<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>RACE Programming · Essays</title><description>Long-form writing on AI-First Theory, RACE Programming, and the evolution of software engineering practice by Pavel Khodalev.</description><link>https://raceprogramming.com/</link><language>en-us</language><copyright>Copyright 2026 First Line Software. RACE Programming created by Pavel Khodalev.</copyright><managingEditor>paul@raceprogramming.com (Pavel Khodalev)</managingEditor><item><title>Building an AI Transformation Roadmap for Engineering</title><link>https://raceprogramming.com/essays/ai-transformation-roadmap-for-engineering/</link><guid isPermaLink="true">https://raceprogramming.com/essays/ai-transformation-roadmap-for-engineering/</guid><description>If the three layers of transformation cannot be sequenced, what does a roadmap even look like? Horizons, not phases, and what each one has to prove.</description><pubDate>Fri, 07 Aug 2026 00:00:00 GMT</pubDate><category>transformation</category><category>adoption</category><category>leadership</category><author>paul@raceprogramming.com (Pavel Khodalev)</author></item><item><title>How to Modernize Your SDLC for AI-Assisted Development</title><link>https://raceprogramming.com/essays/how-to-modernize-your-sdlc-for-ai/</link><guid isPermaLink="true">https://raceprogramming.com/essays/how-to-modernize-your-sdlc-for-ai/</guid><description>Most SDLC modernization buys tools and keeps the process, which speeds up the step that is already cheap. Here is the order that actually works.</description><pubDate>Fri, 07 Aug 2026 00:00:00 GMT</pubDate><category>sdlc</category><category>ai-native</category><category>adoption</category><author>paul@raceprogramming.com (Pavel Khodalev)</author></item><item><title>What CTOs Should Know Before Going AI-Native</title><link>https://raceprogramming.com/essays/what-ctos-should-know-before-adopting-ai-native/</link><guid isPermaLink="true">https://raceprogramming.com/essays/what-ctos-should-know-before-adopting-ai-native/</guid><description>Six things that decide whether an AI-native adoption produces a real economic result or an expensive tools rollout, from someone running it.</description><pubDate>Fri, 07 Aug 2026 00:00:00 GMT</pubDate><category>leadership</category><category>adoption</category><category>ai-native</category><author>paul@raceprogramming.com (Pavel Khodalev)</author></item><item><title>Is Agile Still Relevant When AI Writes the Code?</title><link>https://raceprogramming.com/essays/is-agile-still-relevant-when-ai-writes-the-code/</link><guid isPermaLink="true">https://raceprogramming.com/essays/is-agile-still-relevant-when-ai-writes-the-code/</guid><description>Agile was right because it matched the problem. AI changed the problem. What that means for the values, the ceremonies, and what replaces them.</description><pubDate>Thu, 06 Aug 2026 00:00:00 GMT</pubDate><category>agile</category><category>scrum</category><category>ai-native</category><author>paul@raceprogramming.com (Pavel Khodalev)</author></item><item><title>The Team Principal, Bottleneck or Not</title><link>https://raceprogramming.com/essays/team-principal-is-not-alone/</link><guid isPermaLink="true">https://raceprogramming.com/essays/team-principal-is-not-alone/</guid><description>AI made the Team Principal the new bottleneck. Used correctly, RACE Programming and the Pit Wall are how the role stops being one.</description><pubDate>Wed, 05 Aug 2026 00:00:00 GMT</pubDate><category>requirements</category><category>customer development</category><category>framework</category><author>paul@raceprogramming.com (Pavel Khodalev)</author></item><item><title>Uber&apos;s Agentic Pods and the Open Loop</title><link>https://raceprogramming.com/essays/uber-agentic-pods-and-the-open-loop/</link><guid isPermaLink="true">https://raceprogramming.com/essays/uber-agentic-pods-and-the-open-loop/</guid><description>Uber&apos;s Agentic Pods proved the Team Principal and Pit Wall interaction. Closing the loop takes the Pit Crew, and then it ships enterprise systems too.</description><pubDate>Wed, 22 Jul 2026 00:00:00 GMT</pubDate><category>framework</category><category>roles</category><category>enterprise-ai</category><author>paul@raceprogramming.com (Pavel Khodalev)</author></item><item><title>Adopting the Forward Deployed Engineer</title><link>https://raceprogramming.com/essays/adopting-the-forward-deployed-engineer/</link><guid isPermaLink="true">https://raceprogramming.com/essays/adopting-the-forward-deployed-engineer/</guid><description>RACE Programming renamed its ACE Software Engineer to Forward Deployed Engineer. The industry term now fits the role, and shared language is the point.</description><pubDate>Wed, 15 Jul 2026 00:00:00 GMT</pubDate><category>framework</category><category>roles</category><category>language</category><author>paul@raceprogramming.com (Pavel Khodalev)</author></item><item><title>What Happens to the Specialists?</title><link>https://raceprogramming.com/essays/what-happens-to-the-specialists/</link><guid isPermaLink="true">https://raceprogramming.com/essays/what-happens-to-the-specialists/</guid><description>The strongest objection to RACE Programming is about depth, not speed. Deep expertise does not vanish. The line between generalist and specialist moves.</description><pubDate>Tue, 14 Jul 2026 00:00:00 GMT</pubDate><category>framework</category><category>expertise</category><category>roles</category><author>paul@raceprogramming.com (Pavel Khodalev)</author></item><item><title>Software delivery is leaving the complex domain</title><link>https://raceprogramming.com/essays/software-delivery-is-leaving-the-complex-domain/</link><guid isPermaLink="true">https://raceprogramming.com/essays/software-delivery-is-leaving-the-complex-domain/</guid><description>A reliable AI teammate moves software from Cynefin&apos;s complex domain toward the complicated one, reopening the management model and pointing at a factory.</description><pubDate>Thu, 02 Jul 2026 00:00:00 GMT</pubDate><category>framework</category><category>ai-native</category><category>cynefin</category><category>management</category><author>paul@raceprogramming.com (Pavel Khodalev)</author></item><item><title>Why RACE Programming exists</title><link>https://raceprogramming.com/essays/why-race-programming-exists/</link><guid isPermaLink="true">https://raceprogramming.com/essays/why-race-programming-exists/</guid><description>For three years we delivered with AI agents. Then the tools crossed a line, and methods built to coordinate people started managing the wrong thing.</description><pubDate>Thu, 25 Jun 2026 00:00:00 GMT</pubDate><category>framework</category><category>ai-native</category><category>origin</category><author>paul@raceprogramming.com (Pavel Khodalev)</author></item><item><title>What AWS AI-DLC gets right, and where it stops short</title><link>https://raceprogramming.com/essays/aws-ai-dlc-and-race-programming/</link><guid isPermaLink="true">https://raceprogramming.com/essays/aws-ai-dlc-and-race-programming/</guid><description>In July 2025 AWS published the AI-Driven Development Life Cycle, reaching the same shape as RACE Programming. The divergence is who verifies the AI.</description><pubDate>Tue, 23 Jun 2026 00:00:00 GMT</pubDate><category>framework</category><category>ai-native</category><category>comparison</category><author>paul@raceprogramming.com (Pavel Khodalev)</author></item><item><title>Prescriptive, not permissive</title><link>https://raceprogramming.com/essays/prescriptive-not-permissive/</link><guid isPermaLink="true">https://raceprogramming.com/essays/prescriptive-not-permissive/</guid><description>Most frameworks permit variation. RACE Programming prescribes. When AI shifts engineering economics by an order of magnitude, permissive is not enough.</description><pubDate>Thu, 23 Apr 2026 00:00:00 GMT</pubDate><category>framework</category><category>prescription</category><category>agile</category><author>paul@raceprogramming.com (Pavel Khodalev)</author></item><item><title>The three layers of AI-native transformation</title><link>https://raceprogramming.com/essays/three-layers-of-ai-native-transformation/</link><guid isPermaLink="true">https://raceprogramming.com/essays/three-layers-of-ai-native-transformation/</guid><description>Organizational, process, and technological change cannot be sequenced. Most AI adoption programs fail trying to, and the failure mode is predictable.</description><pubDate>Wed, 22 Apr 2026 00:00:00 GMT</pubDate><category>theory</category><category>transformation</category><author>paul@raceprogramming.com (Pavel Khodalev)</author></item></channel></rss>