Verification register Energy, Climate, Space & Materials
Current reading
The readiness gap, in one scan
AI-assisted assembly · derived results
- Claimed
- 80
- Reported
- 76
- Verified
- 67
- Gap
- +13
Public ambition and stated capability
Observed practitioner reporting
Independently supported evidence
Claimed minus verified
Evidence strength Strong
Decision
What the current evidence supports
Human editorial judgment · 2026-06-28
Adopt with guardrails
- Why
- Strong primary-source benchmark wins (8%+ over 12 FMs, only FM to beat U-Nets on PANGAEA), open Apache-2.0 weights, real edge/on-satellite deployment proof, and active downloads — but the 'mental images' caveat and distillation tradeoffs mean generation outputs need guardrails and per-task validation.
- Next
- Pilot TerraMind-base/large or TerraMind.tiny via the open Hugging Face weights (Apache 2.0) for a segmentation/classification task; treat any generative any-to-any output as a prior, not ground truth, and benchmark against a task-specific U-Net
Constraints
Blockers
No named blocker is present in the current public projection.
Evidence summary
Derived counts
AI-assisted assembly
- Total
- 7
- Tier 1
- 0
- Tier 2
- 0
- Tier 3
- 7
- Supports
- 4
- Contradicts
- 2
- Context
- 1
- Latest observed
- 2026-06-01
Counts and dates only. Raw signals, private excerpts, trust records, and internal corpus material are not published here.
Publication record
Revisions
Initial public reading
This is the initial public reading. No earlier readiness change is recorded.