If AI can code, what will Software Engineers do?

I've worked as a software engineer professionally for 9 years. AI has taken the world by storm this year and it can code. How well depends on the usecase, technologies (and available training data), and your own expertise - but it can certainly pump out small features in mainstream programming languages.

Of course it's gotten me thinking about what the future of this profession holds. After all, if AI can code why do companies need software engineers? If those companies aren't employing software engineers, do I need to find another job?

That's what we'll be exploring in this post.

New Technologies and Job Change

We've seen this pattern numerous times throughout history. A new technology comes along that fixes a common problem and wipes out whole categories of human roles because now it can be done faster, cheaper, and with more consistency.

Quality is another matter entirely but I think most will agree with me that humans are not the most consistent entities on the planet. If anything, we resemble non-deterministic pattern matchers with a bit of emotion splashed in for personality. Hm, what does that remind me of?

You've probably noticed this when you see old movies / news reels and there's background characters doing smth odd. What was normal then is now extinct because modern life has changed so much.

Some examples include:

• Personal Computer - Everyone can type so no need for typists
• Calculator - Everyone can do perfect calculations, no need for human calculators
• Global supply chain - Everyone can get milk at their grocery store, no need for a milk man
• Email - Everyone can email anyone else, no need for in-office couriers (my parents recently told me ab office couriers and how mailrooms were necessary to send messages to other departments)

And that's just a few of them.

The point is that new technologies regularly deprecate human roles because they are faster, cheaper, and more consistent. In most cases, we look back and barely remember those roles a generation on. This change happens but it's been happening for decades.

How AI is different

I think AI is mostly the same. It raises the floor on common tasks which removes the need for whole categories of roles cause now anyone can do them w new tools.

It feels different in a few ways:

• It's software
  • Can be updated and deployed quickly around the globe
  • Has been repurposed to impact many industries
  • Has potential to be plugged into basically anything that's software - and the world has been eaten by software
• The amount of money going in is unfathomable - multiple $100B deals whipped around

But it's not like we've never seen stuff like this before (and recently!):

• Rise of the internet
• Dotcom bubble
• Rise of smartphones

It's scary because it's exponential but we've been on an exponential growth curve for centuries so it's not that weird we'd unlock additional exponential growth to keep the curve going.

I got this idea from Andrej Karpathy's (OpenAI cofounder) Dwarkesh podcast. He believes AGI will be exponential but it will blend into the 2.5 centuries of 2% GDP growth we've seen.

AI, Engineers, and Coders

Software engineers will go a similar route. AI is a new technology that raises the floor on coding. Anyone can go to a mainstream AI, tell it to build them an app, and AI will spit out a thousand line working webapp in seconds.

This is a huge change! Previously you'd need to hire someone to do that and humans are expensive. Not to mention most of the world still doesn't really know how code works so the code was a black box bottleneck and therefore could charge extra for its creatoin.

So of course it's going to impact this profession. But I think it's going to impact different roles differently.

• Software Engineer - Takes problems / requirements and executes the technical side e2e
• Coder - Handed a task and codes it

The former solves a problem e2e with technology. The latter is given instructions and follows them.

AI can already code so I think there's little present or future for coders. That's like having a personal typist in 2025 - you can already type (or dictate to your device) - so having a human in the loop is generally wasteful.

What's going to set engineers apart is what they can do that the AI / new technology cannot yet do:

• Deciding on priorities, direction, and tradeoffs
• Planning large projects
• Splitting, sequencing, and delegating tasks
• Interfacing with the human world - humans, tools, intuition
• Upholding quality

This is why I'm so big on vibe engineering over vibe coding - if you can vibe code it then there's no reason for you to be in the loop. Vibe engineering is leveraging the tool for what it's good for while adding value yourself.

AI and other roles

As we've mentioned, AI seems different from previous technology advancements because its impacts are so widespread - across industries, B2C, and B2B.

The generalization for how this affects human roles across industries I think is:

• AI raises the floor so jobs that were only doing the floor work are unnecessary - AI can do it faster, cheaper, and likely better than them
• Those who use AI will be able to leverage it to achieve greater outputs than they could individually
• The market will rearrange itself to this new equilibrium, shutting out jobs that are unnecessary and growing in areas where bottlenecks exist.
• New jobs will unfold around those bottlenecks - we're not sure what they are but they will appear (think computer repair, mobile app developers, and influencers - all things that didn't exist widely until the rise of computers, smart phones, and the internet respectively)
• As AI continues to improve, the bottleneck will continue to be identifying opportunities, executing on them, and redirecting based on the outcomes
• AI will continue to raise the floor so the staying factor is to raise the ceiling - in depth (specialize super deep in a few areas such that you beat the AIs) or breadth (become a good generalist so you can leverage these tools across layers to achieve outcomes they can't)

Next

AI is going to change everything. But everything is already changing. It's been changing for hundreds and thousands of years. This time will be different but it will also be the same.

The strategy remains: Learn the new tools, adapt to the new environment, and carry on.

If you liked this post you might also like:

• Stop Vibe Coding, Start Power Coding - How To Write Quality Software Faster With Agentic AI
• My Career in Programming Languages
• I Vibe-Coded a C# Library with Claude Code - Here's 6 Things I Learned