In this episode of Ship Happens, host Per Krogslund talks with Eric Bowman, CTO of King, about how AI is changing the fundamental constraints of software development. As AI makes writing and producing code dramatically faster, Eric argues that the bottleneck is shifting: the harder problem is increasingly deciding what to build, why to build it, and whether you're building the right thing in the first place.
In this episode of Ship Happens, host Per Krogslund talks with Eric Bowman, CTO of King, about how AI is changing the fundamental constraints of software development. As AI makes writing and producing code dramatically faster, Eric argues that the bottleneck is shifting: the harder problem is increasingly deciding what to build, why to build it, and whether you're building the right thing in the first place.
Eric compares AI's impact on software development to the earlier acceleration brought by open source, while exploring what happens when teams can produce software faster than ever. They discuss the risks of the “myth of instant expertise,” how AI can reinforce existing cognitive biases, and why greater coding velocity doesn't necessarily translate into better products.
The conversation also explores how AI could reshape engineering organizations, enabling smaller, more full-stack teams and pushing the platform and domain split higher in the stack. But while AI can accelerate development, Eric argues that operations, incident response, judgment, and understanding context remain deeply human.
Per and Eric dig into the importance of prioritization, flow, and learning loops in an AI-powered development environment. Eric explains why teams need to get better at predicting outcomes, creating psychological safety, and shortening the distance between shipping something and learning whether it worked. He also looks ahead at how software development could change over the next five years, including his prediction that traditional code review may largely fade—and why human creativity could remain one of the hardest things for AI to replicate.
00:00 AI Changes the Bottleneck
00:39 Meet Eric Bowman, CTO of King
01:34 Is Coding Still the Bottleneck?
04:15 AI and the Open Source Acceleration
06:10 Rethinking Engineering Teams
08:50 The New Platform and Operations Divide
10:46 Beyond Productivity: Are We Building the Right Thing?
14:00 Making Software Development Fun Again
15:39 Priorities, Flow, and Shipping What Matters
18:15 Shortening the Time to Learn
22:25 Building Better Learning Loops
26:42 Where Software Goes From Here
27:34 Why Creativity May Stay Human
28:50 Closing Thoughts
The bottleneck is moving. When AI makes producing code easier, deciding what deserves to be built becomes increasingly important.
Velocity isn't the same as value. Teams can ship more software without necessarily creating better products. AI makes good judgment and prioritization even more important.
AI amplifies what already exists. Strong engineering practices can become more powerful with AI, but weak or mediocre practices can also scale faster.
Learning speed matters. Teams that can make predictions, ship, get feedback, and adapt quickly will have an advantage as the cost of producing software continues to fall.
Smaller teams may do more. AI could allow engineers to operate across more of the stack and reduce the need for some organizational layers.
Some work remains fundamentally human. Incident response, judgment, context, and creativity aren't simply eliminated by faster code generation.
The goal isn't to build faster. It's to build better. As AI changes the economics of software development, the teams that can consistently identify and build the right things may have the biggest advantage.
Eric Bowman is the Chief Technology Officer at King, the game development company behind globally recognized titles including Candy Crush. He has spent his career building software organizations and platforms and exploring how engineering teams can work more effectively as technology and development practices evolve.
In this conversation, Eric brings that experience to one of the biggest changes facing software engineering today: what happens when AI dramatically reduces the cost and time required to produce code?