Greg Brockman's courtroom appearance in the Elon Musk vs OpenAI case turned into a masterclass in how not to be a witness. The OpenAI president got cross-examined first in an unusual procedural move, and his performance had serious "well, actually" energy.
Brockman spent most of his time on semantic gymnastics. He'd respond with "I wouldn't characterize it that way" or demand to see written evidence "in context" before confirming he wrote it. When Musk's attorney read evidence aloud, Brockman would correct him for skipping articles like "a" or "the."
The irony is that Brockman's own journal entries have been one of the strongest pieces of evidence for Musk's side. Sometimes the best witness against you is your past self with a diary.
For anyone watching the AI industry's power dynamics play out, this case matters beyond the legal outcome. The testimony reveals how OpenAI's early leadership viewed their mission and commitments, which directly impacts trust in how frontier AI labs operate today.
The courtroom drama also highlights a broader pattern in tech. When founding teams split and billions are at stake, those old Slack messages and journal entries become legal ammunition. Document everything, but remember it might get read aloud in court someday.
