AI doesn’t lower the cost of building the wrong thing
AI is rapidly lowering the cost of building software. But it doesn’t lower the cost of building the wrong thing.
I had a client ask me recently whether they should use AI to port an aging application from a language that has largely fallen out of favor to JavaScript.
Today, I came across an interesting project where someone used an LLM to port Command & Conquer: Generals to macOS, iOS, and iPadOS.
Technically, it’s impressive.
What caught my attention wasn’t the engineering. It was this line:
“No game assets are included or distributed. You need your own copy (Steam, ~$5 on sale).”
It reminded me that building software is only one part of creating value. As software becomes cheaper to produce, the quality of decisions becomes more valuable than the quality of implementation.
AI is rapidly lowering the cost of building software. But it doesn’t lower the cost of building the wrong thing.
Whether you’re modernizing a legacy application or porting a classic game, the first question shouldn’t be, “Can we build it?”
It should be:
- Should we build it?
- Does it solve a real problem?
- Is this the right tradeoff?
- Will customers actually care?
I think that’s where experienced engineers, product leaders, and executives create the most value now. Not because they write code faster, but because they help organizations make better decisions before the code is written.
I think this is one of the lasting changes of the AI era. As implementation becomes cheaper, judgment becomes more valuable. I’m seeing it over and over again.
The repo I mentioned: Generals for Mac, iOS, and iPad.
Originally posted on LinkedIn.