Developers Are Missing the Point About AI

Everyone is using AI now, so much so that the Pope is telling priests to stop using it.

MSN: Pope implores priests to stop writing sermons using ChatGPT

And software developers are having a bit of an identity crisis about it.

AI has taken over my day-to-day. I have multiple ChatGPT threads open almost constantly: dev work, random ideas, even running chats on my houseplants.

I recently spoke with someone leading data science at a major multinational company, it’s basically him and four AI agents running the show.

Usage is exploding, but at the same time we’re seeing social media fill up with AI slop, creators getting frustrated, and developers are split between excitement and outright hostility.

I get both reactions, but this shift feels inevitable.

I think the obvious fear is losing their jobs, with massive tech layoffs right now, that fear is probably founded.

There’s also quality concerns. AI-generated code can be sloppy, especially if they only tried it 6+ months ago, but it’s improving fast.

I’ve talked to a number of dev friends about this, and the deeper issue is craftsmanship.

For decades, coding has been a hard-earned skill. Developers built the most valuable industry in the world by learning how to make computers do useful things. To non-technical folks, coding is black magic, and good devs are basically wizards.

So what happens when that barrier drops, and anyone can just ask a computer to build something?

I think that’s at the core of the resistance for a lot of devs, a loss of identity and hard-earned status.

At the same time, I think a lot of devs are reacting to the current state of AI, not the direction it’s heading.

You see people joking about vibe coded apps and how now there’s going to be jobs just cleaning up AI-generated messes. And yeah, today, that’s sometimes true.

That feels a lot like making fun of early AI images for having six fingers.
That didn’t last very long.

The mistake is focusing on the rough edges today and ignoring the curve.

Because right now, small teams are shipping real products with mostly AI-generated code. Large companies are saying the majority of new code is already AI-assisted. Entire systems are being built faster than teams could have done even a year ago.

So yeah, you can laugh at cleaning up vibe code… but if that’s your takeaway, you might be missing what’s actually happening.

One thing worth remembering: in computing, we’ve always been climbing the abstraction ladder.

No one writes ones and zeros anymore. We moved to assembly, then high-level languages, then frameworks.

How software development works today:
High-Level Code → Compiler → Assembly → Machine Code

Now we’re adding another layer:
Natural Language → LLM → High-level Code → …

The interaction is changing, from writing code to using natural language, but the direction is the same as it’s always been. This isn’t new.

For prototypes and smaller projects, it can generate most (sometimes all) of the code. That changes the game. I test ideas in minutes or hours instead of days or weeks, and ship small stuff way faster.

It’s also a great thinking partner. Not perfect, but good enough to unblock you, challenge you, or help you design.

And it lowers the barrier to trying new things. New languages, new stacks, things that used to sit on a “maybe someday” list are now lunchtime experiments.

The time savings can be ridiculous. I built a small app for a friend, and he later wanted Spanish support. Previously, that would’ve been a painful refactor. Codex handled it in a couple minutes.

Between speed, brainstorming, and offloading a lot of the manual work, it’s hard not to lean in.

AI doesn’t remove the need for developers, but it is changing the job.

If your identity is “I write code and others can’t,” this will feel threatening.

If your identity is “I use technology to solve problems,” this is leverage.

We’re moving up a level, from writing code to directing systems. Less typing, more thinking. Less implementation, more orchestration.

You can still go deep. You can still write and review code. But now you also manage agents, shape outputs, and define problems more clearly.

In a way, we’re all becoming a bit more like product managers, whether we like it or not.

Power tools didn’t replace construction workers, they made them faster and more capable.

AI is the same.

Use it when it helps. Ignore it when it doesn’t. Write code by hand when it makes sense.

But don’t confuse today’s rough edges with the long-term direction.

This isn’t going away.

So you can fight it for a while…
Or you can embrace it, and go build everything on your weekend project list.

Tech Spot: BYD Breakthrough