AI is like SEO

Developers don’t like AI for the same reason they don’t like SEO; they resent being forced into a black-box discipline with rules that change without notice. They have spent years or decades honing their craft, becoming masters of logic and exactness. Then a new tool arrives that produces the same output, but only 99% of the time. 99% is not exact, and so it must be incorrect.

I’ve noticed that employees and employers fall into two camps.

The staff engineer is frustrated that their company is going all in on code-slop generators which will likely lead to maintenance and security nightmares. They have honed their craft over years or decades and can see the subtle mistakes these AIs make, and they feel their craft is being devalued as now “anyone” can be a developer. Not only is their craft being disrupted but their tools are too; they now have to use “co-pilot” for everything, or respond to walls of AI-generated text in bug reports and email chains.

Then we have the business leader who sees the hockey-stick uptick in AI’s software engineering prowess. The Anthropic bulls are saying it will only be years before 80% of white-collar work is replaced, and if they don’t jump on the trend their business will go bankrupt. Do they really need to be paying developers this much money? Do they actually need this many developers now that one developer with five agents is a force multiplier on their own?

How I’m dealing with this

I’ll be the first to admit I’m not a strong developer, but I’m a good generalist, and I’ve spent a lot of time doing SEO for companies previously. My plan is to continue as is: treat these AI models as a powerful addition to the toolbelt, and keep experimenting, pushing its boundaries to see where it shines and how much leash to grant it. Ignoring it is not an option. Knowing what you want from it is the whole game.