Qubit technologies, error correction challenges, and what's actually useful today
Quantum computing has been "ten years away" for thirty years. Yet the field has made genuine progress: quantum computers exist, they outperform classical computers on specific problems, and quantum advantage has been demonstrated in narrow domains. Understanding what quantum computers actually do—and don't do—is essential for assessing their practical prospects.
This article explains the current state of quantum computing: different qubit technologies, what quantum supremacy claims actually mean, the error correction challenges, and realistic applications in the NISQ (Noisy Intermediate-Scale Quantum) era.
Classical computers use bits—0 or 1. Quantum computers use qubits, which can exist in superposition of 0 and 1 simultaneously:
Classical bit: 0 OR 1
Qubit: α|0⟩ + β|1⟩ where α,β are complex amplitudes
|α|² + |β|² = 1 (probabilities must sum to 1)
When you measure a qubit, you get 0 with probability |α|² or 1 with probability |β|². The power of quantum computing comes from manipulating these amplitudes before measurement.
Quantum circuits apply gates to qubits, evolving their quantum state:
Entangled qubits share quantum state regardless of distance:
Bell state: (|00⟩ + |11⟩) / √2
Measuring first qubit as 0 → second qubit is definitely 0
Measuring first qubit as 1 → second qubit is definitely 1
Correlation is stronger than any classical explanation allows
Quantum computers offer speedup for specific problem structures:
The dominant approach, used by Google, IBM, and Rigetti:
| Company | Qubit Count | Connectivity | Coherence Time |
|---|---|---|---|
| IBM | 1,000+ (Eagle) | Heavy-hex lattice | ~300-400 μs |
| 70+ (Sycamore) | Nearest neighbor | ~100 μs | |
| Rigetti | 84 | Chiplet-based | ~50 μs |
Superconducting qubits operate at millikelvin temperatures (~15 mK, colder than outer space). They use Josephson junctions as nonlinear circuit elements.
Pros: Fast gate times (~nanoseconds), mature fabrication from semiconductor industry
Cons: Extreme cooling required, short coherence times, susceptibility to noise
Used by IonQ and Honeywell (now Quantinuum):
| Company | Qubit Count | Gate Fidelity | Coherence Time |
|---|---|---|---|
| IonQ | 32 (system) | 99.9% (2-qubit) | >1 hour (physicist measures coherence by identity operations) but algorithmic coherence is much shorter, limited by entanglement fidelity|
| Quantinuum | 32 (H2) | 99.8% (2-qubit) | >1s |
Trapped ions are held in electromagnetic traps and manipulated with lasers. All-to-all connectivity is native—no qubit routing required.
Pros: Higher gate fidelity, all-to-all connectivity, longer coherence
Cons: Slow gates (microseconds), complex laser systems, vacuum requirements
Xanadu and PsiQuantum are developing photonic quantum computers:
Photons (light particles) carry qubits as polarization or time-bin encoding. Photonic systems operate at room temperature.
Pros: Room temperature operation, photonics integrates with fiber optics
Cons: Hard to make photons interact (needed for 2-qubit gates), photon loss
Google claimed "quantum supremacy" in 2019 by running a random circuit sampling problem on Sycamore in 200 seconds that would take ~10,000 years on classical computers.
Chinese researchers using photonic systems (Zuchongzhi) have also claimed quantum advantage for specific problems. These claims are similarly narrow—they demonstrate quantum capability on designed benchmarks rather than practical tasks.
True quantum advantage for practical problems requires:
We're currently in the NISQ era: Noisy Intermediate-Scale Quantum computers with 50-1000 qubits that cannot yet perform error-corrected computation.
Quantum states are fragile. Environmental noise causes errors. Current qubits have error rates of ~0.1-1% per gate—far above the ~0.001% needed for practical algorithms.
To run Shor's algorithm to factor a 2048-bit RSA key, you need approximately:
Current machines have 100-1000 physical qubits. We're 3-4 orders of magnitude short.
The surface code is the leading approach:
Physical qubits arranged in 2D grid:
- Data qubits (blue) store the logical qubit state
- Measurement qubits (red) detect errors
To create 1 logical qubit with error rate < 10⁻¹⁰:
Requires ~1000 physical qubits at current error rates
Other codes (color codes, LDPC codes) aim for better efficiency but are less mature.
Estimates for when fault-tolerant quantum computing arrives vary wildly:
Current quantum computers can run variational algorithms for specific optimization and simulation problems:
The jury is out on whether NISQ algorithms provide genuine advantage over classical heuristics.
The most promising near-term application is simulating quantum systems—molecules, materials, drugs—that are exponentially hard for classical computers:
| Framework | Provider | Language |
|---|---|---|
| Qiskit | IBM | Python |
| Cirq | Python | |
| Braket | Amazon | Python |
| Circuit | IonQ | Python |
| Custon SDK | Xanadu | Python (Strawberry Fields) |
All major quantum computing companies offer cloud access:
Pricing is free for basic access; premium plans for dedicated hardware access run thousands to millions of dollars per year.
Quantum computing is a genuine technology—quantum computers exist and can solve specific problems faster than classical computers. However, the transformative applications (cryptography, drug discovery, optimization) require fault-tolerant quantum computers that don't yet exist.
The path from current NISQ devices to fault-tolerant quantum computers requires:
This is a 10-30 year challenge. Organizations should monitor progress, engage with quantum software platforms to build expertise, and identify problems where quantum simulation might provide value—but avoid overinvesting based on near-term promises.