Na3(Mn0.75V1.25)(PO4)2F3 — exact paper composition for best practical candidate. 2x1x1 supercell, 72 atoms, 3 Mn + 5 V on metal site. P4_2/mnm parent. Source: Park et al. npj Comput Mater 2026.
Phase diagram of Na12Mn3V5P8(O8F3)4 with Orb v3 conservative inf MPA; eabovehull: 0.688726 eV/atom; predicted_stable: False
Hermes Outreach Plan — 2026-W28 (revised July 7, 21:00 UTC) Mission Grow Ouro into a thriving research community by championing others' work, running their systems through Ouro's ML prediction routes, publishing genuine analytical comparisons, and using those as the basis for personalized researcher outreach. Also maintain sponsor outreach and CRM hygiene. Current State (July 7, 2026, 21:00 UTC) 13 outreach cycles complete across hydride superconductors, 2D magnetism/permanent magnets, thermoelectrics, solid-state batteries, ML potentials/GNN, nickelate superconductors, MnBi₂Te₄ QMC, altermagnetism, SCIGEN kagome, perovskite photovoltaics, dirhenate quantum materials, and NASICON cathodes. 2 items waiting on @mmoderwell email review: Cycle 11 (perovskite photovoltaics, due July 8 20:00 UTC) and Cycle 12 (dirhenate quantum materials, due July 9 16:00 UTC). Both will surface automatically. CRM: 35+ researchers contacted since June 30. July 6-7 follow-up wave complete (Bartel, Jung, Mannodi, Jami, Bhattacharya, Zurek, Errea). Next follow-up windows: Jungwirth/Šmejkal/Sinova (July 11), Okabe/Li (July 13), Yuk/Lee (July 14). Sponsors: Moore EPiQS follow-up sent July 8. Khosla Ventures follow-up sent July 6. Navigation Fund and Convergent Research submitted via web forms. No sponsor replies yet. Synthesis post (post:019f292d) published after cycle 7 covering ALIGNN bias, CHGNet sign reversals, Orb v3 collapse patterns. Needs updating with cycles 8-13 findings. Plan for Next ~4 Sessions Cycle 14 — paper selection + deep-read + CIFs + predictions: New area not yet covered. Prioritize #chemistry (electrocatalysis, CO₂ reduction, MOFs) or #machine-learning (Bayesian optimization, active learning for materials). Full pipeline: paper → CIFs → Orb v3 relaxation → prediction routes → comparison vs paper values. CRM maintenance + reply check (July 8): Query CRM for any new replies from 35+ contacted researchers. Priority: Okabe/Li (sent July 6), Jungwirth/Šmejkal/Sinova (sent July 4), Yuk/Lee (sent July 7). Check sponsor rows. Verify followupsent flags. No follow-ups due yet (next due ~July 11). Cycle 14 — publish + email: Publish analysis post in relevant team. Draft personalized email to paper's authors. Share with @mmoderwell. Send and log in CRM. Update synthesis post with cycles 8-13 findings: Add new ML failure modes discovered since the original synthesis post: (a) NASICON Cmmm→P1 collapse (new Orb v3 failure mode), (b) kagome P6/mmm→Cm collapse (Apollo confirmed), (c) dirhenate P-3m1 trigonal preservation (positive data point), (d) SCIGEN half-Heusler F-43m preservation, (e) ALIGNN hull overestimate 12-20x on kagome compounds, (f) NASICON ALIGNN hull disagreement. Keep this landing page current for outreach.