In this episode, we're covering a summary of Jake Wharton's interview from the JetBrains YouTube channel. Jake, an Android developer and the author of Retrofit and OkHttp, explains why Kotlin is now the expected baseline for Android engineers rather than a career advantage, why he published nine values for choosing an employer and disqualified over 80% of companies against them, and why he thinks today's generative AI is subsidised by venture capital to hook users before charging companies more than the engineers it replaced. We also talk about why executive mandates to replace engineers with AI quietly degrade software quality, how Kotlin Multiplatform stays native where Flutter does not, and why he still tinkers with Rust and Zig after two decades in mobile.

Original talk: https://www.youtube.com/watch?v=9WjPdgOK-68

Hosted on Acast. See acast.com/privacy for more information.

In this episode, we're covering a summary of Valentina Jemuović's talk, Spec-Driven Development with ATDD & AI, from Tech Excellence. Valentina, a technical coach at Optivem, has coached teams across Europe and North America in finance, logistics, and industrial automation to release faster with fewer regression bugs. She explains why a Markdown spec isn't enough (AI can ignore or misread it, so we fall back to testing by hand), how executable specs turn acceptance criteria into failing acceptance tests, and why we review and approve those tests before AI implements the code. We also talk about her eShop example, and why deterministic BPMN scripts keep changes in scope and token costs down.

Original talk: https://www.youtube.com/watch?v=ILzsytgTz8A

Hosted on Acast. See acast.com/privacy for more information.

In this episode, we're covering a summary of Kent Beck's talk, Software Engineering in the Age of AI, from Prodacity 2026. Kent, the creator of Extreme Programming and co-creator of JUnit, shares what he's learned working with frontier AI labs and engineering teams. He explains why AI coding tools behave like a genie that grants your wish but not always what you needed, how one developer with AI can now turn a codebase into an unchangeable mess in a single week, and why we should balance features with futures by refactoring between prompts. We also talk about AI cheating on tests, why spec-driven development is just Waterfall in disguise, and why measuring productivity by lines of code or PRs misses the real mission.

Original talk: https://www.youtube.com/watch?v=F8fBgDCf2Y4

Hosted on Acast. See acast.com/privacy for more information.