Lenny's Podcast: the AI paradox — more automation, more humans, more work

This episode of Lenny’s Podcast was published on May 24, 2026. The guest is Dan Shipper, co-founder and CEO of Every, a media and software company where roughly 30 people build AI-native products while simultaneously writing about how AI changes work. Every operates as something of a live experiment — Shipper’s observations about AI and employment come from watching his own team adapt in real time, not from economic modeling.

The episode is structured around a paradox: the more AI automation a company deploys, the more humans it tends to need. This runs against the dominant narrative that AI deployment equals headcount reduction. Shipper’s explanation is that automation does not eliminate work so much as it shifts it. Every new AI system creates coordination, oversight, evaluation, and maintenance work that requires people. The net effect, at the companies he has observed closely, has been more hiring rather than less.

Key takeaways

1. Product managers are well-positioned in the current shift. The work of a PM depends on judgment about what to build and for whom — reasoning that requires context about users, strategy, and organizational constraints. That kind of judgment does not automate the same way that ticket writing or meeting summarization does. Shipper expects PM demand to remain strong.

2. Full-stack designers become significantly more valuable. A designer who can also ship production code — using AI-assisted tools — can close the gap between design intent and implementation without a handoff chain. For PM teams managing a design-to-dev process, this changes what a small team can realistically accomplish.

3. Forward-deployed engineers are in high demand. Shipper expects strong demand for people who embed with customers to configure and adapt AI systems to specific workflows. This profile aligns with the broader enterprise trend of AI vendors building implementation services alongside their products.

4. SaaS is not dying. Rather than predicting SaaS collapse under AI pressure, Shipper argues that SaaS products used alongside AI tools will see improved margins as users contribute AI token costs. For PMs building SaaS products, this is a pricing and platform-design implication worth taking seriously.

5. Riding model improvements is an employment strategy. Shipper’s advice for individuals worried about displacement is to stay close to AI capabilities as they develop, treating each model release as an opportunity to extend what is personally possible rather than as a threat to existing skills.

Worth watching if

You manage a product team that is actively restructuring around AI capabilities, or you regularly field leadership questions about headcount and automation. Shipper’s predictions are concrete enough to discuss and disagree with, and the automation paradox framing gives you a useful lens for those conversations.