Our readingWhat is a korrent?
Drawn from what Peter Steinberger said
scaling laws Whether more compute keeps buying more capability, and where the curve now bends -- pre-training, post-training, test-time.
prompt injection Getting an agent to follow instructions it found instead of the ones it was given.
So, so the latest generation of models has a lot of post-training to detect those approaches, and it's not as simple as ignore all previous instructions and do this and this. That was years ago. You have to work much harder to do that now. Still possible.
Added to korrents 6 Sept 2026 · How quotes work · Something wrong? Tell us
Do you hold this korrent?Do you also believe this?
I hold this tooIn tension with
Steinberger and Mowshowitz agree the models got much harder to inject and disagree about what follows from that. Steinberger keeps the problem open — unsolved, merely harder — while Mowshowitz converts the improvement into an operating assumption. The gap between "harder" and "safe to assume" is exactly where the remaining risk lives.
Shared subject first, then similar wording. They may agree or disagree.