podcast · Organized Programming · #92
How AI Will Change Software Development: Programmers, Teams, and IT Companies
September 13, 2026 · Organized Programming | Kirill Mokevnin · #92 · 1 h 53 min
Kirill Mokevnin and Alexander Polomodov discuss how agents reshape the whole development system: team composition, enterprise platforms, knowledge, business value, and junior training. Alexander draws on his experience introducing AI at T-Bank. A small team supported by agents is a direction of change, rather than an outcome already achieved in every company.
Host and guest
The conversation at a glance
- 01Engineering roles remain, while one person with agents can cover more stages. The aim is to reduce handoffs and carry a task through to an outcome.
- 02Model subscriptions are only a starting point: teams need shared tools, data access, and someone responsible for changing each product area's workflow.
- 03Specifications, ADRs, and plan reviews clarify the task before code generation. Quality requirements need automated checks behind them.
- 04Agents need usable internal platforms and knowledge. Incompatible forks and implicit rules increase integration costs.
- 05AI adoption, time savings, and business value are different measurement layers. Additional tasks help only when they solve a worthwhile problem.
- 06Working code no longer proves a junior engineer has learned. Review task framing, the work process, and understanding of the result; complex legacy systems still demand investigation.