On this episode of Data Driven, we venture into the ever-shifting landscape of software engineering, AI-assisted coding, and the sometimes chaotic future of development teams with special guest Thanos Diacakis—no, not the Marvel villain, but a powerhouse in software leadership whose career spans scrappy startups to tech giants like Uber.
Hosts Frank La Vigne, Andy Leonard, and resident AI BAILeY welcome Thanos for an insightful, witty, and sometimes philosophical conversation about the promise and perils of AI in software engineering. From the rise (and risks) of “vibe coding” to the changing definition of good engineering, Thanos shares firsthand experiences, pitfalls to avoid, and why solid fundamentals matter now more than ever—even with AI churning out code at lightning speed.
We also explore how AI is reshaping workflows for both solo developers and large teams, the difference between local and global efficiency, and how bottlenecks are shifting in this new era. Thanos breaks down his “seven mental models” for effective engineering, demystifies the right way to work with AI tools (instead of letting them run wild), and explains why junior devs still deserve a seat at the table.
If you’re curious about the real impact of AI-assisted coding, worried about the next wave of software “apocalypse,” or just eager for practical strategies to keep your teams productive and sane, this episode is packed with the wisdom, humor, and actionable takeaways you’ll want to hear. Buckle up as we dive deep into the code base—this is Data Driven, where the only thing disappearing is inefficiency.
Links
Time Stamps
00:00 "Exploring AI Coding with Thanos"
05:20 Revolutionizing Productivity with Code
08:18 Cross-Platform Utility Development
12:14 Database Costs Drop, Demand Surges
14:43 Unauthorized Copying of Software Manuals
17:54 "Chunked Narratives Enhance LLM Performance"
19:43 Grox Impressive Despite Incompletion
24:40 Reviving a Startup with AI
28:53 Bottleneck Identification and Resolution Strategy
29:37 Incremental Improvements Drive Success
34:40 Agile's Evolution: Remembering Its Roots
38:41 Balancing Features and Tech Debt
42:50 Software Estimations: A Management Trap
45:55 Managing Project Time Expectations
47:47 Trust Building Through Incremental Delivery
51:23 Language Models: Human Cognition Parallel?
56:04 "Exploring Intelligence: Brains vs. AI"
58:06 "Data Driven Episode Conclusion"
