What Greg Brockman said, newest first. Changes are found by AI and checked before they show; every quote opens at its moment.
2026-09-14 · Bloomberg Podcasts · 42:59
“I actually think that this is a very optimistic story as well, because what it means is that these graders are now like, how do you judge what good writing is? The answer is you use an eye to evaluate if you have good writing.”He believes improving AI capability enables much more reliable graders to prevent reward hacking.
2026-09-14 · CrowdStrike · 1:15:41
“for us a lot of the way that we make development progress is by finding bugs right in these graders and kind of the the rewards that you give to your models very very important”He says finding bugs in graders and rewards is a key way development progress happens, and that train-of-thought monitoring during evaluation and development is now needed.
Blumify can make mistakes. Please check important info.