Intel pulled off a great marketing trick: they made you care about a chip you would never see. The sticker went on the outside of the laptop. The processor was sealed inside. You had no idea what it looked like or how it worked, but “Intel Inside” meant the thing was engineered well, and that somebody was accountable for the core.

I want a version of that sticker for my own work. Call it “Alex Inside.” Not because I wrote every line, because increasingly I didn’t, but because there’s a person behind the judgment and the direction and the calls that matter. The chip is impressive. Someone still has to be the computer.

The sticker is easy to print. Earning it turned out to be harder than I expected.

The part where I lost the thread

A while back I handed off a significant chunk of a project to AI. Not boilerplate. The kind of logic that shapes how the thing behaves. It went smoothly, the output was good, and I moved on.

A few weeks later I needed to change it. New constraint, new information, rethink the logic. And I realized, with a quiet sort of dread, that I couldn’t. Not because the AI had written something bad. It hadn’t. I couldn’t change it because I had never owned it. I approved it, I didn’t understand it. I agreed, I didn’t decide.

How do you evolve an algorithm you didn’t write, and didn’t even ask for very carefully?

That question stuck with me, because the drift isn’t dramatic. You don’t wake up one day unable to do your job. One project you hand off 20%. The next one, 40%. Then you notice you’re reviewing AI output more than you’re thinking, and you’ve quietly outsourced the part of the job that keeps you sharp. The part that builds intuition. The part that lets you change direction when the direction needs to change.

The algorithm doesn’t drift. You do.

AI for AI’s sake

There’s a less personal version of the same drift: bolting AI onto things that didn’t ask for it.

Someone in the meeting says, “Should we add a chat interface so users can query their data?” It’s 2026 and there’s an LLM API key sitting right there, so the answer comes out as “yes, obviously.” The question worth asking first is whether it helps anyone.

Information has never been easier to find, which means finding it isn’t the skill anymore. Distilling it is. Not asking Jeeves, but knowing which question to ask and what to do with the answer. A chat box that lets users type “what happened last Tuesday?” demos well. If what they need is a weekly summary that surfaces the three things they should act on, the chat box is a party trick with a monthly bill.

Someone has to decide what matters. Someone has to make the call about what the user needs versus what they typed. You can’t put that in a prompt.

The best AI integrations I’ve seen are invisible. Real work is happening under the hood and the person using the product just gets a cleaner experience, like the product understood them. The worst ones are AI with a voice skin on top. Somebody decided the experience should be AI because AI is the thing right now, not because it was the right answer.

Where are you in the loop?

There’s a version of “human in the loop” that is mostly decorative. You’re in the loop in the sense that you clicked approve. You read the summary. You nodded. The loop would run the same without you. You’re there for liability reasons.

Then there’s the version that counts.

The difference is asking and agreeing versus directing and evaluating. Asking and agreeing looks like: “Write me a strategy for X.” Read it. Seems reasonable. Ship it. Directing and evaluating looks like: knowing enough about X to shape the input, reading the output with a critical eye, pushing back on the parts that don’t sit right, and building on the result in a way you can adapt later when the situation changes. It’s the difference between an opinion you can defend and a summary you half remember.

This matters more when other people are involved. Clients, users, teammates. People don’t want polished output so much as they want to feel heard, and trust doesn’t transfer through an AI’s eloquence. It transfers through the sense that a real person thought about their specific situation and gave a damn. A beautifully worded recommendation that clearly came out of a template, even a good template, does not land the way a recommendation from someone who was paying attention does.

You can automate a lot of the generation. You can’t automate the relationship.

Alex inside

So here’s where I’ve landed.

AI is in nearly everything I do now. Research, drafts, code, proposals. It’s in the stack whether I name it or not, and I think that’s fine. Better than fine. The output is usually better and faster than what I’d have produced grinding through it alone.

What I keep coming back to is the sticker on the outside.

“Intel Inside” worked because Intel stood for something. The chip did a specific thing, did it well, and the company was on the hook for it. The sticker was a promise.

“Alex Inside” is a promise too. It means I directed this. I evaluated it. I can defend the reasoning, and I can change it when the situation changes. I’m not the person who approved the output. I’m the person who owns it.

The AI does real work in there. I’m still the processor.

Using AI well has less to do with how much you hand off and more to do with staying the kind of person who can hand things off and still know what’s happening. A loop with no human actually in it isn’t a loop. It’s a machine running until it breaks, with nobody quite sure how to fix it.

Don’t lose the thread.