Measure the Gap

Every quantum headline, vendor pitch and board briefing contains a claim, such as a qubit count, a date, a detection range or a threat. Someone has demonstrated all of it, part of it, or nothing yet. Measuring the gap means finding out which, in numbers, before you act on it. We teach that habit in every course, in every field, and it’s our motto.

Hype and dismissal

People make two kinds of misleading claim about quantum technology. Some overstate what it can do, and how soon. Others call it decades away, or impossible, and wait. We measure both kinds the same way.

Both kinds cost money. Buyers who believe the first pay for products that can’t do what the brochure says. Organizations that believe the second miss deadlines that are already set and overlook capabilities that work today.

Security and post-quantum cryptography

Security teams hear that RSA-4096 is already broken in a secret lab, and that RSA-2048 will never be factored. We measure both claims against the public record.

Already broken in secret

In September 2026, two teams set new public records for factoring RSA challenge numbers, on classical GPUs. The larger number, RSA-896, is 896 bits long. Both teams proved their results by publishing the factors, which anyone can multiply to check. A lab that had broken RSA-2048 or RSA-4096 could prove it the same way, with the RSA-2048 challenge number published in 2001. Nobody has. The NSA and the UK’s NCSC, which see classified intelligence, still set migration deadlines as late as 2035. Those deadlines make sense only if RSA still works today.

Never

The opposite claim contains no date to check. We measure it against what has been demonstrated and what has been estimated. In December 2024, Google showed error correction working below threshold: each step up in code size halved the logical error rate. Researchers have cut the estimate for breaking RSA-2048 from about 20 million noisy physical qubits in 2019 to under a million in 2025. Measured against that record, “never” is the claim that needs evidence.

For planning, you need three numbers more than you need a date:

how long your data must stay confidential
how long your migration will take
which deadlines already apply to you

The PQC Migration Framework gives four to fifteen years as the migration time for most large enterprises. With ten years of confidentiality and a ten-year migration, your data is safe only if no capable quantum computer exists within twenty years. Nobody can guarantee that. Mosca’s inequality is the name for this calculation, and the deadlines apply whatever year such a machine appears. The Global PQC Migration Timeline lists them for more than 20 jurisdictions.

Sensing and metrology

Sensing has both kinds of claim in one field. Someone who calls all quantum sensing hype is wrong about clocks: every GPS satellite carries several atomic clocks. Quantum magnetometers and gravimeters are also in use, in medical imaging, surveying and defense. Someone who sells quantum radar for air defense is selling a laboratory result.

In 2020, researchers used entangled microwaves to detect a room-temperature object one meter away, and their receiver beat a comparable classical noise radar. The advantage exists only for extremely faint signals, at powers far below what long-range radar transmits. Jeffrey Shapiro of MIT, who helped develop the theory, has concluded that its practical value is severely limited. Vendors still market quantum radar. We teach you to ask for the demonstrated range and the conditions behind it.

Computing, networking, AI and leadership

Quantum computing

The overstated claim is that a thousand qubits make a useful quantum computer. The dismissive one is that error correction will never work. We ask whether each qubit count means physical or logical qubits. We also ask whether each capability exists today, in simulation or only on a roadmap. Error correction below threshold has been demonstrated. No one has demonstrated a fault-tolerant machine running useful algorithms at scale.

Communications and networking

Some call quantum key distribution (QKD) unhackable. Others say it has no use. The security proof covers the physics of the link. The equipment still requires classical authentication, and operators add trusted relay sites on long routes. Researchers have also attacked commercial QKD systems through their detectors. We teach where QKD fits beside post-quantum cryptography, and what it costs.

Quantum AI

The overstated claim is that quantum machine learning already beats classical models. The dismissive one is that AI and quantum computing have nothing to offer each other. We compare every quantum benchmark against a serious classical baseline. Engineers already use machine learning to calibrate quantum computers and decode their errors. In 2024, Google DeepMind’s AlphaQubit decoder made fewer errors than the previous leading decoders on Google’s processor data.

Leadership and strategy

Boards hear both kinds of claim as a date. One briefing puts Q-Day, the day a quantum computer can break RSA, three years away. The next puts it thirty years away. We teach leaders to ask for the assumptions behind any single date. They then plan from the obligations that apply whatever the date, such as regulatory deadlines, contract terms and questions from insurers.

What you do in our courses

We build every course around the questions practitioners ask in their own work. You check qubit counts and roadmap dates in the computing courses. In sensing, you check specification sheets against classical baselines. In networking, you check security proofs against equipment. In the post-quantum courses, you work out urgency from your own data and deadlines, without relying on vendor-driven fear.

In the seven Quantum Technology industry courses, you leave with a table of every use case taught, each labeled demonstrated, announced or estimated, with its sources. In the Quantum-Safe sector intensives, your team builds an initial priority list and separates what is evidenced from what is assumed.

How we apply it to our own work

We apply the same test to our own material. Before publication, a reviewer checks each qubit count for whether it means physical or logical qubits. The reviewer also checks each capability claim for whether the result has been demonstrated or only announced. Someone on the team reads every source before it is cited. The AI Editorial Policy sets out the full review.

If you find an error in a course, a page or an article, email support@quantumacademy.com. We check every report and correct confirmed errors.

Six questions for vendor claims

The article A Baloney Detection Toolkit for Quantum Claims sets out six questions you can send to any vendor, with a worked example. It also covers the opposite failure, reflex dismissal, and how to test a critic’s claims the same way.

Where to start

Each field has its own page and courses:

See the training schedule

Browse all courses

Quantum Academy programs are educational and informational only, and are not legal, compliance, or engineering advice.