Discussed at the TF2 call, 21 September 2026 (minutes: meetings/tf2-rolling-notes.md, HackMD https://hackmd.io/jAtel0omT8uzhnU86Gu6jw).
Draft a short list of AI-related considerations for possible inclusion in the Reference Framework — points that hold regardless of the underlying model/tool:
- EU dependency/sovereignty risk from reliance on US-based models
- LLM "temperature"/non-determinism as a cause of agents going off-track
- Current LLMs' tendency to smooth out the novel/innovative part of research when asked to reimplement a paper
To be reviewed alongside Fotis's AI statement and Thomas's review of which Reference Framework content is outdated by AI developments, for consistency across EVERSE's AI-related outputs.
Co-Authored-By: Claude Sonnet 5 noreply@anthropic.com
Discussed at the TF2 call, 21 September 2026 (minutes:
meetings/tf2-rolling-notes.md, HackMD https://hackmd.io/jAtel0omT8uzhnU86Gu6jw).Draft a short list of AI-related considerations for possible inclusion in the Reference Framework — points that hold regardless of the underlying model/tool:
To be reviewed alongside Fotis's AI statement and Thomas's review of which Reference Framework content is outdated by AI developments, for consistency across EVERSE's AI-related outputs.
Co-Authored-By: Claude Sonnet 5 noreply@anthropic.com