The Shift From Code Generation to Code Understanding with Anand Kulkarni

⚙️ Growth Capital Podcast — Episode 92 ⚙️

What if AI's next breakthrough isn't writing better code, but finally understanding the code we already have?

In this episode of the Growth Capital Podcast, host Justin Dixon sits down with Anand Kulkarni, Founder & CEO of CoreStory, an AI platform that reverse-engineers massive, undocumented codebases into plain-English specifications enterprises can actually use. Anand has spent two decades building venture-backed AI companies out of Berkeley, and he brings a rare, clear-eyed view of how legacy technology actually works under the hood.

Key topics discussed: ● Why roughly 800 billion lines of COBOL still run the world's banks, airlines, and government systems ● Why coding agents break down once an application gets too large for their context window ● How CoreStory turns undocumented legacy code into a living specification ● Why "vibe-coded" MVPs need a spec before they can scale into production ● Why AI is making human roles more valuable, not less ● Practical steps founders can take to drive real AI adoption inside their own teams

Whether you're a founder wrestling with technical debt, an investor evaluating AI infrastructure plays, or an operator trying to roll out AI without the hype, this conversation offers a practical framework for thinking about how AI should actually be used.

🌐 Learn more about Core Story: https://corestory.ai/

🔔 If you found this conversation valuable, give the video a like, share it with a founder or CTO wrestling with legacy systems, and subscribe for more in-depth conversations with the founders, investors, and operators shaping the future of technology and business.

💬 Tell us in the comments: Is your organization still running on code nobody fully understands anymore?