What finally made me commit to raising my lobster (running OpenClaw) was reading Dan Martell’s Buy Back Your Time. The only difference is that I’m not hiring people. I’m hiring AI.
There’s one core idea in the book that I really like: don’t hire people to grow your business, hire people to buy back your time.
Sounds obvious, right? But think about it and it’s actually very different. Most people hire with the logic of “I can’t keep up, so let’s get someone to share the load.” What Martell says is “you should free your time from low-value tasks, then spend it on the things that really matter.”
The book has a concept called the “Replacement Ladder”: you hand off tasks level by level. The first things to go are the most draining Lv1 tasks: admin chores, the inbox, calendar scheduling. The core logic is not to skip levels. Clear out the busywork that drains you most first, and only then will you have the room to deal with higher-level strategy or building.
OK, that’s the theory from the book. Back to my own hands-on experience.
Once I started running my lobster, the very first thing I did was hand it the Lv1 admin tasks. The morning newsletter digest, cron reminders, data cleanup: things that have to get done every day but don’t take much skill. I gave all of them to the agent.
The result was obvious. I save about two hours a day, which comes to somewhere over ten hours a week.
What do I do with those ten-plus hours? Write, improve the product, think about strategy. I used to feel like I never had enough time, but it really wasn’t a lack of time. It was being eaten up by admin chores.
Of course, AI isn’t the same as a human assistant. It won’t think ahead for you, and it won’t take one example and apply it to others. In Martell’s terms, it’s worse than an intern, because an intern at least asks you questions. My lobster doesn’t ask. It reads your mind on its own and hands you a result that looks complete but may be missing bits here and there 😂
So you have to manage it with the book’s 10-80-10 rule: you define the first 10% (the goal and the spec), let the AI run the middle 80% (the execution), and then you spend the last 10% checking and fixing.
Here’s what I actually do:
- Do it by hand once to confirm the process
- Write it up as a Skill for my lobster to follow
- Run it once and see whether anything breaks
- Review every day and fix whatever isn’t clear enough in the prompt
This process is really Martell’s “Buyback Loop”: Audit (see where your time goes) → Transfer (build an SOP and hand it off) → Fill (spend the freed-up time on high-value work).
The only difference is that he’s talking about hiring an admin assistant, and I’m raising a lobster. The cost differs by more than ten times, but the core logic is exactly the same.
Personally, I think the most valuable thing about the book isn’t teaching you how to hire. It’s giving you a framework for deciding: is this worth doing yourself? If the answer is “no, but someone still has to do it,” then find a way to hand it off. In 2026, there’s one more option for who you hand it to: AI.
To me, my lobster is an admin assistant on a salary of US$20 a month (maybe more). It makes mistakes, needs watching and has to be taught step by step, but it works 24 hours a day without a break. And as your Skills and org structure get better and better, it gets more and more reliable. That said, there’s no denying that a lot of things can still only be done by people.
In the AI era, “buying back your time” doesn’t have to mean spending a lot to hire someone. Sometimes spending a bit of time raising a good lobster works just as well. These days I’ve also started giving it tasks that aren’t purely admin. It’s an addictive process.
Do any of you have similar experiences? How much time have you bought back with AI? Come chat in the comments 🙌
#BuyBackYourTime #AIAgent #OpenClaw #BuildInPublic