> Markdown version of [/videos/1365-wearedevelopers-live-the-weekly-developer-show-with-chris-heilmann-and-daniel-cranney?t=1364](https://www.wearedevelopers.com/videos/1365-wearedevelopers-live-the-weekly-developer-show-with-chris-heilmann-and-daniel-cranney?t=1364). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

#  WeAreDevelopers LIVE - the weekly developer show with Chris Heilmann and Daniel Cranney

Microsoft's Julia Cardi warns that AI assistants create a hidden productivity tax for senior developers. Discover why, and see how multi-agent frameworks are successfully reverse-engineering legacy COBOL codebases.

- **Speakers:** [Chris Heilmann](https://www.wearedevelopers.com/@chris-heilmann), [Daniel Cranney](https://www.wearedevelopers.com/@daniel-cranney), [Julia Kordick](https://www.wearedevelopers.com/@julia-kordick)
- **Event:** WeAreDevelopers LIVE
- **Published:** August 6, 2025
- **Duration:** 59:27
- **URL:** https://www.wearedevelopers.com/videos/1365-wearedevelopers-live-the-weekly-developer-show-with-chris-heilmann-and-daniel-cranney

## Summary

Microsoft Global Black Belt Julia Cardi joins the show to dissect the evolving landscape of developer productivity, bridging the gap between cutting-edge AI code generation and the pressing reality of legacy system modernization. The conversation uncovers the hidden "productivity tax" of AI assistants like GitHub Copilot and Cursor. While these tools make rapid "vibe coding" accessible, they frequently produce code that requires more time to debug and intercept than if a developer had written it from scratch. This dynamic is fundamentally shifting engineering team workflows; junior developers increasingly rely on AI to generate initial drafts, which inadvertently shifts the heavy lifting onto senior developers who must safeguard the architecture and review massive, potentially hallucinated pull requests.

Rather than blindly trusting large language models with production code, the speakers emphasize the necessity of deterministic security measurements to prevent vulnerabilities—a lesson highlighted by recent AI-driven application outages. The discussion challenges the industry's default belief that generating more lines of code equates to higher productivity, advocating instead for using AI time-savings to let engineers learn language fundamentals and practice thoughtful system design.

This critical lens on AI capabilities is then applied to a looming industry threat: transitioning COBOL mainframes to the cloud as the original maintainers reach retirement age. Because legacy code is heavily tied to specific hardware architectures and lacks the public open-source repositories needed to train LLMs, off-the-shelf agent models struggle to translate it. To solve this, Cardi details her work building a multi-agent framework utilizing Semantic Kernel. Rather than attempting a naive line-by-line translation, this framework sequentially reverse-engineers undocumented COBOL business logic, separates out non-functional hardware requirements, and prepares the logic for cloud-native Java environments.

**Keywords:** ai productivity tax, github copilot limitations, vibe coding challenges, legacy code modernization, cobol to cloud migration, mainframe infrastructure, multi-agent semantic kernel, reverse-engineering business logic, deterministic code security, llm training data scarcity, cursor configuration files, developer productivity metrics, software architecture design, tech stack migration strategies

## Chapters

1. **Moderating developer conferences and preparing for stage recordings** (00:00) — How updating speaker profiles and considering the video recording improves event quality and digital reach.
1. **Navigating technical clarity as a global black belt** (03:44) — How subject matter experts support strategic projects and evaluate emerging AI solutions for large enterprise customers.
1. **Integrating GitHub and Microsoft enterprise sales and product teams** (05:56) — How previously separated product organizations are merging to deliver cohesive enterprise tools and unified sales motions.
1. **Evaluating tech stack migrations driven by industry hype** (08:54) — Why technology-driven language choices often fail without the proper organizational alignment and available technical talent pool.
1. **Managing the hidden productivity tax of AI code generation** (13:22) — The challenges of intercepting AI models mid-generation and correcting hallucinated dependencies before they clutter a repository.
1. **Enforcing engineering guidelines through AI system prompts and instructions** (18:20) — How establishing coding rules and boundaries for AI requires heavier upfront preparation before actual development begins.
1. **Preventing application outages caused by unvetted AI generated code** (22:44) — Why deterministic security measurements must be enforced rather than implicitly trusting automated code generation tools.
1. **Balancing AI tool mandates with developer trust and productivity** (26:46) — How the forced adoption of coding assistants shifts senior engineering workloads away from architecture and toward tedious debugging.
1. **Reinvesting saved time into developer learning and technical training** (30:31) — Why productivity gains from AI should fund skill development rather than merely extracting a higher volume of unoptimized code.
1. **Building cross-functional engineering teams instead of full-stack developers** (33:46) — How specialized experts collaborating effectively build better products than individuals stretched thinly across entire systems.
1. **Understanding the enduring enterprise reliance on legacy JavaScript frameworks** (35:38) — Why older abstraction libraries like jQuery remain heavily utilized to maintain stable, long-running enterprise web applications.
1. **Customizing code editor themes and visual focus environments** (38:53) — How extreme visual complexity in editor features impacts developer focus and neurodivergent accessibility during deep work.
1. **Managing open source side projects and dormant domain portfolios** (41:50) — The realities and cognitive overload of returning to unmaintained personal repositories after long periods of inactivity.
1. **Modernizing legacy COBOL mainframe systems using AI agents** (44:36) — How deploying a multi-agent framework attempts to reverse-engineer aging enterprise business logic out of unmaintained mainframes.
1. **Overcoming training data scarcity for mainframe modernization models** (51:12) — The difficulty of preserving hardware-specific non-functional requirements when migrating logic from closed mainframes to cloud architecture.
1. **Designing complex software architecture in the era of AI** (54:40) — Why defining accurate technical specifications and abstract architecture remains a critical human skill while automating boilerplate code.
1. **Participating in open source mainframe modernization projects** (57:35) — Where developers can locate sample Azure code repositories and contribute to ongoing enterprise multi-agent framework development.

## Related Moments

- [Shifting developer workloads and realistic AI productivity gains](https://www.wearedevelopers.com/videos/1830-wearedevelopers-live-speculaitions) (from "WeAreDevelopers LIVE - SpeculAItions")
- [Navigating technical debt generation in the era of AI](https://www.wearedevelopers.com/videos/1342-your-code-as-a-crime-scene) (from "Your Code as a Crime Scene")
- [Motivations for adopting AI to enhance developer productivity](https://www.wearedevelopers.com/videos/1266-navigating-the-ai-revolution-in-software-development) (from "Navigating the AI Revolution in Software Development")
- [Using AI copilots to explain and debug legacy codebases](https://www.wearedevelopers.com/videos/1302-wearedevelopers-live-dishonest-charts-britcss-debugging-with-ai) (from "WeAreDevelopers LIVE - Dishonest Charts, BritCSS, Debugging with AI")
- [Balancing developer autonomy with the adoption of coding agents](https://www.wearedevelopers.com/videos/100198-the-last-mile-of-ai-from-prototype-to-production) (from "The Last Mile of AI: From Prototype to Production")
- [Security integration and AI skepticism in developer tooling](https://www.wearedevelopers.com/videos/1830-wearedevelopers-live-speculaitions) (from "WeAreDevelopers LIVE - SpeculAItions")

## Related Articles

- [Exploring AI: Opportunities and Risks for Developers](https://www.wearedevelopers.com/magazine/522-exploring-ai-opportunities-and-risks-for-developers)
- [What is Software Engineering in the Age of AI?](https://www.wearedevelopers.com/magazine/640-what-is-software-engineering-in-the-age-of-ai)
- [How we Build The Software of Tomorrow](https://www.wearedevelopers.com/magazine/120-how-we-build-the-software-of-tomorrow)
- [Dev Digest 137 - AI'm not sure about this](https://www.wearedevelopers.com/magazine/485-dev-digest-137-ai-m-not-sure-about-this)

## Related Jobs

- [Staff Software Engineer, Copilot Experiences](https://www.wearedevelopers.com/jobs/ext/164361-staff-software-engineer-copilot-experiences) at **GitHub**
- [Principal Software Engineer, Enterprise AI Platform](https://www.wearedevelopers.com/jobs/ext/1467292-principal-software-engineer-enterprise-ai-platform) at **GitHub**
- [Principal Product Manager, Agent Platform](https://www.wearedevelopers.com/jobs/ext/277541-principal-product-manager-agent-platform) at **GitHub**
- [Senior Software Engineer, Enterprise Products](https://www.wearedevelopers.com/jobs/ext/1841248-senior-software-engineer-enterprise-products) at **GitHub**
- [Staff Developer Advocate, GitHub Security Lab](https://www.wearedevelopers.com/jobs/ext/1921051-staff-developer-advocate-github-security-lab) at **GitHub**
- [AI Software Engineer (Germany)](https://www.wearedevelopers.com/jobs/48317-ai-software-engineer-germany) at **Sunhat**