AI Scaling & Capability
What drives capability in learned systems — and what limits it.
Architecture, data, compute and evaluation: separating what actually produces capability from what merely accompanies it.
Drag to pan · Ctrl + scroll to zoom
Questions being explored
Expeditions in progress
Signals at the edge
- Active research frontierWhat fundamentally limits transformer scaling?Confidence 0.40 · medium
- Active research frontierWhich questions became approachable only recently?Confidence 0.30 · medium
- Contradiction clusterWhat does a capability jump measure?Confidence 0.38 · medium
0 charted findings and 0 bridge candidates touch this field. See what is charted →
Why this field matters
Enormous resources are allocated on the assumption that current scaling relationships continue. Knowing which constraint binds next — and whether our measurements of capability measure capability — is decision-relevant far beyond the field.
How your support is allocated
Backers who chose a focus are honoured first; the rest follows these weights (demonstration values), then each target’s expected usefulness. Every batch is recorded with its reasons — most recently A-428, $7,274.76 routed.
- Active expeditions34%
- Replication20%
- Open questions12%
- Frontier signals10%
- Scouting8%
Popularity chooses attention. Evidence chooses knowledge.