April 15, 2025

AI Reveals What Actually Requires Thinking: Reflections on the Future of Development

Tim O'Reilly's recent article on AI's impact on programming offers a refreshingly nuanced perspective on how AI is transforming our craft. Rather than the apocalyptic "AI will replace all developers" narrative, O'Reilly sees something far more interesting: a fundamental shift in how we work.

Beyond the Doom and Gloom

Just read Tim O'Reilly's brilliant article on AI's impact on programming, and what resonated most wasn't another prediction about job displacement, but rather his clear-eyed view of transformation.

O'Reilly doesn't see this as the death of development careers but rather a fundamental shift in how we work. This perspective feels both more realistic and more actionable than the binary "AI will replace us all" versus "AI changes nothing" debates we often see.

The reality, as O'Reilly frames it, is far more nuanced and interesting.

Engineering Wisdom in the Age of AI

My favorite insight from the article: "When you watch a senior engineer work with AI tools... They're applying years of hard-won engineering wisdom to shape and constrain the AI's output."

This observation perfectly captures what I've seen in practice. The developers who are most effective with AI tools aren't the ones who blindly accept whatever the AI generates. They're the ones who know how to frame problems, recognize good solutions, spot potential issues, and guide the AI toward better outcomes.

Years of debugging obscure edge cases, understanding system architecture, and grappling with the consequences of technical decisions—this experience doesn't become obsolete. Instead, it becomes the lens through which we evaluate and refine AI-generated solutions.

What Actually Requires Thinking

But perhaps the most profound insight from O'Reilly's piece is this gem: "I don't think AI introduces a new kind of thinking. It reveals what actually requires thinking."

This reframing is powerful. Instead of asking "What can AI do?" we should be asking "What requires uniquely human cognitive abilities?"

AI excels at pattern recognition, code generation from clear specifications, and handling well-defined problems. But it struggles with:

  • Understanding ambiguous or incomplete requirements
  • Making strategic technical decisions with long-term implications
  • Balancing competing priorities and constraints
  • Debugging complex system interactions
  • Communicating technical concepts to non-technical stakeholders
  • Designing solutions that account for organizational context

These aren't new skills—they're the skills that have always separated junior developers from senior ones, good engineers from great ones. AI just makes this distinction more visible.

We're Still at the Beginning

As O'Reilly notes, we're still in the early days—"we're not even at the locomotive stage of AI yet." This historical analogy is particularly apt.

Early automobiles were literally called "horseless carriages" because people could only understand them in terms of what came before. Similarly, many of our current AI tools are essentially "programmer-assisted coding" rather than fundamentally new ways of building software.

We're still figuring out the equivalent of paved roads, traffic lights, and highway systems for the AI era. The infrastructure, patterns, and best practices are still being invented.

The Transformation is Already Underway

But the transformation of our craft is already underway, and it's happening faster than many anticipated.

I've watched development teams dramatically accelerate their prototyping cycles using AI-assisted coding. I've seen complex data transformations written in minutes rather than hours. I've witnessed AI help developers work in languages they're not fluent in, expanding their ability to contribute across different parts of a system.

At the same time, I've seen the limitations become clear. AI-generated code often lacks the subtle considerations that come from experience. It can create technical debt in ways that aren't immediately obvious. It sometimes solves the wrong problem entirely.

Adapting Our Practice

So how do we adapt our practice as developers in this new landscape?

Develop better problem definition skills: If AI is going to generate solutions, we need to get better at defining problems clearly and completely.

Focus on system thinking: Understanding how components interact, where bottlenecks emerge, and how changes propagate through complex systems becomes even more valuable.

Cultivate judgment: The ability to quickly evaluate whether a solution is good, maintainable, and appropriate for the context is increasingly important.

Embrace the collaborative model: Rather than seeing AI as a replacement, learn to work with it as a highly capable but sometimes misguided colleague.

The Craft Evolves

Every significant technological shift in software development has changed what we spend our time on without eliminating the need for skilled practitioners.

High-level languages didn't eliminate programming—they made it more accessible and shifted focus from memory management to algorithm design. Visual development tools didn't replace developers—they enabled rapid prototyping and shifted emphasis to business logic.

AI represents another such shift. It's automating the mechanical aspects of coding while highlighting the importance of strategic thinking, problem-solving, and human judgment.

As O'Reilly suggests, this isn't the end of development careers—it's the evolution of what it means to be a developer. And for those who embrace this evolution, the future looks remarkably bright.

The question isn't whether AI will change our field. It's whether we'll thoughtfully shape that change or simply react to it.

Aaron Wallis is a founder and technology leader who positions himself as a "Technical-Creative Bridge." With over two decades of experience building technology that bridges creative vision with technical excellence.

Share this article

More Articles

AI Reveals What Actually Requires Thinking: Reflections on the Future of Development

Exploring Tim O'Reilly's insights on how AI is transforming programming - not killing it, but revealing what truly requires human thinking.

Read More

When AI Becomes Your Best Referral Partner: The New Era of Organic Discovery

How AI recommendations are reshaping business discovery and what this means for the future of SEO and organic search.

Read More

From R&D to Production: Our GenAI Journey at AWS Summit

Reflecting on Lexer's journey from early GenAI research in 2023 to production deployment and optimization, as presented at AWS Summit.

Read More

When Conversation Becomes the Interface: The GenAI UX Revolution

How GenAI chat agents are fundamentally changing our UI/UX expectations and why traditional forms are starting to feel obsolete.

Read More

GEM: The New SEO for the GenAI Era

How Generative Engine Marketing is reshaping digital discovery, with insights from Vercel's success getting 10% of signups from ChatGPT.

Read More

Vibe Coding: The Self-Driving Car for Development

How LLMs are transforming development by handling the tedious work, freeing developers to focus on architecture and decisions that actually matter.

Read More

Authentic Selves: Reflections on Pride, Community, and the Retail Industry

A personal reflection on diversity, inclusion, and the privilege of showing up authentically in professional spaces during Pride Month.

Read More

Being a Verb: The Power of Action in a World of Overthinking

Explore the concept of 'being a verb' and why prioritizing action over overthinking leads to better outcomes.

Read More

AI Strategy: Balancing Innovation with Implementation

Practical considerations for implementing AI in your organization without getting lost in the hype.

Read More

From COVID to AI: The Next Phase of Social Distancing

How AI might create a more profound shift in human interaction than COVID social distancing, further abstracting our connections.

Read More

On Avalanches and Snowballs

A reflection on balancing short-term and long-term projects in product teams, using debt payment strategies as an analogy.

Read More

A Recipe for Resilience in Pride Month and Beyond

A personal reflection on maintaining pride and resilience during difficult times, with a meaningful connection to Pride Month celebrations.

Read More

War and Peace

An exploration of how businesses operate either 'at war' or 'at peace' and how this mindset affects their strategies and operations.

Read More