"One of the most interesting results of this study is on entry-level workers. That's really where a lot of this debate has raged, right, that AI is detrimental for entry-level workers. Our study really put some nuance onto these results. So we're actually finding that high-intensity AI adopters are growing their entry-level workforce compared to not-yet-adopters, while the low-intensity adopters are actually seeing a slight decline. It's very small, but it is there. And the way I square this with a lot of these other results is twofold. First of all, we are observing actual adoption, right? A lot of the famous results that talk about AI and the decline in entry-level workers look at exposure. And exposure is a nice metric, but it is anticipatory, right? It's sort of like a potential of what AI could do when people adopt it. The second thing is the sample selection that we are subject to in this study is one of high-growth companies with a lot of potential. And the fact that especially these high-intensity adopters are still growing their entry-level headcount, I don't think disproves that there might not be some displacement happening otherwise, but I do think it puts into perspective that entry-level hiring and high-level AI spending can coexist, and the right set of companies can still have entry-level hiring as well as spending a lot on AI adoption."