Jim VandeHei says CEOs should delete half their work — and AI is making that possible now

Jim VandeHei's book Simplify is built around a system Axios has been running internally for years: identify the three things you actually need to do to excel at your job, delete the three things you do out of habit that add no value, and offload everything else to AI or outsourcing. Axios says the results produced 70% higher revenue per employee while also making staff happier by eliminating work they didn't want to be doing in the first place.

VandeHei argues the system applies from the CEO down to the newest hire. He frames it as a survival requirement for the AI era, not a productivity nice-to-have. Companies that don't build a lean, focused operating model will accumulate what he calls the "bureaucratic blob" — layers of inertia that make it impossible to keep pace with the velocity of change AI is forcing across every business function.

“We basically forcing me down to the newest employee to focus only on those things we're actually good at and that you want to do and should do, and then deleting the hell out of all the things we do at work we shouldn't do. The results for us internally were magic. We had 70% higher revenue per employee, but more important, employees were happier.”

AI adoption gap

VandeHei estimates that only 1–2% of people use AI deeply enough to understand what it can actually do. The rest see a slightly better search engine that occasionally gets things wrong. That gap feeds fear, and fear gives politicians an opening to propose blunt solutions to a problem that demands precise ones. His worry isn't just competitive (China) but distributional: if the productivity gains concentrate at the top, the political backlash will be severe.

His advice for individuals is pragmatic. People still forming their thinking — college students in particular — should protect time for deep concentration and original thought. People already in the workforce should treat AI as the most powerful power tool ever built and at minimum adopt it passively, the way earlier generations adopted the calculator or the internet. The bigger unlock comes when you invest enough time to train a model on your own style, blind spots, and preferences. Most people don't have that time, which is why businesses are stuck in what he calls a "weird messy middle."

Where AI is actually moving Axios's business

On the revenue side, VandeHei is direct that AI isn't meaningfully changing how reporters write or research. The gains are behind the scenes. He gives a concrete example: one junior employee built a system in 36 hours that automatically converts Axios newsletters into six-to-eight card Instagram and Facebook posts. The feature was live on Instagram within six days, opening new distribution and potential revenue without touching editorial quality.

That pattern, he says, is where Axios sees the real upside — automating the operational scaffolding around journalism so cost can be redirected into the reporting itself.

On automating the CEO role

Asked whether AI agents could effectively replace a CEO's routing-and-alignment function, VandeHei's answer is that the tools are already useful but the core job doesn't disappear. He uses AI today to prototype new products without pulling in his team, and for internal alignment, but argues that hiring well, setting strategy, and keeping people motivated aren't functions a model takes over. His more pointed observation is that AI probably makes it easier to start a company without heavy finance, operations, and marketing overhead — but the result is more companies and new use cases for existing people, not fewer roles. Axios, which has been using AI aggressively for a couple of years, hasn't cut headcount. It has found new uses for its designers and operators that didn't exist before.

Axios's 70% revenue-per-employee advantage is the number that anchors everything else. Whether that holds as the productivity tools become commodity is the question VandeHei doesn't address.