coherent.
Chapter I / First signal
LAB 00:00

One qubit.
A world of possibility.

A room. A refrigerator. A very fragile beginning.

1physical qubitsuperconducting laboratory
Your first experiment

Start with
one fragile thing.

The first apparatus

SC / Q01
Your laboratory instrument is drawn here. All controls and measurements are available as text.
○ Physical qubit   ▪ Control portIllustrative schematic
Optional. Reopen Help anytime with ?

One qubit is a beginning. A qualified experiment is progress.

On the research desk

Every breakthrough has a beginning.

0 discoveries
From the journal

One qubit. An entire room to keep it cold.

The science behind the discovery

Your laboratory

A few quiet controls.

Appearance
20%

Original apparatus sounds: cooling air, control pulses, decoder ticks and mechanical cues. Select equipment to hear its signature. Sound is enabled by default and begins after your first interaction. Mute it in the header or adjust its volume here.

Visible, unpaused play only. Hidden and closed tabs receive no offline progress. Export if browser storage is unavailable.
New laboratory

Begin again?

This replaces this browser’s current laboratory. Export your save first if you want to keep the run.
A model, with its assumptions intact

Fragile things.
Careful claims.

Coherent is an educational economy about superconducting qubits and surface-code engineering. Prices, grants, laboratory seconds, noise coefficients, factory layouts, and workload recipes are game scenarios.

Research trust is divided between staff and notebook space. Full notebooks give a design-generation bonus. Engineering designs and classical analysis stations are laboratory game abstractions, not quantum states.

Services share post-calibration apparatus and controller capacity with classical automation. A delivered customer batch represents 1,024 known Bell preparations and computational-basis measurements. Prices, demand, rates, construction, and maintenance are fictional economy presets.

Finite frontier goals and precision requests require a fresh classically sampled trial. The interval is per trial; adaptive tuning and repeated attempts do not create a campaign-wide confidence guarantee. More acquisitions cost funding and apparatus laboratory time. Received prefab equipment adds capacity only after funded workshop commissioning.

The 2026 controller profiles are authored throughput/latency tradeoffs, and adaptive control lowers one selected maintenance requirement. They do not improve every quantum error metric. Compact and parallel schedules retain the same 32-data-site task; workspace, waiting and fresh-state supply remain in the budget.

The small two-spin tutorial and toy certificates are really computed classically in your browser. Larger workloads are resource scenarios. They do not produce a large quantum answer, measured hardware accuracy, or evidence of quantum advantage.

Ramsey T₂*, echo T₂, sampling uncertainty, residual bias, RB characterization, threshold-model noise, protected memory, and qualified operations remain distinct. Syndrome events do not reveal an unknown data state.

Ideal rotated patches use 2d²−1 physical qubits. Factories reserve fictional patch-sized allocation units, not real tile layouts. Preparation credits are rehearsal currency, not stored magic states.

Workload previews include idle memory while waiting, operations, accepted factory-output errors, preparation, repetitions, decoding, feedback, and complete modeled time. Their error budget is conditional on selected estimates, not an experimentally certified guarantee.

All paper links are available in Research before and after a discovery. The numerical model received independent AI scientific review; no credentialed human specialist or hardware experiment is implied.

Read the implementation model ↗