When Garry Tan revealed that some Y Combinator startups have 95% of their code written by AI, many read this as a death knell for software development careers. The reality is more nuanced and, frankly, more interesting than simple replacement narratives suggest.

What’s happening in software development isn’t the elimination of the profession. It’s a dramatic reshaping of what it means to be a developer in the first place.

The traditional career path for developers typically moved from junior roles focused on implementing straightforward features to mid-level positions handling more complex implementation work and finally to senior roles focused on architecture and systems thinking. This ladder is now compressing rapidly.

Two distinct profiles

I spoke with several development team leads last month about how AI coding tools have changed their hiring practices. The consensus was striking: they’re increasingly looking for two distinct profiles rather than the traditional spectrum of junior to senior developers.

The first profile is the systems architect: developers who excel at understanding complex business requirements, designing scalable solutions and making critical technical decisions. These roles require deep technical knowledge combined with business acumen and strategic thinking.

The second profile is what several described as the “AI orchestrator”: developers who may not have decades of coding experience but who excel at effectively prompting AI systems to generate useful code, evaluating the output and integrating it into larger systems.

The missing middle tier

What’s notably missing? The traditional mid-level developer role focused primarily on implementing features according to specifications. This work – writing authentication systems, building standard UI components, setting up databases – is precisely what AI tools now excel at generating.

This mirrors the transformation we’ve tracked in executive assistance (EA), where the middle tier of competent but not exceptional EAs is facing similar pressure as AI handles routine tasks. Just as with developers, the EA profession is splitting into highly strategic relationship managers on one end and AI tool managers on the other.

AI coding tools as an amplifier

For developers navigating this shift, the path forward isn’t about competing with AI on coding speed or syntax knowledge. It’s about developing the skills that complement rather than compete with what AI can do: systems thinking, business domain knowledge and effective AI orchestration.

The most successful developers I’ve spoken with describe their relationship with AI coding tools not as a threat but as an amplifier. They’re writing less code directly but accomplishing more, focusing their human creativity and problem-solving on the aspects of development that truly require it.

This shift demands a fundamental reconsideration of how we train developers and build technical teams. The days of large development teams with many mid-level implementers may be ending. Instead, we’re moving toward smaller, more senior teams augmented by AI tools that handle much of the implementation work.

Honest self-assessment

For individual developers, particularly those in the vulnerable middle, this transformation requires honest self-assessment.

Where do you truly add value that AI cannot? Is it in your understanding of complex systems? Your ability to translate business needs into technical solutions? Or perhaps in your skill at effectively deploying AI tools themselves?

The developer’s role isn’t disappearing. It’s evolving into something potentially more impactful but undeniably different from what many built their careers around. The question isn’t whether to adapt, but how quickly you can reposition yourself for this new reality.