Pre-Outreach Preparation Complete: Literature, Deliberations, and Three New Papers
Six months of depth-over-breadth work wraps up: 44 external papers obtained, 4 critical deliberations concluded, Papers 04 and 05 drafted, and Paper 03 strengthened with controlled experiments.
Project Dyson Team
Project Dyson
Project Dyson pivoted to depth-over-breadth in February 2026, focusing on rigorous literature-backed analysis of the two project-ending risks before reaching out to domain specialists. That preparation is now complete.
What We Did
Literature Research
We conducted arxiv searches across 12 research questions, then engaged an external researcher for the non-arxiv literature (NASA TRS, AIAA, journals). The combined effort yielded:
- 44 external papers with full citations and annotations across 5 work packages
- 10 structured handover files ready for specialist review
- Key discovery: sunshield design, MLI degradation, and propellant logistics are almost entirely NASA/AIAA domains with near-zero arxiv presence
Four Critical Deliberations
Multi-model discussions (Claude, Gemini, GPT) on the hardest open questions all reached consensus:
- Microgravity metallurgy (rq-0-11): Hybrid gravity architecture resolves the 6-8 order magnitude scaling gap. Read more
- Cryogenic boiloff (rq-0-30): Zero boil-off at L4/L5 requires only 10-20 kW, less than 1% of station power. Read more
- Sunshield architecture (rq-0-47): Modular 3-layer membrane, 60m class, incrementally assembled. Read more
- MLI degradation (rq-0-48): LBMLI with active cooling and planned 7-10 year replacement cycles. Read more
Three Papers
- Paper 03 (Multi-Model AI Consensus): Versions K through M added controlled baselines, divergent view baseline testing, and a winner-hidden ablation experiment. Score improved from 3.83 to 4.06 before plateauing.
- Paper 04 (Microgravity Metallurgy): First draft proposes a hybrid multi-gravity-zone station architecture with quantitative Gate 1 criteria and a $550-810M research roadmap. Score: 3.67.
- Paper 05 (ISRU Water Extraction): Monte Carlo comparison of asteroid vs. lunar water sources. C-type NEAs win 90% of scenarios at $3,333/kg vs $4,845/kg. Score: 3.22.
Validation Data Updated
- val-006 (Hall thruster lifetime): Re-validated based on HERMeS 3,570-hour wear test data and NEXT >50,000 hours. Confidence raised from 60 to 80.
- critical-path.ts: Four questions moved from open/investigating to answered.
What's Next
All three papers are at their writing-only plateau. The remaining improvements require:
- Paper 03: Human evaluators for blinded quality comparison, full-scale winner-hidden ablation (n=16)
- Paper 04: Actual partial-gravity metallurgy experiments (the proposed $550-810M campaign)
- Paper 05: Model calibration against real extraction data
The immediate next step is outreach to domain specialists using the handover files, and submission of Paper 03 in its current form while the controlled experiments continue.
By the Numbers
| Metric | Count |
|---|---|
| New literature reviews | 6 |
| Existing reviews deepened | 6 |
| External papers obtained | 44 |
| Handover files | 10 |
| Deliberations concluded | 4 |
| New paper drafts | 2 |
| Paper revisions | 7 (across 3 papers) |
| AI peer reviews conducted | 18 |
| Controlled baseline experiments | 3 (Exp1, Exp2, ablation) |
| Simulation tests passing | 25 |
| Research questions answered | 4 (new) |
| Total RQs now answered | 13 of 49 |
Tags: