In a recent LinkedIn post, Marty Cagan discusses the profound impact of generative AI on product development, cautioning against its potential to accelerate the creation of ineffective products. Cagan, a prominent voice in product management, highlights a critical shift in the industry’s challenges.
Cagan points out that while generative AI offers unprecedented speed, its current application often leads to an increase in “feature factories,” resulting in more poorly conceived products being shipped faster than ever. This outcome, he suggests, is far from the transformative breakthrough many anticipated.
“Many in the industry have already noticed how AI is being used to turbo charge feature factories, and they are shipping more bad product, faster, than ever before.”
The Output vs. Outcome Dilemma
As Marty Cagan notes, the advent of generative AI is making a crucial distinction increasingly apparent: the difference between mere output and meaningful outcomes. While teams can now produce more features at a breakneck pace, Cagan argues that this surge in output does not necessarily translate into desired business results.
According to Cagan, the bottleneck in product development is no longer the engineering team’s ability to deliver features. Instead, the primary challenge has shifted to the more complex and strategic task of identifying what products or features will actually generate the necessary business outcomes.
“But two things are becoming increasingly obvious: first, this is making very clear the real difference between output and outcomes; and second, the bottleneck is no longer your engineers building the product in delivery; it is figuring out what to build that will generate the necessary outcome.”
Building to Learn vs. Building to Earn
Marty Cagan further elaborates on this strategic shift by drawing a distinction between two fundamental approaches to product development: building to learn and building to earn. He suggests that strong product teams intuitively understand this difference, and it is central to their success.
In Cagan’s view, the speed offered by generative AI can be a dangerous distraction if not guided by a clear understanding of learning objectives. Teams that are merely “building to earn” without a robust learning framework risk churning out features that fail to resonate with users or achieve business goals. Conversely, teams that prioritize “building to learn” use AI as a tool to accelerate their understanding of customer needs and market dynamics, leading to more impactful products.
“What strong product teams understand is that there’s a very important difference between building to learn, and building to earn. This article tries to highlight these differences.”
Cagan’s analysis underscores the need for product leaders to refocus their strategies. The power of generative AI, he implies, lies not just in its ability to accelerate delivery, but in its potential to expedite the learning process, provided teams are equipped with the right mindset and methodologies to discern valuable outcomes from mere output.
📝 About This Content
This article is based on insights shared by Marty Cagan on LinkedIn.
📅 Originally posted on April 16, 2026 | View original post on LinkedIn →