Self-Harness: Harnesses That Improve Themselves

4 comments

• 7e 27 minutes ago
  Pretty obvious stuff; see Terminator for the conclusion (SkyNet). Or the Matrix. We really need more work on model alignment, trustworthiness, and control.
• behnamoh 1 hour ago
  What else is new? Put it in emacs and let the model improve the harness over time.
• tlarkworthy 12 minutes ago
  [flagged]
• mncharity 1 day ago
  [dead]