AI Boosts Productivity, but Not in Production with Nathen Harvey

Why It Matters

The divergence between perceived developer productivity and production stability signals risks for companies scaling AI-assisted development: faster delivery can raise operational costs, customer disruption, and technical debt without stronger ownership and production-readiness processes. Organizations must invest in end-to-end responsibility and quality controls to capture the true business value of AI.

Summary

A 2025 study finds that developers report higher individual productivity as they adopt AI tools, and teams are pushing more software changes into delivery pipelines. However, the surge in throughput has been accompanied by increased delivery instability—more rollbacks and hot fixes—indicating gains in coding output don’t reliably translate into production quality. The mismatch appears tied to limited end-to-end ownership: developers feel more productive while working with AI, but that does not guarantee durable improvements once software reaches production. The result is modest production gains despite headline-grabbing productivity claims.

Original Description

AI is making developers feel more productive than ever, but DORA's 2025 research shows that higher delivery throughput is paired with higher instability, more rollbacks, and more emergency hotfixes. Here's DORA's Nathen Harvey on why the productivity boost developers report doesn't automatically follow your code all the way to production. #shorts

Follow O'Reilly on:

LinkedIn: https://www.linkedin.com/company/oreilly/

Facebook: http://facebook.com/OReilly

Instagram: https://www.instagram.com/oreillymedia

BlueSky: https://bsky.app/profile/oreilly.bsky.social

TikTok: https://www.tiktok.com/@oreillymedia