Chris Slowe, CTO of Reddit and technical architect behind the sixth most visited website on the internet, believes one of product development’s most trusted tools is quietly doing damage.
A B testing.
In this week’s episode of The Library of Minds, Chris shares how blind optimization can make products worse, especially when you are building for a deeply opinionated community that hates change.
Reddit did not scale by chasing local maxima. It scaled by understanding its users, respecting the culture, and knowing when data should inform decisions and when it should not.
What’s Inside the Episode
How Reddit grew from a scrappy startup into a global platform
What it means to build for users who actively resist change
What tech debt actually signals about your organization
Why A B testing can lead teams in the wrong direction
Why Reddit refuses to label AI generated content
How anonymity creates better conversations
Chris’s worst architecture decision
Inside the first YC batch with Sam Altman, the Twitch founders, and Aaron Swartz
The one tech bet Chris believes never fails
To understand why optimizing for metrics can erode trust and product quality, you have to hear how he connects these decisions over time.
The Core Tension
A/B tests are great at answering small questions.
They are terrible at answering the most important ones.
When teams optimize only for short term engagement, they risk breaking long term trust. On a platform like Reddit, that tradeoff is fatal. The community notices. And once trust is gone, it does not come back.
Chris argues that engineering judgment, cultural intuition, and long term thinking matter more than incremental gains.
Step Inside Chris’s Mind
Chris has created a Digital Mind on Delphi where you can ask him questions directly and hear his answers in his own voice.
Watch the Episode and Talk to Chris’s Delphi
Try asking:
When should teams ignore A/B test results?
How do you balance data with engineering intuition?
What makes a technical decision durable for decades?
