Every few months, a quantum computing press release drops that seems to announce the imminent end of classical computing as we know it. IBM announces a 1000-qubit processor. Google claims quantum supremacy. Some startup promises to break RSA encryption "within five years." The breathless coverage follows, then the inevitable backlash from physicists who explain that it's all more complicated than reported.

Let's cut through the noise. What can quantum computers actually do today? When will they matter for software developers? And what's hype versus genuine progress?

What Quantum Computers Actually Are

Classical computers store information in bits — 0 or 1. Quantum computers store information in qubits, which exploit quantum mechanical properties to exist in superpositions of 0 and 1 simultaneously.

But superposition alone isn't magic. The real power comes from two other quantum phenomena:

  • Entanglement — Two qubits can be correlated such that measuring one instantly tells you the state of the other, regardless of distance. This allows quantum computers to represent correlations between variables exponentially more efficiently than classical systems.
  • Interference — Quantum algorithms manipulate probability amplitudes, causing wrong answers to cancel (interfere destructively) while correct answers reinforce (interfere constructively). This is the actual mechanism of speedup.

The result: for specific problem types, quantum algorithms can find solutions exponentially or quadratically faster than the best known classical algorithms.

What Quantum Can (and Cannot) Speed Up

This is where most popular coverage goes wrong. Quantum computers are not faster computers in general. They're specialized_ accelerators for specific problem classes.

Problems where quantum has proven or likely speedup:

  • Factoring large integers (Shor's algorithm — threatens RSA encryption)
  • Searching unsorted databases (Grover's algorithm — quadratic speedup)
  • Simulating quantum systems (chemistry, materials science, drug discovery)
  • Certain optimization problems (quantum approximate optimization)
  • Linear systems of equations (HHL algorithm — with significant caveats)

Problems where quantum provides little or no speedup:

  • Most machine learning training (the data loading bottleneck kills any quantum speedup)
  • Video encoding, compression, database queries — classical algorithms are near-optimal
  • Sorting, string processing, virtually all everyday computing tasks
  • Running classical software — quantum computers can't just "run your app faster"
⚠️
The "Quantum AI" Myth
"Quantum machine learning" is heavily marketed but largely unproven. Most QML papers describe speedups on artificially small datasets or make assumptions that don't hold in practice. Be very skeptical of quantum AI claims.

The Error Problem: Why We're Not There Yet

Today's quantum computers are what researchers call NISQ devices — Noisy Intermediate-Scale Quantum processors. "Noisy" is the key word. Qubits are extraordinarily fragile. Thermal vibrations, electromagnetic interference, even cosmic rays can flip a qubit or corrupt a computation. Current error rates are roughly 0.1-1% per gate operation.

Shor's algorithm to break 2048-bit RSA would require millions of near-perfect logical qubits. We currently have thousands of physical (noisy) qubits. Turning noisy physical qubits into reliable logical qubits via quantum error correction requires roughly 1,000-10,000 physical qubits per logical qubit.

"We are where classical computers were in 1955 — the hardware exists and works, but building useful applications requires solving the reliability problem first."

What Actually Works Today

Despite the limitations, there are genuine applications where NISQ computers provide value:

  • Quantum simulation of molecules — IBM and Google have demonstrated quantum simulations of small molecules (like lithium hydride) that are difficult for classical computers. This has real drug discovery and materials applications.
  • Quantum random number generation — True hardware randomness from quantum measurement, useful for cryptography.
  • Quantum key distribution (QKD) — Provably secure communication channels based on quantum mechanics, commercially deployed in some finance networks.
  • Optimization on small problems — Quantum approximate optimization algorithms show promise on certain portfolio optimization and logistics problems.

The Cryptography Threat: When to Worry

The one area where developers should be actively planning: post-quantum cryptography.

A cryptographically relevant quantum computer (one capable of running Shor's algorithm on real keys) is likely 10-20 years away. But here's the problem: data encrypted today with RSA or ECC can be stored by adversaries and decrypted retroactively once quantum computers mature. This is the "harvest now, decrypt later" attack.

NIST finalized its post-quantum cryptography standards in 2024 — ML-KEM (formerly CRYSTALS-Kyber) for key encapsulation and ML-DSA (formerly CRYSTALS-Dilithium) for digital signatures. Start planning your cryptographic migration now, especially if you handle data with long security requirements.

🔐
Action Item for Developers
Audit your cryptographic dependencies. If you're using RSA, ECC, or Diffie-Hellman for anything sensitive, add post-quantum migration to your technical roadmap. OpenSSL 3.5+ and most major TLS libraries now support hybrid classical/post-quantum key exchange.

The Players Worth Watching

  • IBM Quantum — Largest publicly accessible quantum network, open Qiskit SDK, strong research output
  • Google Quantum AI — Demonstrated quantum supremacy (contested), superconducting qubit approach
  • IonQ — Ion trap approach with high fidelity qubits, publicly traded
  • Quantinuum (Honeywell) — Record-breaking fidelity, serious enterprise focus
  • PsiQuantum — Photonic approach, targeting fault-tolerant quantum at silicon fab scale
  • Microsoft — Long-term bet on topological qubits (still unproven but potentially transformative)

Should You Learn Quantum Computing?

If you're a software developer: not urgently, but the fundamentals are worth understanding. Spend a few hours with IBM's Qiskit documentation or the Quirk circuit simulator. Understand superposition and interference at a conceptual level. Know which problem types have quantum speedup and which don't.

If you work in cryptography or high-performance computing: quantum literacy is increasingly important. Follow NIST's post-quantum standards. Experiment with hybrid approaches.

If you work in chemistry, drug discovery, or materials science: quantum simulation is genuinely coming for your field within 5-10 years. Start building quantum intuition now.

For everyone else: watch the space, ignore the press releases, and bet on 2030-2035 for the first commercially relevant quantum advantages in non-specialized domains.

TF Editorial

TF Editorial

Editorial Team · Tomfoolering

We write about technology with depth and without condescension.