Enter FAULTLINE
Campaign node 1 and commit a baseline transaction.
OPEN STOP →FOUR ROUTES / THREE MINUTES EACH
Choose a claim.
Attack it.
Each route is a deliberately short sequence: mechanism, adversarial condition, oracle, boundary. A reviewer can stop after three minutes or follow any claim into code and tests.
Create a minority
Isolate the leader and observe what cannot commit.
OPEN STOP →Ask the oracle
Compare a witnessed linearizable history with a stale read.
OPEN STOP →Read the boundary
Separate logical ticks from production latency.
OPEN STOP →Maneuver inside bank
Compare single-CV and IMM localization.
OPEN STOP →Cross two tracks
See why lower localization can still mean worse total OSPA.
OPEN STOP →Remove parallax
Replay the VIO hover failure.
OPEN STOP →Follow the limits
Read why each method loses differently.
OPEN STOP →Open the tape
Start the deterministic market replay.
OPEN STOP →Control inventory
Compare naive and inventory-skew regimes.
OPEN STOP →Raise toxicity
Make stale quotes easier to attack.
OPEN STOP →Test the headline
Inspect P&L tradeoffs and the digital-payoff failure.
OPEN STOP →Compare shapes
Open KERNELARENA’s live compiler matrix.
OPEN STOP →Replay passes
Inspect canonicalization, fusion, DCE, and memory reuse.
OPEN STOP →Read WGSL
Connect scheduled kernels to emitted source.
OPEN STOP →Probe locally
Measure CPU and WebGPU with numerical comparison.
OPEN STOP →