I’m seeing a lot of companies say that they are measuring AI usage as a performance metric.

The assumption is AI usage == efficiency.

I think this is a poor metric. They should try to measure actual efficiency metrics, not tool use. See exhibit A:
https://arstechnica.com/ai/2026/05/amazon-employees-are-tokenmaxxing-due-to-pressure-to-use-ai-tools/

Executives bragging about coding again and shipping production code with AI because they can misses a huge massive point.

Is this really the BEST use of an executives limited time and mental energy?

Sure, it makes them feel productive and “in touch” but the role of a executive should be strategy, setting direction and priorities, hiring and mentoring their staff, holding teams accountable, and enabling their teams to be successful.

This is ego and shiny toy syndrome talking. It’s a distraction from critical high leverage work only they can do.

If you want to fast, go alone. If you want to go far, go together.

So, Brian Armstrong, CEO of Coinbase, published a letter on X about his companies future and his planned layoffs.

You can find the full post here: https://x.com/brian_armstrong/status/2051616759145185723?s=20

So many folks, rightly so, have zeroed in on this sentence with serious angst:

"Non-technical teams are now shipping production code..."

I think this is the inevitable outcome of the past 30 years. First cloud, then SaaS, now vibe coding has moved IT ownership to the masses.

I don't think this is great for security, governance, or oversight, but it's AMAZING for CEOs and boards who just want to go fast and break things and "empower their people."

I'm not belittling "the masses." But they aren't technologists, by and large.

And what is being demanded of them by misguided leaders is to run some massively complex SaaS/Cloud/Coding tool that "Looks Easy Enough" but all of the devil is in the details that only hard core technologist would know or care about.

I believe this is why we have seen so many breaches based on misconfigurations and poor secret management and poor API/Token/Oauth management. The people making those design decisions aren't equipped with the skills to understand the consequences of their design choices.

They are marketing people, or sales people, or HR people, or whatever. They have other important skills, but we have forced IT onto them because leadership massively underestimates the complexity, risk, and specialized knowledge required to run it safely.

"I mean, how hard can a surgery robot be? You just push buttons right? Get the front desk guy to do it!"

This is inevitable, but stupid. Good luck to us all.