Jason M. Lemkin argues that AI is rapidly transforming work: many teams resist change (he estimates 20–40%), and companies—even healthy ones like Microsoft—are executing small targeted layoffs (he cites a "3%" Microsoft layoff) to exchange roles rather than shrink net headcount. He highlights that AI already handles roughly 20–40% of support tasks in some vertical SaaS, and that AI tooling (e.g., Cursor) has driven large productivity gains in coding — he cites RevenueCat achieving ~50% increased software productivity. His core claim: AI-driven reskilling is brutal but necessary; gains compound year-over-year and will accelerate layoffs and role shifts.
Questions this video answers
3% (mentioned verbally).
20–40% (speaker's estimate).
Approximately 20–40% (spoken estimate).
40% of all mobile subscriptions (as stated).
About 50% (speaker's claim).
Chapters
So this is. Not only does everyone have to move twice as fast, but it's tough. I don't think 20 to 40% of our teams want to change. And we're going to see. People are like, shocked that Microsoft did a 3% layoff this week. 3% layoff, okay, but those are folks that can't change. They need new people to exchange. We're going to see this trend accelerate. We're going to see more and more layoffs of companies that are doing well. And the Net headcount is not going to go down. They've just got to reskill people. It's brutal. But if you don't do it, you're going to fall behind. You're going to fall behind. And. And if you're still skeptical, hopefully you're not skeptical. I just want to highlight this one thing that's maybe obvious, and I'll tie it all together. If you haven't seen what's happened in coding and support, it's radically changed. If you're not seeing your vertical SaaS, if you're not seeing it in your segment, literally 20 to 40% of support is all done on AI today. And it's great. It's great. It's not perfect. Everyone that is doing support at scale, and there's folks here, you can talk to decagon tomorrow. The CEO's here, which is at the cutting edge, but almost everyone has humans in the loop. Okay, that's not the point. The point is almost 40% of humans are often already replaced with AI. Look at coding. I keep picking on Miguel because he's in the front row, but in Revenuecat, which has 40% of all mobile subscriptions use them, they've increased software productivity about 50%. Using AI tools, using cursor and friends up 50%. 30, 40. That's a lot, though. So I was literally at an investor meeting not too long ago, and the VP of engineering, I'm not sure. I don't know this stuff's really a good idea. I don't know if it's going to work. I'm like, imagine you get 30% gains this year and 30% next year, and it just compounds at an epic rate. So my real point is it's here.
