Recommended Reads

APRIL 2026

Ethan Mollick (via The Economist) on why companies should embrace the weirdness of AI and avoid the temptation to treat it the way they've historically treated enterprise software. Why are companies "slotting [AI] into existing processes, assigning KPIs, and handing it to IT"? The instinct is understandable. Mollick argues this comes from training and pattern recognition - it's a new technology, so let's apply the same playbook we've used before.

It also comes down to risk tolerance.

Many established companies, understandably and by design, have lower tolerance for risk. They are less likely to make big moves to fundamentally change how they operate and go to market. The rub is that by not doing so, they fall into Clayton Christensen’s Innovators Dilemma (another favorite read). Meanwhile, AI-native competitors have a higher risk threshold and are doing the opposite, experimenting aggressively, rebuilding workflows, and challenging assumptions from day one.

So what should incumbents do?

Mollick outlines a three-prong approach: (1) leadership, (2) crowd, and (3) lab. To borrow a phrase from improv - yes, AND leadership must consider big shifts like:

• Fundamentally changing how product development operates

• Fully rebuilding software to modern, AI-native software

• Training and incentivizing the workforce on vibe coding and creating agents across all functions

• Determining whether there is a completely different delivery and/or monetization model to adopt