AI Removes the Cover

The question isn't whether AI helps your company. It's who inside it actually benefits - and who quietly absorbs the cost.

A quick note before we get into it.

I’m slightly changing how I write here. Shorter pieces, more often. More of what I’m actually thinking about as it happens, less waiting until I have something fully formed.

Some context on where I’m coming from: I’ve been vibe coding for more than three months now. It started with building things in an IDE, watching ideas become real faster than I ever thought possible. By now, much of my work has moved into the IDE, triggering skills and workflows for non-coding tasks.

At the same time my main job is to help each of the PIRATE companies be better off six, twelve, eighteen months from now. That means thinking about AI constantly. What it means for the businesses. For the people in them. For our clients.

It keeps me up at night, honestly. Not because I think it’s bad - but because I think it’s the biggest change I’ve faced in my working life, and I’m still figuring out what that means. Some days I look at all of this and feel genuine excitement. Other days the lens clouds over - not with fear, but with the quiet pressure of knowing this actually matters and not yet having all the answers. Usually both.

I read a lot, build a lot, have a lot of conversations - founders, leaders, people who keep calling and asking where this is headed and what it means for their business. I don’t have clean answers. I have a perspective, some useful frameworks, and a lot of honest uncertainty.

That’s what I want to share more of here.

I write in three areas now: longer essays on making business more human, shorter pieces on AI and the human organization, and a founder’s lens on building. What follows is one of the shorter ones. Would love to hear what you think.

AI doesn’t help or hurt companies. It helps or hurts specific people inside them. Those two things are not the same, and the distinction matters.

The efficiency gains from AI tend to flow upward. They show up as faster output, leaner headcount, or compressed timelines. And they eventually show up in margins and board decks. The person whose tasks just got automated doesn’t usually capture those gains. That’s not a controversial observation. It’s how value has been distributed through most technology shifts so far.

But there’s something less discussed happening underneath that.

AI accelerates whatever behavior a company already exhibits. Good judgment compounds. Poor judgment scales into expensive mistakes faster.

But this assumes companies have a clear direction that AI simply speeds up. Most don’t. What AI amplifies isn’t a company’s strategy - it’s its current behavior, including the politics, inertia, and confused incentives that strategy documents tend to paper over.

More precisely: AI accelerates exposure.

The manager who seemed valuable because they processed information quickly and produced polished decks has lost that specific edge. What’s now visible is whether there was any actual judgment underneath. AI doesn’t create judgment. It just removes the cover that hid the absence of it.

This plays out differently depending on where your edge actually lives.

If your edge was execution speed, AI compresses it - your competitors get the same tools. If your edge was taste, relationships, domain knowledge built over years, or data nobody else has, AI becomes amplification. The gap between you and others tends to widen, not close.

The engineer who knows where all the bodies are buried. The salesperson whose relationships run deep. The designer with genuine taste. These people tend to get more powerful. Meanwhile, roles built around executing well-defined tasks quickly are under real pressure.

Same technology. Same company. Very different experiences depending on where you sit.

The question worth asking isn’t whether AI helps your company. It’s who inside it actually benefits - and who is quietly absorbing the cost.

It’s an uncomfortable question. Which is probably why it doesn’t come up enough.

🙏

Be kind,

Manuel

PS: Do you prefer shorter posts or longer posts? What would you want me to write about more? Would love to know. Just send me a quick message.