Why quantum computers won't replace your laptop (and what they'll actually do)

There’s a persistent fantasy that quantum computers will eventually replace classical ones. Faster email. Snappier spreadsheets. Better video games.

This won’t happen. Quantum computers are not a better version of your laptop. They’re a different tool entirely — like a submarine versus a car. One isn’t “better” than the other. They solve different problems.

What quantum computers are genuinely good at

Quantum advantage exists for a specific set of problems. Here are the main ones, in order of how close they are to being practical:

1. Simulating molecules and materials

Why it matters: Drug discovery, battery design, fertiliser production, materials science.

Why quantum helps: Molecules are quantum systems. Simulating them on classical computers requires exponentially more resources as molecules get bigger. A quantum computer can simulate quantum behaviour directly, using the same physics.

Where we are: This is probably the first domain where quantum computers will be genuinely useful. Companies like Algorithmiq are already demonstrating quantum simulations of cancer drugs on today’s hardware. The simulations are limited, but they’re real.

The practical future: Pharmaceutical companies call a quantum API to simulate how a drug candidate binds to a protein. The simulation takes minutes instead of months. This changes which drugs get developed and how fast.

2. Optimisation

Why it matters: Logistics, supply chains, financial portfolio management, scheduling, network routing.

Why quantum helps: Many optimisation problems involve searching a vast landscape of possible solutions. Quantum algorithms (like QAOA — the Quantum Approximate Optimisation Algorithm) can sometimes navigate these landscapes more efficiently.

Where we are: Still mostly theoretical for practical-sized problems. D-Wave’s quantum annealers tackle optimisation directly, but whether they outperform the best classical algorithms is still debated.

The practical future: A logistics company calls a quantum optimisation service to route 10,000 deliveries across a city, considering traffic, time windows, vehicle capacity, and driver schedules simultaneously. The quantum service returns a better solution than classical solvers — not because it’s faster, but because it can handle more constraints.

3. Breaking (and fixing) cryptography

Why it matters: Everything encrypted with RSA or elliptic curve cryptography (which is most internet security) is theoretically vulnerable to quantum attack.

Why quantum helps: Shor’s algorithm factors large numbers exponentially faster than any known classical algorithm. Current encryption relies on the difficulty of factoring.

Where we are: Very far from practical. Breaking RSA-2048 would require millions of physical qubits. But the threat is taken seriously enough that governments and companies are already migrating to “post-quantum” cryptography.

The practical future: By the time quantum computers can break current encryption, most systems will have migrated to quantum-resistant algorithms. The real impact may be quantum key distribution (QKD) — using entanglement to create provably secure communication channels.

4. Machine learning (maybe)

Why it matters: Everyone wants faster, better AI.

Why quantum helps: Some quantum algorithms could speed up specific ML operations (like kernel methods, sampling, or linear algebra).

Where we are: This is the most speculative category. Quantum ML is a hot research area, but practical advantages over classical ML haven’t been demonstrated yet. Classical hardware (GPUs, TPUs) is very fast and keeps getting better.

The practical future: Uncertain. Some researchers believe quantum ML will find its niche in specific problem types. Others think classical hardware will scale fast enough to make quantum ML irrelevant. This is genuinely unknown.

What quantum computers will NOT do

• Replace GPUs for AI training. The data loading problem alone makes this impractical.
• Speed up web browsing, gaming, or office apps. These are already perfectly handled by classical computers.
• Solve NP-complete problems instantly. Quantum computers provide at most a quadratic speedup for search (Grover’s algorithm), not an exponential one. NP-complete problems remain hard.
• “Try all answers at once.” As we explained in the interference article, quantum speedups come from clever amplitude manipulation, not brute-force parallelism.

The API prediction

Here’s what we think actually happens:

Quantum computing will become a cloud service. You won’t own a quantum computer any more than you own a GPS satellite. Companies like IBM, Google, Amazon, and specialised startups will offer quantum services via API:

# This is roughly how it'll work (conceptually)
result = quantum_service.simulate_molecule(
    molecule="aspirin_derivative_47b",
    accuracy="chemical"
)

You’ll call the API, specify your problem, and get results. You won’t need to understand qubits, error correction, or gate operations — just like you don’t need to understand transistor physics to use a spreadsheet.

The value will be in knowing which problems to send to the quantum service and how to interpret the results. That’s why understanding the field now matters — not to build quantum hardware, but to recognise when quantum services become relevant to your work.

The honest summary

• Quantum computers solve specific problems: molecular simulation, optimisation, cryptography, possibly ML
• They will NOT replace classical computers for everyday tasks
• The practical form factor will be cloud APIs, not personal quantum laptops
• The competitive advantage comes from knowing which problems benefit from quantum
• Molecular simulation is closest to practical; cryptography impact is furthest but most consequential

What’s next?

Understanding what quantum computers can do leads naturally to understanding who’s building them and why their approaches differ. Five fundamentally different technologies are competing to win the quantum hardware race.