AI Coding Shifts Development Bottleneck from Execution to Strategy

Anthropic's AI coding tool Claude Code is reportedly increasing engineer productivity by a factor of 3. In response to this change, Anthropic instructed its growth team to expand hiring of product managers (the role responsible for deciding product direction and priorities). With code writing speed no longer a bottleneck, the shortage of talent capable of deciding 'what to build' has become apparent.

To understand this shift, it helps to look back at how engineers' workflows have changed in recent years. Before 2022, the standard solution when developers got stuck was to search for answers on the Q&A site Stack Overflow. However, since ChatGPT's launch in November 2022, new posts to Stack Overflow have declined by approximately 77%. This demonstrates that engineers are increasingly choosing generative AI as their problem-solving tool.

Subsequently, how AI tools are used has evolved in stages. In the initial stage, engineers repeatedly asked AI questions in a browser and manually pasted the responses into their code editor. In the next stage, tools like Cursor and Claude Code embedded AI directly into editors and enabled direct access to repositories (the complete codebase storage). Rather than Bash, the command-line tool that veteran engineers long described as the 'most continuously used tool,' many developers by 2026 are launching the 'claude' command first when starting new work.

Furthermore, from 2025 to 2026, as the volume of context that AI can handle (context window) expanded, a 'specification-driven' development style became widespread where work that previously required multiple design documents or weeks of development cycles could be completed in short timeframes from a single specification. Amazon's code editor Kiro development team reported shortening feature development time from two weeks to two days using this approach. Additionally, an AWS engineering team completed a system redesign originally planned to take 30 people over 18 months in just 76 days with 6 people.

In April 2026, Anthropic released Claude Code Routines. This is an autonomous AI agent feature that operates continuously on a defined schedule, further automating development work. This capability enables AI to continuously handle repetitive tasks without people constantly providing instructions.

What these changes demonstrate is that the 'rate-limiting step' in software development (the point that determines overall speed) has fundamentally shifted. Whereas code writing speed once determined development speed, whether one can clearly define 'what and how to build' now determines project-wide speed and quality. This reflects a growing requirement that engineers possess not only technical skills but also the thinking capability to consider product direction.

Anthropic's decision to increase product manager hiring in its recruitment strategy also signals insight for the broader industry. The more AI becomes capable of substituting for code-writing work through increased speed, the more human roles concentrate on 'giving direction' and 'determining strategy.' This structural change is positioned to continue affecting engineer career development and how companies structure development teams in the future.