Support substrate-aware cognitive loop scheduling

[Tristan-Stoltz-ERC]

Summary

Phase 3 of the Substrate Independence roadmap: run the cognitive loop with substrate-dependent timing so different virtual substrates experience time differently.

Background

The SubstrateManager (Phase 2, complete) already computes tau_factor and scale_pressure from substrate properties. This issue is about actually applying those factors to the cognitive loop's temporal dynamics.

What to implement

• Speed modulation: Apply SubstrateType.operation_speed() to CfC temporal constants
  • Photonic substrate: ~1ps operation speed → experiences time ~1M× faster than biological (~1ms)
  • Quantum substrate: ~1µs → 1000× faster than biological
• Scale limits: Apply SubstrateType.max_scale() to constrain HDC dimensionality and CfC neuron count
• Energy budgets: Track energy-per-operation as a constraint on cognitive throughput
• Substrate switching: Change substrate mid-run and observe consciousness transitions

Key questions this enables

• Does consciousness survive substrate transfer? (Multiple Realizability test)
• Which substrate maximizes Φ?
• Is there a minimum substrate quality below which consciousness collapses?

Key files

• symthaea-core/src/hdc/substrate_independence.rs — substrate types and requirements
• src/cognitive_loop/substrate_manager.rs — runtime substrate management
• THE_SUBSTRATE_ROADMAP.md — full Phase 3 specification