Precision systems for adaptive intelligence.
Quarks Lab
State change • Pipeline flow • Dynamic adaptation.
TwoQuarks explores stability-aware reinforcement learning: controlled environments, phase transitions, transient operators, and measurable structural adaptation.
DOWN — Transitions
State change • Pipeline flow • Dynamic adaptation. (Oct-Nov 2025)
Flavor Down
A reinforcement learning agent that remains functional when reward signals are deceptive or adversarial.
STRANGE — Color Confinement
Composite agents • Emergent reconstruction. (Nov 2025)
Flavor Strange
A phase-aware RL architecture for environments with hidden regime shifts.
TOP — Heavy States
High-mass architectures • Full-scale Instant stacks • Fermi Paradox Env. (Nov - Dec 2025)
Flavor Top
A transient operator that activates high-precision behavior only at critical moments, then disappears.
CHARM — Elementaries
Kernels • Lion meta-control • CUDA-level reasoning. (Oct - Dec 2025)
Flavor Charm
A stability-preserving RL variant designed to keep coherent behavior when the signal-to-noise ratio collapses.
UP — Truth 𝑻
DDQN • High Frequency • Observable results. (Oct 2025 - Jan 2026)
Flavor Up
An early-warning RL controller that tracks pre-critical instability and avoids collapse before it happens.
BOTTOM — orchestrates the collapse of interactions
Adaptive Derivation of the Emerging Agent. (Jan 2026)
Flavor Bottom
A minimal starting point: probes collapse interaction and establishes baseline stability measurements.