এই এপিসোড সম্বন্ধে
Lovely listener, share with us your thoughts!
9 out of 10 professional developers now use AI coding agents every week. This year, the best agent in a benchmark built to look like real work passed fewer than 15% of its checks. Both are true at once.
A solo episode of Curiosity Code on which engineering work a financial firm should hand to an AI-driven team and which it should keep with conventional developers. My company, Productera, sells both, so I make money whichever way you decide. My answer: writing the code got cheap, being accountable for it did not. Hand over the support work, new builds a senior engineer can read end to end, reading old code and small bounded changes. Keep long-lived core systems, regulated and security code, and anything that needs an independent sign-off. Europe's DORA already draws that line.
The three questions for your backlog: Will this code still be running in three years? Does it touch a regulation, money movement or security? And is there someone senior who didn't write it, with time to read all of it? Short-lived, low stakes and reviewable goes to an AI-driven team. Long-lived, regulated, or not properly reviewable stays with conventional developers.
Chapters
0:00 90% use AI agents, the best passes under 15%
0:41 Why this is a solo episode, and where I stand
2:16 Writing code got cheap, accountability didn't
2:48 What I'd hand to an AI-driven team
4:33 What I'd keep with conventional developers
6:36 The rule that already draws the line
7:35 Three questions for your backlog
Sources
JetBrains, AI coding agent adoption in 2026 (survey of 15,000+ developers, August 2026): https://blog.jetbrains.com/research/2026/08/ai-coding-agent-adoption-2026/
Orlanski et al., SlopCodeBench, University of Wisconsin-Madison (arXiv, May 2026): https://arxiv.org/abs/2603.24755
Brett Codes, "Real risks of using AI as a dev team" (September 2026): https://youtu.be/g0jjNoR09Tc
Fred Brooks, The Mythical Man-Month (1975): https://en.wikipedia.org/wiki/The_Mythical_Man-Month
Morgan Stanley's DevGen.AI, 9 million lines and 280,000 hours (Business Insider, June 2025; first reported by the Wall Street Journal): https://www.businessinsider.nl/a-team-of-engineers-saved-morgan-stanley-more-than-280000-hours-this-year-the-bank-says-their-tool-wont-take-jobs/
Veracode, GenAI Code Security Report (July 2025): https://www.veracode.com/blog/genai-code-security-report/
Faros AI, The AI Productivity Paradox (July 2025): https://www.faros.ai/blog/ai-software-engineering
DORA, Commission Delegated Regulation (EU) 2024/1774, Article 17 on ICT change management: https://www.springlex.eu/en/packages/dora/rts-rmf-regulation/article-17/
Federal Reserve SR 26-2, model risk management guidance (April 2026): https://www.federalreserve.gov/supervisionreg/srletters/SR2602a1.pdf
🎙️ Subscribe for weekly episodes on fintech, AI, and innovation
🌐 Website: https://curiositycodepodcast.com/
📱 Follow on LinkedIn: https://www.linkedin.com/in/alexkhomyakov/
🎧 Listen on Spotify: https://open.spotify.com/show/4ePKf2qCLzcKATS7o6w2U5
🍎 Apple Podcasts: https://podcasts.apple.com/podcast/id1797349967
Disclosure: Productera supplies both AI-driven teams and conventional developers to financial firms.
Music: "Soundbed" by Blue Saga, via Epidemic Sound.
🎙️ Subscribe for weekly episodes on fintech, AI, and innovation
🌐 Website: https://curiositycodepodcast.com/
📱 Follow on LinkedIn: https://www.linkedin.com/in/akhomyakov/
🎧 Listen on Spotify: https://open.spotify.com/show/curiosity-code
🍎 Apple Podcasts: https://podcasts.apple.com/podcast/the-curiosity-code-podcast/id1748596092
