Resource allocation beats feature prioritization

Instead of management, senior management refereeing every item on a roadmap, they're refereeing which teams have how many resources, which is more of like a once or twice or three times a year decision versus refereeing everything on the product roadmap. And then all the resource contention issues, that's a daily issue.

AI research progress is bottlenecked by management, research management. It's because you have constrained set of compute and you need to allocate the compute to the research paths that you feel the most convinced about.

We now have what I call the fast thinking group, which officially is called AI platform, but it really means we want to just ship a bunch of new capabilities on a near weekly basis.

It's also 10x more painful than you think to cut mild successes than anything else, and just organizationally painful. And there's teams that have whole roadmaps, and it's a really painful experience, so you have to think really hard before you kick those off.