It’d be really nice to see research in this area from somewhere without a financial interest in hyping AI.

That incentive doesn’t invalidate research, but AI results are so easy to nudge in any direction that it’s hard to ignore.