Will Engineers Still Matter?

AI is redefining what engineers do, and who you need to build a great product team.

Over the past two weeks, I’ve had the same conversation—again and again.

It came up with friends over coffee. With clients during roadmap reviews. With colleagues deep into hiring discussions.

The question always landed in the same place:

What will our engineering team look like next year? What about five years from now?

It’s a real, pressing concern. And nearly everyone is asking because of one thing—AI.

The AI Impact

Software engineering may be the domain most visibly impacted by AI right now.

And the changes are coming fast.

On the optimistic end:
Check out this conversation with Dan Shipper: The AI-native startup: 5 products, 7-figure revenue, 100% AI-written code.
Dan’s team is shipping faster than ever—with AI doing most of the coding heavy lifting.

But there’s also serious disruption:

• Zuckerberg Says AI Will Replace Mid-Level Engineers Soon (Forbes)

• Goldman Sachs Hires an AI Engineer Instead of a $180K Human (Fortune)

• And about 50 more examples if you google it...

These aren’t outliers—they’re part of a clear pattern.

The Problem

If AI can now write high-quality code faster, what does that mean for engineers?

In short: writing code isn’t enough anymore.

AI dev tools are making one engineer as productive as two or three. (IBM, New Horizons)

So the value of a great engineer is shifting up the stack.

Insight From the Field

Here’s what I’m seeing firsthand:

• Engineers are taking on product and project work. They're managing scope, architecture, and progress—while AI handles much of the coding.

• QA is increasingly automated. Traditional QA teams are shrinking. AI tools now handle test generation and validation. Engineers simply review the results.

• PMs are getting squeezed. Two of my clients have no plans to hire more product managers. Instead, they’re focused on enabling AI-augmented engineers to build more, faster.

The result? Faster development, fewer bottlenecks, and leaner teams shipping more product.

How to Adapt

For Engineers:

• Master “agentic” AI tools
  At Robinhood, "close to 100%" of engineers use AI editors like Cursor and Windsurf, with roughly 50% of new code AI-generated.
  Business Insider

• Become a code review expert
  Perplexity’s team made AI tools mandatory—prototype development dropped from days to hours—but bugs still sneak in.
  Business Insider , Entrepreneur

• Think beyond syntax
  Replit’s Agent lets devs describe apps in English now—they steer architecture, UX, and logic, while AI scaffolds code.
  Source: Wikipedia (Replit):

For Startups:

• Use “vibe coding” workflows
  Companies like Vanguard and Choice Hotels use natural‑language AI (OpenAI, Claude) to prototype UI pages ~40% faster.
  Wall Street Journal

• Reduce PM headcount
  Teams mandating AI-based architecture and scope planning are cutting dependence on PMs—engineers now drive end-to-end delivery.
  Business Insider

For Enterprises:

• Pilot AI in modernization
  Pegasystems + AWS offer agentic automation tools to speed legacy-to-cloud migrations with minimal code rewrites.
  ITPro:

• Appoint internal AI champions
  Over 100 CIOs are forming AI task forces to set adoption standards and manage governance.
  a16z

• Automate test generation
  Qodo (formerly Codium) raised $40 M to build enterprise-grade AI test and review tools.
  Wikipedia (Qodo):

Pitfalls to Avoid

• Assuming AI is faster
  A METR study shows experienced developers were ~19% slower with AI tools, spending time fixing and interpreting code.
  Time , Axios:

• Cutting teams too soon
  Capgemini warns many AI initiatives fail to deliver ROI—losing oversight too early can backfire.
  Times of India

• Skipping governance
  Gartner notes ~40% of AI code in enterprises comes from non-devs—ground rules are critical.
  Wall Street Journal:

• Neglecting training
  McKinsey highlights that without proper training in prompt engineering and governance, AI tools slow people down.

• Skipping user validation
  Faster shipping doesn’t guarantee value—without rigorous user feedback, speed can be misleading.

Closing Thoughts

The question isn’t whether AI will reshape engineering teams—it already has.

From startups skipping product managers to enterprises automating legacy migrations, the shift is happening in real time. Engineers are no longer just coders. They’re architects, product thinkers, and AI supervisors.

But here’s the thing: this isn’t about replacement. It’s about reinvention.

Yes, AI is automating the grunt work. But the teams thriving in 2025 are the ones doubling down on judgment, creativity, and clarity. They’re using AI not to do less—but to do more of what matters.

So whether you're an engineer, founder, or exec, the challenge is the same:
Adapt fast. Stay curious. Build better.

That’s the real edge in the AI era.