Is AI Going to Replace Programmers?

We’ve seen the rise of OpenAI’s ChatGPT, Github Co-pilot, Google Gemini, X Grok, and more. They’re incredible in generating code so it’s not out of the question that programmers might not be needed as much as before. But does that mean that all programmers will be replaced by AI in the future? It’s highly likely that a ton of them will be replaced, but it is unlikely that all programmers will be replaced, here’s why:

The Value Generative AI Brings to Coding

Generative AI is only as good at coding as the source data it has trained from and let’s face it, there’s a lot of great data available for it to train on. There’s also a lot of bad data for it to train on which makes the output of the AI freakishly good, since it’s been trained on both the good and the bad.

One thing that generative AI can’t do though, is think creatively about a solution. It’s really great at giving you snippets or creating something as a whole, but it’s not great at determining the best piece of code to generate based on the entire codebase you’re working with. If you’re a seasoned programmer, then you know how frustrating it can be to come into a codebase that’s basically a bunch of random parts glued together to work. If you’re not, imagine an old used car with several off-color parts installed to keep it looking like a working car - sure it works but holy smokes is it ugly.

The Value Programmers Bring to Coding

Programmers are creative thinkers that bring real-world experience to their craft. Just like how you can’t expect a chat bot to always solve your customer service issue, you can’t expect generative AI to always be able to give you the right code to use in your application. Given enough scale, it’ll work for the majority of cases (laws of probability), but it won’t work for edge cases. Let’s be real, if you’re trying to use AI to create your own software product - the chances of it knowing how to output the right code for your unique case is going to be rare. You’re going to need someone who understands programming, code architecture, and has experience designing software applications to get something new built. Generative AI can help, but it won’t replace your need for a programmer.

Seasoned programmers know what’s possible and what’s not possible. They understand what good code looks like and what bad code looks like. They’ll know what to ask the AI to help with so that the end product works as expected. Without a programmer, you’re basically throwing darts at a board - you’ll have the general direction but will lack accuracy. Think of it like this - You can ask AI to generate a Shakespearean piece, but it’s not going to be a brand new work of art that has never been seen before and would pass as Shakespeare’s latest masterpiece.

What Will the Future Look Like?

There is no doubt that generative AI for coding has and will continue to change the landscape of code development. It won’t fully replace the need for programmers, but it will likely put a lot of programmers out of the job. Bluntly put, only the truly creative thinkers will remain while the ones that simply code by instruction will get left behind.

Every smart, capable programmer will become exponentially more productive with AI helping them write code faster and fix syntax issues with ease. Successful programmers will use that time saved to be more critical thinkers and planners in how they design and architect their solutions. The creme of the crop will become leaders that shape the direction of their products, becoming more focused on solving their users’ issues rather than writing code to fulfill feature requirements.

If you’re a programmer in today’s world that simply works off of ticket requests without any thought about the “Why” of what you’re building, then your time is finite given the rise of generative AI. I strongly recommend that you begin thinking about how to level up your thinking from a simple individual contributor to a dev leader or product leader. If you’re new to this way of thinking, there are two books that I’ve treated as gospel in my 12+ years of product development: Inspired by Marty Cagan and Hooked by Nir Eyal (these are Amazon affiliate links, but the price is the same for you either way, it just helps support me if you use my links).