BeachBook as Citizen Science Infrastructure

BeachBook → Research Platform pivot

The contact said "too early-stage" about the grant, but he also said BeachBook. The Vinebrooke introduction isn't just an academic contact — it's a path from consumer app to institutional research infrastructure.

Universities, watershed councils, and parks agencies need ground-truth environmental monitoring tools. BeachBook already does this. The pivot costs nothing to frame — it's a different landing page and pitch deck, not a different codebase.

The Salmon Arm 20+ parks lead is exactly the institutional proof-of-concept a UAlberta research proposal needs.

HailStorm + Farm satellite data = commodity signal layer

HailStorm retrained successfully on April 17. The farm reports system generates NDVI/NDMI time series by crop region. There's a known correlation between crop health indices and agricultural futures prices (wheat/canola).

The ML pipeline in HailStorm already consumes time-series features. Feeding nimpact's satellite data as an additional feature vector is a natural extension — and it would make HailStorm actually differentiated from generic trading bots.

Sparky brain → MCP server product

The @driftwest/mcp-server is already on npm at v1.3.0. The brain architecture is the actual innovation — not the Sparky app. Packaging the memory system (hippocampus, amygdala, pattern cortex, dreams, predictions) as an installable MCP server that any Claude Code user can drop into their project is a real product.

The "We Built a Brain" blog post is already the marketing copy. This is the thing that could go viral on Hacker News.

BeachBook 401 bug is blocking the Vinebrooke demo

The download URLs bypass API key headers via launchUrl, returning 401. The contact or Vinebrooke may ask for a demo before the next conversation. Right now you can't actually show report downloads working in the app. Small fix with disproportionate stakes.