The Algorithm Gap
Here’s the open secret of quantum computing: even a perfect one would help with surprisingly few problems — we simply don’t know many quantum algorithms that win.
Almost all quantum discussion assumes the hardware is the only obstacle — build a big enough, clean enough machine and quantum computing transforms everything. This lesson reveals a second obstacle that gets far less airtime and is just as important: even a perfect, fault-tolerant quantum computer would help with a surprisingly small set of problems. The bottleneck isn’t only building the machine — it’s that we simply don’t know many quantum algorithms that actually beat classical ones, and finding new ones is brutally hard. This is one of the most honest things you can understand about the field. Hold the question: if you had a flawless quantum computer tomorrow, what could you actually do with it?
A quantum computer is not a faster computer
The foundational misconception: people imagine a quantum computer as a universally faster computer — like a super-fast classical PC that speeds up everything. It is not. A quantum computer only offers a speedup for problems with a specific mathematical structure that its tricks — superposition and, crucially, interference (lesson 6) — can exploit. For the vast majority of everyday computing (email, spreadsheets, browsing, most software), a quantum computer offers no advantage at all — it would be slower and more expensive than your laptop. This follows directly from what you learned: the quantum magic is using interference to concentrate probability on right answers (lessons 9, 35), and that only works when the problem has the right shape. Most problems don’t. Quantum computing is a specialized tool, not a universal upgrade — the same conclusion the crypto track reached about blockchains, and just as important here.
We know only a handful of algorithms that win
Here’s the honest, striking part. After decades of research, the quantum algorithms with a proven, meaningful advantage over classical ones number in the handful. The famous families: Shor’s (factoring / breaking certain cryptography, lesson 10) — a dramatic exponential speedup, but narrow; Grover’s (searching, lesson 9) — a modest quadratic speedup, broadly applicable but not game-changing; quantum simulation (chemistry/materials, lesson 11) — the big, genuinely broad one; and a scattering of specialized algorithms. And that’s roughly it for proven wins. Enormous swaths of computing — including most optimization (lesson 35) and machine learning (36) — have no known quantum algorithm with a clear advantage. This is the algorithm gap: the distance between “problems a quantum computer could theoretically help with” and “problems we actually know how to make it help with,” and it’s wide. The hardware race gets the headlines; this quieter gap may matter just as much.
What a perfect quantum computer would (and wouldn’t) speed up: • Factoring large numbers / breaking certain crypto → yes (Shor), exponential — but narrow. • Simulating molecules & materials → yes (simulation), broad and valuable. • Unstructured search → modestly (Grover), quadratic. • Email, spreadsheets, video, most business software, most optimization, most ML → no known advantage. • So even a flawless machine transforms a few domains, not computing in general.
Why finding new algorithms is hard — and what it means
Why don’t we just discover more quantum algorithms? Because thinking in quantum interference is genuinely, deeply unintuitive — you have to design a computation where the wrong answers cancel out and the right ones reinforce, across a space too large to picture, using rules that defy everyday logic. Very few problems have an obvious way to do this, and cleverly finding the structure that allows a speedup is some of the hardest work in computer science; major new quantum algorithms are rare events, years apart. So the honest, calibrated takeaway that this lesson adds to the module: quantum computing’s promise is real but bounded — even with perfect hardware, its impact is concentrated in the specific domains where we have (or find) algorithms with real advantage, above all simulation of quantum systems (its most robust, broad application, lesson 34). The two open frontiers are therefore both the hardware (fault tolerance, lesson 39) and the algorithms (finding new wins). Anyone who tells you a working quantum computer will speed up “everything” either misunderstands this or is selling — and now you know exactly why. (Honest and educational — grounded expectations, not dismissal.)
Think of a quantum computer as an extraordinary but highly specialized key. It opens a few very important, very specific locks — the “factoring” lock, the “simulate-a-molecule” lock — that no ordinary key can budge, which is genuinely remarkable. But most doors in the world aren’t those locks; they’re opened fine by the ordinary keys (classical computers) we already have, and jamming the fancy quantum key into them does nothing useful. The hardware race is about forging this key well enough to actually turn (fault tolerance). The algorithm gap is the separate, quieter problem that we’ve only discovered a handful of locks it fits — and finding new ones is painstaking, rare work. A magnificent key for a few crucial doors is not the same as a master key for all of them.
Reading a “quantum will revolutionize X” claim: 1. Does X have a known quantum algorithm with a real advantage? (Factoring, simulation → yes; most things → no.) 2. If no known algorithm, the claim rests on one being found — a rare, hard, uncertain event. 3. Is the speedup exponential (rare, e.g. Shor/simulation) or modest quadratic (Grover) or merely hoped-for (optimization/ML)? 4. Conclusion: quantum transforms specific domains where algorithms exist — especially simulation — not computing in general. The algorithm gap, not just hardware, bounds the impact.
This is the reading. The interactive version — active-recall quiz, a hands-on experiment you run in your own AI, and an earned mastery check — is free in the app.
Start this lesson free →