I think AI will create more "software engineers," not fewer. Only they will have knowledge akin to a superintendent in the construction industry, or a general contractor. It will be a high-skilled job, but it won't require as much depth in the specifics of any given programming language, networking, or server configurations. Instead, it will be focused on "higher-level" configuration and tooling, and more focused on the data layer and its relationship to the business need.
This is similar to how the early 2000s generation of programmers are less focused on hardware constraints, programming writing-style, and how things work from the operating system level down to memory and processing. For instance, it is entirely possible that a JavaScript programmer can build an "app" and understand very little about system memory or how the OS interacts with hardware. The builders of operating systems and the apps that newer programmers use truly are giants on which JavaScript programmers stand.
The concept of the "program used to program" is the basic idea here. As Newton said: “If I have seen further, it is by standing on the shoulders of giants.” So it is that AI-enabled programmers will build their apps.
I think we are very far from AGI and AI building apps on its own — if that will ever happen — hopefully not. I think we are still pretty far from the first AI-empowered developers being able to build successful products or contribute meaningfully to successful products. Thinking that because AI tools can write code in JavaScript or any other language that they can therefore build apps misunderstands what it takes to build a viable product.
It may be able to “build an app” in the sense that it can create a simple video game, todo list, or note-taker. I think AI can copy existing apps fairly well and maybe run and deploy them on its own. But when it comes to modifying those apps in a meaningful way or translating business needs into a working system, it requires a human layer of judgment and experience.
An AI-enabled “programmer” who doesn’t have deep understanding of programming will need to memorize techniques, processes, and develop a mental model of how parts fit together. They will need domain-specific knowledge that requires the same “10,000 hours” that engineers and other professionals need to master their craft.
Therefore, the new programmer will likely still be a programmer or even a "software engineer" — not because they code everything manually, but because they understand what to build and how to direct the tools that build it. Manual DevOps or low-level coding may become unnecessary, and even distracting, from actual product-building.
Just as JavaScript developers don’t need to understand the compiler, runtime, OS, and hardware layers in depth, so too might AI developers focus on shaping data flows, user experiences, and system configurations rather than building from scratch. Each layer is very "deep" — to master any layer can take years.
Will the AI-enabled developer need to know how to write code, or just what kind of code is needed? I think they will focus on building Software-as-a-Service (SaaS) and applying LLMs in SaaS contexts. They will articulate user needs, UX minutiae, permissions, and personify data — enabling rich metaphors about relationships between data structures.
We’ll likely see hybrids of games and apps emerge. AI will enable expressive, imaginative systems, but building those systems will still require a deep technical imagination.
This should not be mistaken for “easy.” Even with AI as a helper, building serious software still demands deep knowledge. I’m hard-pressed to think of any engineering effort that has become fundamentally easier due to tech. It just changes the expectation of output.
Sure, building an app with HTML, CSS, and JS may seem simpler than writing a compiler or innovating in cryptography. But today's app developer may exert just as much effort as those who built the layers they stand on. This has always been true — from C devs to early hardware hackers.
With powerful tools, others expect more. Businesses focus on results. If a tool exists, it’s assumed you’ll use it to produce something great. That’s why the pace of progress increases. AI developers may run 1,000 mph while JS devs ran at 10 — but the new destination is a million miles away.
Expectations scale with power. The paradox is: as tools make more possible, the average developer must still give the same or more effort to stand out or contribute meaningfully.
The rise of AI doesn’t reduce the need for skilled engineers — if anything, it increases it. Layoffs at mega-corporations may stem more from supply-demand mismatches than from AI efficiency gains. As more people enter tech, competition intensifies.
These new entrants may be more talented, more educated. But that doesn’t mean the job is easier. AI hasn’t lowered the barrier — it has simply brought more attention to the field.
Ultimately, AI won’t replace engineers. It will redefine the skillset and raise the bar for what’s possible — and what’s expected.
Sincerely,
Buckley Mower
February 13th 2025