About
Hi, Spring fans! In this episode, Josh sits down with longtime friend and JetBrains developer Anton Arkhipov for a fascinating look at how AI is reshaping the way software gets built. From the halls of Devoxx to the future of IntelliJ IDEA, the conversation explores how AI has moved from being a novelty to becoming a core part of modern development workflows.
Anton shares how JetBrains is experimenting with a new “agent-first” way of working, where chat becomes the central interface and AI agents can help plan features, generate code, run tests, and even debug problems. They dive into JetBrains AIR, the company’s experimental environment for orchestrating multiple agents, and discuss how it blends with the familiar IntelliJ experience rather than replacing it.
The discussion gets especially exciting when Anton demonstrates how specs, rules, and reusable skills can guide agents to build software more consistently. He also shows how IntelliJ’s MCP and ACP integrations can give AI agents access to powerful IDE tools, including refactoring and debugging capabilities. One standout moment is a live example where an agent diagnoses a failing test, uses lock points and debugger support, finds the root cause, and fixes the issue automatically.
If you’re curious about where AI-assisted development is heading—and how much of the coding, testing, and debugging loop can now be shared with intelligent tools—this episode offers a compelling glimpse into the near future of software engineering.
Anton shares how JetBrains is experimenting with a new “agent-first” way of working, where chat becomes the central interface and AI agents can help plan features, generate code, run tests, and even debug problems. They dive into JetBrains AIR, the company’s experimental environment for orchestrating multiple agents, and discuss how it blends with the familiar IntelliJ experience rather than replacing it.
The discussion gets especially exciting when Anton demonstrates how specs, rules, and reusable skills can guide agents to build software more consistently. He also shows how IntelliJ’s MCP and ACP integrations can give AI agents access to powerful IDE tools, including refactoring and debugging capabilities. One standout moment is a live example where an agent diagnoses a failing test, uses lock points and debugger support, finds the root cause, and fixes the issue automatically.
If you’re curious about where AI-assisted development is heading—and how much of the coding, testing, and debugging loop can now be shared with intelligent tools—this episode offers a compelling glimpse into the near future of software engineering.