Code Faster, Review Smarter

How do we ship code faster without sacrificing quality or accountability? Greg Foster, co-founder and CTO at Graphite, joins the show to unpack how AI is reshaping code reviews, developer workflows, and the very definition of software engineering. From AI-assisted reviews to the challenge of maintaining context in a world of auto-generated code, Greg shares hard-won insights from the front lines of dev tools innovation. If you care about shipping fast, staying secure, and evolving your engineering org for what’s next — this one’s for you.

Key Takeaways

• Code review is becoming more about collaboration and less about bug catching

• AI-generated code introduces a new challenge: how engineers maintain context without writing the code themselves

• Developer experience is shifting toward orchestration, not just authorship — prompting, reviewing, shipping, and owning

• Stack-based workflows are essential for speed, safety, and parallel progress in an AI-assisted world

• Even with AI in the loop, human accountability — especially for security and architecture — remains critical

Timestamped Highlights

2:10 – Why Graphite calls itself “code review for the age of AI”

4:50 – What code review really means today (hint: it's not just about bugs)

8:40 – The hidden cost of losing context when you’re not writing the code

12:05 – How the developer experience is evolving with AI-generated code

16:10 – Is tech debt still a problem if code becomes disposable?

21:00 – Inner vs. outer loops of development — and why the bottleneck is shifting

26:10 – Why we hold AI to a higher standard than human engineers

Quote of the Episode

“We used to get context for free — just by writing the code. But in a world of AI code gen, we’re going to need new ways to absorb and maintain that context.” – Greg Foster

Resources Mentioned

Graphite: https://graphite.dev

Greg on LinkedIn: https://www.linkedin.com/in/gregmfoster

Email Greg: [email protected]

Pro Tips

Stack your PRs to keep shipping fast and safely. Whether it’s AI or human writing the code, small, parallelized changes are easier to review, test, and deploy — especially when you're operating at high velocity.

Call to Action

Enjoyed this episode? Share it with a fellow engineer, follow the show, and leave a review on Apple Podcasts or Spotify. For more insights like this, connect with us on LinkedIn or subscribe to our newsletter.