LLM Wiki Index

This wiki exists to make Alex’s research memory operational: source-traceable notes, stable names, and durable cross-links give agents one maintained interface for ingestion, retrieval, and follow-up work while the raw corpus remains the source of truth.

North Star

The central frame for this wiki is Foundation Time-Series Model Research Agenda. It defines the slots into which source pages, topic pages, entities, and durable ideas should be mapped.

The practical North Star is to build digital-world robots: action-conditioned agents for digital, organizational, and cyber-physical systems. They observe telemetry, event streams, business signals, topology, process context, and action history; maintain latent state; reason over plausible futures; and choose typed interventions. Today the closest testbeds are SRE observability and telecom because they expose dense time-series streams, system structure, and operational controls. The intended scope is broader: digital marketing, sales operations, supply chains, finance, industrial systems, and eventually safety-critical infrastructure such as nuclear power plants.

This wiki treats time series as evidence about evolving system state, not merely as sequences to forecast. Forecasting matters, but the central question is whether a model can maintain useful latent state, use context, reason about plausible futures, and eventually support action-conditioned decisions. What’s Wrong With The Current Time-Series Deep Learning? is the landmark position source for that argument.

Contact / Collaboration

If this research direction resonates with you, I would be happy to talk with like-minded people, collaborate on research, and work on use-cases together.

Ideas are not the bottleneck. Hands are. Time-series modeling should be moving at least as fast as vision, audio, and robotics.

Files

  • contradictions.md - Open tensions, unresolved disagreements, and stale-claim watchpoints across the wiki.

Machine-Readable Surface

This wiki is human-readable, but the best way to read it is to point your agent at the manifest or a specific raw wiki page, then ask it to explain that page or answer a question from it.

Directories

  • entities - Named systems, models, datasets, and benchmarks that recur across the corpus.
  • ideas - Alex-curated project ideas, research agendas, and system-design sketches.
  • sources - One generated source page per ingested paper.
  • topics - Cross-source concepts, methods, and research themes.