Techniq World
Software

The Rollercoaster Reality of AI Agents in Engineering Workflows

Learning a new coding workflow only to watch it become obsolete by the next model release is the constant tax of modern software development. Coding agents have made it...

In brief

Learning a new coding workflow only to watch it become obsolete by the next model release is the constant tax of modern software development. Coding agents have made it...

Angie Jones of the Agentic AI Foundation explains how coding agents are forcing engineering teams to rethink junior onboarding and workflow stability.

Learning a new coding workflow only to watch it become obsolete by the next model release is the constant tax of modern software development. Coding agents have made it simpler for people to propose a fix or a feature, shifting the landscape of engineering organizations.

The Block Experiment and Shifting Roles

A few years ago, putting new hires in front of unfamiliar codebases with access to AI agents would have sounded impossible. A lead developer at Block once handed 10 new hires a codebase they had never seen before, provided them with access to AI agents, and gave them a day and a half to solve real customer problems. Even though it was their first day on the job, every single participant delivered solutions.

This experiment raises pressing questions for engineering teams regarding where junior developers fit into modern pipelines. It also forces experienced engineers to consider what they might learn from those newer to the field. Angie Jones, the lead developer from that experiment who now serves as vice president of the Agentic AI Foundation, described the current environment as a rollercoaster ride during a conversation in San Jose last month at AGNTCon + MCPCon North America.

Engineering Adoption and Open Source Maintenance

Engineering organizations face ongoing friction as they attempt to integrate AI tooling into daily operations. Jones has navigated this firsthand, having led AI enablement at Block while also helping maintain the open source coding agent Goo.

The rapid pace of model updates creates a persistent tension across teams. Developers must constantly adapt their habits as underlying capabilities shift, making long-term workflow stability difficult to maintain. While coding agents streamline the process of submitting patches and new features, managing the human side of this technological transition remains a complex hurdle for leadership.

What to Watch in Agentic Workflows

Tech consumers, gadget buyers, and developers should monitor how open source foundations and enterprise teams adapt their onboarding processes as agent capabilities evolve. Observers should track how engineering organizations balance the speed of AI-assisted code generation with the unpredictable lifecycle of frequent model updates.

Frequently Asked Questions

Q1: Who is Angie Jones?

Angie Jones is the Vice President of the Agentic AI Foundation, a former lead developer at Block who led AI enablement, and a maintainer of the open source coding agent Goo.

Q2: What was discovered during the Block onboarding experiment?

Ten new hires were given an unfamiliar codebase and AI agents for a day and a half to solve customer problems, and every single one of them delivered results on their first day.

Q3: Where did the discussion between Angie Jones and Frederic Lardinois take place?

The conversation took place last month at AGNTCon + MCPCon North America in San Jose.

Leave a Reply