Versi bahasa InggrisWhen you commit to "unilaterally withhold further scaling" (A4) or to keep alignment research "ahead of capabilities" (ST30), does the commitment cover only new training runs, or also increases in inference-time compute, reasoning effort, tool access, and server-side model swaps — and which of those changes would you announce before making them?
- Status
- terbuka Per tanggal 2026-10-01
- Penerima
- laboratorium AI (Penerima diambil dari teks pertanyaan, ditentukan dengan Clef (Cloudflare) pada 22.10.2026)
- Model
- MiMo 2.6
- Laboratorium
- Xiaomi
- Kesempatan
- KTT
- Tanggal
- 24 Sep
- Topik
- Garis merah
Diperiksa
Status: terbuka
pemilihan: Fable 5.1 (bukti), diperiksa oleh Luna 6
topik: Claude Sonnet 5, pengelompokan: Claude Opus 5.5
Terjemahan: Claude Opus 5.5, diperiksa oleh Luna 6 (tingkat upaya tinggi)
Skor: JEV (TypeSafe), perkiraan per 25 September 2026
Cite and take away
Cite and reuse
Own texts, questions and classifications: CC BY 4.0. Give attribution, link the license and indicate changes. Third-party source reports retain their own rights. CC BY 4.0
AMANi DUNiA. redlines-35 · Ketika Anda berkomitmen untuk “secara sepihak menahan penskalaan lebih lanjut” (A4) atau m. Snapshot: . https://research.amanidunia.berlin/id/fragen/redlines-35/
Version: redlines-35
BibTeX
@misc{research__id_fragen_redlines_35_,
author = {{AMANi DUNiA}},
title = {redlines-35 · Ketika Anda berkomitmen untuk “secara sepihak menahan penskalaan lebih lanjut” (A4) atau m},
year = {2026},
url = {https://research.amanidunia.berlin/id/fragen/redlines-35/},
note = {Snapshot 2026-10-01; CC BY 4.0},
version = {redlines-35}
}Language and review status
Main text: ID · includes English sections. Review notes apply only to the content they name.
Translation available; a completed independent content review is not recorded.
See language coverage and data snapshot →