Hey!

Chris here. Welcome to Blueprint—the newsletter to help you build a winning engineering team.

I've been thinking lately about what it would actually take to disappear on vacation for a week. Not a "working vacation" where you have one eye on Slack the whole time. I’m talking really and truly gone.

Right now, that feels like a pipe dream. And the more I sit with why, the more I think it's pointing at the real bottleneck for founders trying to build with AI right now.

Let me explain. 👇

📒 DEEP DIVE

The Living Context Layer

How AI is making it possible to hand off the judgment, history, and decisions behind your work.

The Reason You Can Never Actually Check Out

I want to be careful here, because I'm not knocking SOPs. My companies run on them, and up until very recently, they were the best answer we had for getting work off your plate.

They're also still structurally correct: Build a process, document it, train your team against it, and create leverage.

That's the playbook, and it works.

But SOPs have a ceiling.

The batons I'd need to hand off before disappearing for a week include:

• Exactly how to work inside a specific client's system right now

• The decisions I've made in their codebase over the last 3 weeks

• The shortcuts I've developed for their environment

• What we've tried, what we're in the middle of, and why we went one direction instead of another

That's a living, accumulating body of context that no static document can keep up with. And for most of the history of this industry, when a founder walked out the door, that context walked out with them.

That’s why the night before every trip, every founder I know runs the same script: okay, I need to go write all of this down. I need to record videos, write notes, and document everything so whoever's holding the fort actually knows what to do.

But nobody has time to prepare all that context the day before they leave. So it doesn't happen.

You take the trip, you're not really there, and you spend half the week waiting for the call about the thing that only you can handle.

That's the tax every founder pays for knowing more than anyone else on their team.

What's Different Now

When I'm working deeply inside a system—building features, navigating a complex codebase, running AI agents through a project—that entire history is being recorded.

So, what if that recorded history is the baton? What if instead of spending the night before vacation, frantically writing documentation, I just hand someone the keys and say, "Take this, ask it what I was working on, and it'll tell you exactly where we left off."

The context is already baked in. Anyone reasonably trained in the toolset can sit down, ask it where I was, and pick up from there.

Proof That This Works

A few weeks ago, I had a meeting with a CEO I've worked with for a long time. He had four questions he wanted to walk through. I had nothing left in the tank—running on 3 hours of sleep, attention completely shot.

So 5 minutes before our meeting, I dropped a bot into our DM thread. Something I've been building that has absorbed everything I know about his company and the work we've done together.

I dropped in his 4 questions. The bot nailed every one.

There weren’t any tricks, smoke, or mirrors. It was exactly what I'd been telling him this could do. What he needed was my knowledge, not my attention. And that’s what these tools are designed for.

That experience hit me hard. I realized what I actually want is my attention back.

Every day, I get pulled in 40 directions by things that, in a world where context was properly shared, would never need to reach me. The bot handled it better than I would have anyway, because it had the full picture.

What It Takes

Getting to that state right now is work.

You need a memory system that’s constantly recording what's being done. You need a context file structured the right way for your AI agents. You need the tooling underneath it so the system can act, not just describe.

And you need all of that coherent enough so that someone reasonably trained in the toolset—someone who doesn't know your domain—can sit down, get oriented in minutes, and run with it.

That's hard to set up. But it's also work that follows a pattern.

I've done it in enough different environments to know that I approach every foreign codebase and every new client system the same way. Which means the process can be generalized.

That's a bold claim. But I've watched myself do it enough times to believe it.

What's becoming possible, for maybe the first time, is delegating the accumulated understanding of everything you've been doing and why. That's what passing the baton should look like.

Why Humans Still Have to Be in the Loop

There's 1 more piece of this worth considering, because it's the whole reason the context problem matters in the first place.

People don't want to talk to AI agents. Even if they get good enough to fool them, they’d be pissed if they ever found out. Something about human connection doesn't transfer to an AI interaction, no matter how good the AI gets. I'm not going to an AI concert. I'll go watch John Mayer play the guitar in a smoky lounge somewhere, because I have some connection to what it takes for a human being to do that thing.

So the equation is simple. Humans stay in the loop. And the only sustainable way to keep them there—without the founder becoming the permanent bottleneck—is to solve the context problem.

That's the work. That's what I've been heads-down on. And every demo I've shown, every conversation I've had with a CEO who's seen a piece of this, ends the same way: when can I have it?

BEFORE YOU GO…

The problem I'm describing is as old as business itself. Founders have been trapped in their own heads since the first business was ever built.

What's finally becoming possible now is a path out. The context you carry around can stay accessible when you step away. Someone else can pick up the thread and keep going, because everything you've been doing is already there, organized, and ready to answer questions.

We're early in figuring out what this looks like in practice, but the direction is clear.

And the founders who solve it first are going to be the ones who can finally do what the rest of us never could: take a vacation.

Talk soon,

Chris.