traffic can be stored now and read later.
Laboratories are named only where a public paper is. This page does not rank them, invest in them, or speak for them.
The interesting number is not how many qubits. It is whether the error falls when you add more.
A single qubit
A bit is a fact: 0 or 1. A qubit is a direction. North is |0⟩, south is |1⟩, and every other point is a superposition — both answers, with different weights, plus a phase that a single measurement will not show you. Drag it. Then measure.
Bloch sphere. Drag to rotate the qubit, or focus it and use the arrow keys. North is zero, south is one.
Drag the sphere, or focus it and use the arrow keys.
0.82 |0⟩ + (0.43 + 0.37i) |1⟩
Ideal state. A laboratory qubit is a noisy cousin of this picture.
θ 69° · φ 40° · Born rule, amplitude squared
Each mark is a fresh copy of this state. The sphere stays put so you can watch the odds gather.
The stakes
If fault tolerance scales, the effects are narrow and serious. They are not a new internet, and they are not a new phone.
01
Molecules are already quantum systems. The cost of simulating them on a classical machine grows brutally with the number of electrons. The honest target is a handful of reactions — catalysts, battery chemistries, the fixation of nitrogen — where today’s approximations disagree with each other. Not every drug. A few mechanisms that matter.
02
The same mathematics covers magnets, superconductors, and awkward interfaces. If this arrives, it will look like a better instrument in a laboratory, used by people who already run difficult simulations. Not like a glowing chip in a pocket.
03
This is the stake that does not depend on a breakthrough you can wait out. Data protected by RSA or elliptic curves can be recorded today. Replacement standards already exist. The hard part is inventory: knowing which locks are fitted to which doors, in which old systems, under which vendors.
04
Routing, portfolios, schedules. Treat the category with patience and suspicion. A surprising number of announced speedups have lost, later, to a careful classical solver. The comparison is the result. The adjective is not.
05
Clocks, gravimeters, and magnetometers do not need a universal computer. Some of them are closer to ordinary use than any machine that could threaten a cipher. Navigation and imaging may feel quantum before general-purpose computing does.
06
The plausible machine is a quantum processor beside a classical supercomputer, not instead of one. Decoders, compilers, control electronics, and the cryogenics are part of the computer. Nothing in the public record suggests the ordinary bit is going away.
The horizon
The record is what has been published. The other three are a way to read the decade. They are not a forecast, and they are not anyone’s plan.
Public record
Physical qubits are a solved demonstration and an unsolved product. A below-threshold error-correction experiment has been published. Post-quantum cryptography has standards with names and document numbers. “Quantum advantage” has been claimed for particular tasks, some of them narrow on purpose. Utility is a different word, and it has not been earned in general.
What stays
01
Mail, photographs, maps, and ordinary applications are not waiting on a qubit. A future in which those move to a quantum processor is not a serious engineering forecast.
02
A quantum result that cannot be checked is a rumor with good lighting. The useful algorithms are the ones a classical machine can verify, or a second laboratory can repeat.
03
It becomes a colleague. Error decoding, compilation, and control are classical jobs sitting on top of a quantum one. The stack gets stranger. It does not get replaced.
04
It has a better publishing record than any press cycle this field has run. Qubit counts are cheap. A falling logical error, on a circuit someone else can specify, is not.
