I’m with Steve Balmer. Software developers are not going extinct. In fact the amount of innovation that will be built on AI technology will make them more indespensible than ever.
The dumb idea that AI is going to replace all developers is being promoted by hypesters and people who are completely ignorant about how software is actually made. It's true that the role and workflow of coders will change. In fact, AI might end up writing all the code—but it will never do all of the thinking.
My brilliant friend Tim Christensen says that we don’t pay developers to write code, we pay them to think. Those mythical 10x developers you hear about spend most of their time thinking and playing the dinosaur game and a minority of their effort goes into actually writing code.
Once again, we've all be fooled by snazzy demos that make us imagine a future that will likely never arrive. Vibe-coding tools like Bolt, Lovable, and Replit give the impression that soon, anybody will be able to chat their way into software products. And that’s already almost a possibility! At Machine & Partners, we use those tools all the time to quickly create prototypes or scaffold apps. But a “real” software application is not a one-and-done artifact—it’s a living, breathing, ever-changing solution to an ever-evolving business problem.
What non-coders don’t understand is that writing the code is the easy part. It’s like being an architect. You don’t pay them to draw plans—you pay them to design a building that solves for space and cost constraints. The plans are just the artifact that comes out of the process of doing the valuable work of thinking, designing, and problem-solving in general.
The Reality: What Developers Actually Do
Development requires creative problem-solving that AI can’t replicate:
- Translating ambiguity into clarity: Turning vague business needs ("make it user-friendly") into precise technical specifications
- Systems thinking: Understanding how changes to one component might affect ten others
- Contextual judgment: Deciding when to optimize for speed, security, maintainability, or cost
- Innovation: Creating solutions for problems that have never been solved before
- Collaboration: Coordinating and merging other’s strengths and weaknesses into a reliable process
AI can write code. And that is awesome, because writing code, especially boilerplate code that every coder has written a thousand times, is tedious and error-prone.
AI can turn thoughts into code. Which is awesome because typing brackets and semicolons in accordance with a strict syntax can be tedious and error-prone.
AI can organize code. Which is awesome because it allows developers to ponder different ideas, weighing their pros and cons before committing.
So, most software engineers welcome their AI coding partners. And the vast majority of them aren’t worried about losing their job because they know what their job actually is.
The 2024 Stack Overflow Developer Survey tells the story:
Almost 72% of developers surveyed have a favorable impression of AI as a code-writing assistant.
And only 12% feel it’s a threat to their job.
That’s despite the near constant headlines that developers are going away and that kids shouldn’t study computer science in college.
A more nuanced and helpful view is that we simply won’t perform computer science the same way when AI’s get good enough to codify our designs without much human intervention.
It’s the leaders and managers we need to worry about. Short-sighted cost cutting opportunities are setting expectations in the wrong places.
AI is nowhere near replacing developers, but that doesn’t mean developers shouldn’t be utilizing AI to its fullest. The developers embracing these new tools are becoming super-coders. The ones who aren’t are probably near retirement or working in niche corners of a domain where AI doesn’t bring a lot of value.
I obviously don’t know the future. AI is moving fast. Is there a chance that some crazy breakthroughs in the next few years will make developers obsolete? Sure. But if that happens, then I’d argue CEOs, CFOs, and CMOs, will be obsolete too. Futurism is fun, but rarely accurate.
Every knowledge worker, developer, marketer, scientist, accountant, etc., is getting an incredible tool that we haven’t figured out how to fully exploit yet. It’s amazing! But my team builds fairly advanced AI systems and I can tell you without a doubt that they are nowhere near replacing the most valuable things we humans do. In general there is waaaaaay too much context that we would need to load into an AI system to get it to replace a human with even basic standards of acceptability.
The demos we all see emerging on a weekly basis are really cool. But when you take a closer look you realize that while their output is unprecedented, it’s also inadequate. AI can produce a sales deck. But is it a good sales deck? It can analyze some data, but is it an insightful analysis? It can write an article, but is it a good article? The answer is: sometimes it’s good enough. And that’s not good enough to replace us.
The Economic Math: Productivity Gains ≠ Headcount Reduction
Let's analyze this from a CFO and CEO perspective. Imagine you lead a technology company competing neck-and-neck with rivals. You have a strategic backlog of features that could differentiate your product and capture market share.
If AI suddenly made your development team 50% more productive, would the optimal business decision be to fire half your developers? That might improve your P&L temporarily and please shareholders for a quarter. But you’re competing against companies like Hooli, and you can’t let your innovation stagnate or your competitive position erode.
The financially sound long-term strategy would be to maintain your team and accelerate feature development, expanding market share and increasing revenue—metrics that ultimately matter more to enterprise value than short-term cost reductions.
It’s funny, because two years ago if you asked a typical software leader what their number wish would be it would be higher velocity—the ability to fix bugs and ship features faster. That wish has come true, but all anyone can talk about is reducing costs by firing engineers!
Unfortunately you have to consider this competitive reality: The AI productivity tide lifts all boats. Your competitors gain the same efficiency advantages. The winning strategy isn't reducing your developer headcount—it's leveraging your newly supercharged development team to out-innovate competitors.
Leadership Strategy: Building the AI-Enhanced Technology Organization
For technology executives planning their future workforce, the developers who will drive the most business value will be those who excel at:
- Thinking Strategically: Articulating which technology problems, when solved, create the most business value, and which business problems have viable technology solutions
- Thinking Practically: Making design decisions that balance immediate needs with the company's long-term strategy
- Thinking Holistically: Understanding how interactions across components impact reliability, scalability, and security
- Thinking Collaboratively: Working productively with other stakeholders and business experts to define and guide successful projects
In fact every knowledge professional in your organization— developers, marketers, data scientists, and finance teams—is gaining access to unprecedented productivity tools. However, as executives who build and deploy AI systems can attest, these tools are nowhere near replacing the strategic thinking that drives the largest enterprise value.
The Executive Takeaway
The competitive advantage will go to companies that strategically pair AI with smart developers—not those who mistakenly view AI as a replacement for technical talent. The creative problem-solving that drives innovation and business value will always require human judgment. The boon of AI isn’t that it replaces people. It’s that it makes your existing teams more impactful than ever before.
Forward-thinking technology leaders are going to need to rethink the development process. This is truly a paradigm shift in how products are defined and produced. Agile will be more agile. Teams will be more cross-functional. The cost of experimentation will plummet. Roles will become less distinct because AI will allow broader participation from non-technical folks. So many things are changing and the playbook for how to do it hasn’t been written.
Here’s my advice:
- Invest in AI tools that amplify your development team's capabilities
- Reallocate developer time from routine coding to thinking and experimentation
- Incorporate subject matter experts directly into the development team
- Measure success by business outcomes not output velocity
- Develop a talent strategy that emphasizes the skills AI cannot replicate
Let’s also be realistic. Not everyone will keep their job—because the jobs themselves are going to change as our AI-enabled processes emerge. At Machine & Partners, our product managers are super-product managers. They are two to three- times as effective as PMs I’ve worked with in the past. Our developers are super-developers. They blow our client’s minds with how quickly they are able to build solutions to real problems. And I aspire to be a super-leader, using AI to help with business automations and strategic thinking.
The age of the AI-augmented organization has begun. The question for executives isn't how much you'll save by replacing your employees with AI—it's how fast you can harness AI to make your employees more effective than your competitors'. Machine & Partners helps our clients define and execute transformational AI strategies. If you're interested in thinking through your strategy together, give us a shout!
