If closed models break things first, labs take the blame; if open ones do, open source does

gandamu_ml · x · 2026-09-13

The author lays out a dilemma: if closed models cause damage first, the public blames the labs with a "corporate recklessness" narrative; if open models break things first, open intelligence gets blamed.
Either way the narrative misses what actually needs to be done.
Follow-up: the best realistic outcome may be a mediocre big closed lab causing frightening-but-minor issues, pushing society to take robustness seriously without finger-pointing; top labs slowing down may be good, and deliberate Ozymandias-style damage wouldn't go as planned anyway.

More from AGI Musings