Key Takeaways

  • Software tools and AI models generate more code and internal experiments than ever, but customer willingness to pay has not changed.
  • Scaling raw output turns product teams into software factories that outsource thinking and lose taste.
  • Linear runs specific rituals like Quality Wednesdays and Feature Roasts to keep engineering judgment sharp as AI accelerates delivery.
  • Product teams must evaluate whether new tools help them make better things, rather than simply increasing ticket throughput.

The Software Factory Trap

Every wave of developer tooling promises faster delivery. AI agents, automated scaffolding, and code generators make shipping features easier every month. Saarinen warns that treating software development as an output maximization problem is a dangerous mistake.

As Saarinen put it: “Don't become a software factory where you outsource all the work and thinking to somewhere else and just focus on how efficient can the output be, because the output is not the product.”

When companies measure productivity by volume, they optimize for the wrong scoreboard. They run dozens of A/B tests and merge hundreds of PRs, yet fail to move revenue. Saarinen notes the disconnect: “Customers are not buying the lines of code. They're not buying the experiments or they're not buying a lot of things. So in the end, if you want to sell something or make some good product, you have to make something someone wants.”

When a team outsources product sense to spreadsheets or automated pipelines, it stops developing intuition. Faster execution on bad ideas only produces bad software faster.

Craft Requires Shared Judgment

Building software that people love requires compounding context inside the team. If engineers and designers only act as ticket-movers, nobody develops the taste needed to solve subtle customer problems.

Saarinen asks the question every engineering leader should ask when adopting new tools: “They're more capable. But the more important question is, are we actually making better things with these tools? And are our teams getting better with these tools?”

To preserve craft while moving fast, Linear protects internal rituals designed to force quality discussions. The team runs Quality Wednesdays to fix product rough edges that rarely show up in user tickets. They hold Feature Roasts, where teams critique work in progress before it ships. They also feed raw customer input into AI-driven feedback briefings, ensuring builders read actual user words instead of filtered summaries. These practices force team members to apply critical thinking directly, rather than letting automated pipelines make product choices for them.

What to Do With This

Audit your last two product sprints. Count how many shipped items were built to satisfy an internal throughput metric versus direct customer demand. This week, run a 60-minute Feature Roast on your next release: put the staging build on screen, ban slide decks, and have your lead engineer and designer critique usability flaws together before writing another line of code.