Vibe-Coding: Will Software Engineers Disappear by 2028?

Okay, but we're in this world where now even needing a developer to essentially come on as, let's say, a software developer who potentially needs to put an idea into an AI software and then obviously like prompted to create this code.

But from an enterprise perspective, what is the change about the way companies should think about software development? Where is the skill that should not go unnoticed?

Yeah, so most enterprises over the last 10, 20, 30 years, especially with the whole digital transformation wave, most of us have not had enough engineers, software engineers.

We're all hiring software engineers very rapidly because the primary way we drive efficiency in our company is via software.

You know, if you look at a Coca-Cola, for example, there's not that much innovation really in how much cheaper Coca-Cola to manufacture on an annual basis.

But Coca-Cola is actually getting more profitable because of software.

They're figuring out how to optimize their processes, deliver more Coke, et cetera.

And so software has kind of been sort of the driving growth lever for enterprises over the last two, three decades.

And so as a result, we've all been hiring so many software engineers.

Well, now, as you noted, what has changed is that suddenly there's this new technology, LLMs, that actually is better at writing code than even software engineers are, actually.

And it's much faster than software engineers as well.

And so what we're seeing now is a lot of enterprises are saying, is it really worthwhile for me to hire so many developers? Because first, my developers could be a lot more productive now with AI and LLMs.

But secondarily, I think probably in the next 18, 24 months, software engineering may itself actually go away.

And what you're going to have is effectively a bunch of knowledge workers, so non-software engineers, people that work in marketing, people that work in sales, people that work in operations, they will all be building software.

And that's actually a pretty scary new world, if you will, because software engineers are very well aware of what guardrails they need when they go build software.

No software engineer will deploy to prod immediately, you know, without CoderView, for example.

But a non-engineer might do that.

And so I think setting up the guardrails then becomes really, really important.

You actually led me to my next question quite easily because you mentioned everyone in the enterprise might suddenly become like this developer of a sort.

So if everyone in that case can suddenly become a software builder, you also have potentially thousands of applications being created outside IT's line of sight.

And that leads us to my next question about guardrails itself.

What happens when the speed of building software outpaces an enterprise's ability to govern it? How do you set up those guardrails?