MCT oil. Creatine. L-theanine. Three things in the Amazon basket this week. I'm enjoying morning coffee with Kirsten, a ritual that is becoming a favourite. I'm still reeling after coming off a four-to-five-month keto run. I say reeling because at first I hated it. Then it started to grow on me. A small experiment in ai for founder productivity, in a way.
It's simple. About ten things you can eat, or that I ate, and I rotated around them. Mornings: bulletproof coffee. Two tablespoons of grass-fed butter, a tablespoon of coconut oil, a tablespoon of MCT, a teaspoon of cacao, a splash of heavy cream, blended hot. Plus strong coffee of course, and that was coming from being a strict tea drinker for many years. Intermittent fast until 1 or 2, usually a skin or a workout before then. (Skinning, for non-skiers, is climbing up a slope with skins stuck to the bottom of your skis. Ski touring around the Swiss alps was 70% of my winter.) I'd check the blood about 45 minutes after a 50 to 60 minute workout, or longer on a skin, and watch the ketones peak. Then some kind of salad with protein, or yogurt with nuts and berries. Dinner the same: a meat rotation on the grill. Salmon, tuna, beef filet, lamb. Chicken was the hardest to come back to. The dirty bird. Four years vegan before the cancer diagnosis put meat back on the plate, and the dying vegan in me had seen too many Netflix factory-farm films (Cowspiracy, Dominion, What the Health) to finally crack on chicken last.
Fortunately, most meat in Switzerland is grass-fed, bio, or locally sourced. The local butcher down in the valley knows the farmer. That feels better than the alternative: factory production with injected hormones and antibiotics.
The main thing I enjoyed about keto was the mental clarity. It is real. I needed less sleep. I'd wake at 5am, ready to start the day, and most mornings I'd start with writing. The crispness expressed itself in wanting to spill thoughts onto a page. For three months of that stretch I was also doing a lot of driving, an hour each way from our mountain ski station down into the valley for salvage radiation therapy for prostate cancer. These were some dark times. The driving was time I couldn't get back, but I filled it with podcasts. Tim Ferriss for the long-form interviews with high performers across science, business, and sport. Alex Hormozi for tactical sales and offer-design. Steven Bartlett's Diary of a CEO for the founder conversations he pulls his guests into. Bright spots in the drive.
The actual reason the times were dark was the toxic work environment I'd let myself get pulled into. A previous role as a COO at a startup. The founder getting more and more exasperated at his own failings, going deeper into investor debt, taking it out on the team. For my own sanity I cut ties. Immediately. The unwind is painful and I am still working through pieces of it today. Ninety days from now it should be behind me, and worth it. Like sitting on the radiation table and getting zapped. My PSA is undetectable at 0.01, the lowest reading the test gives. Touche bois it holds. I am doing my best on the low-glycemic side, carbs under 150g a day, and the rest of the protocol.
Today, in our cosy Seignosse, FR setting, I am trying to find the cocktail without the restrictive keto rules. According to my research, MCT oil, L-Theanine 200mg, and Creatine 5g is the right balance. Some of the clarity is already coming back since starting the L-Theanine, and I want to get as close as I can to the head I had on keto.
This is the smallest of three rearrangements happening at the same time. The other two are bigger.
Long-time readers (4 currently right now, and not even my mother) know the through-line of this build-in-public. My SaaS days are over. I am creating my own way of life through work on my own terms, all in public. The writing inspiration comes and goes, which I suspect is normal. The crazy decisions are the through-line of the story. Walking away from a software company I founded because I could see where agencies were heading. Going all-in on fractional. Saying no to clients who could not pay before I started.
Sunday's MCT order is the smallest of those decisions and still crazy in its own way. A founder running a two-man team, buying his clarity with AI-researched supplements in an Amazon.fr cart, in the year he is trying to ten-x the business.
Fifty-some-odd days into this experiment: pull the rip cord, cut every slightly bad-vibing thing out of my life, reinvent myself completely, live 100% by feel and instinct.
The year ONE Thing for TestVentures is $100,000 a month with four to five specialists underneath me. We are at $11,800 a month and a team of two: me, and Jan part-time, the CTO I hired a week or so ago. The gap is roughly $88,200 a month and three specialists.
That gap closes one of two ways.
We can keep doing what we are doing. One enterprise client. Jan deep in the build as code reviewer and second pair of eyes. Me on strategy, on the AI agents underneath, on the daily writing. Steady. Honest. Two of us doing the work of a small team. The numbers tell you what that path produces: about $11,800 a month, slow growth, more code, more reviews, more posts. Linear. Good. Not $100K.
Or we switch desks.
Starting as soon as revenues justify, Jan goes full-time on fulfilment. He carries the AI Financial Controller build through the next few months for the current enterprise client, then picks up engagements two and three as we land them. His standing job today is to read my code, find what is wrong, and tell me. His new job is the same thing, plus running the build on the next two clients as senior delivery. He has earned this. The eighteen findings he caught yesterday on the live AI agent are the proof. Three were shipped the same day with twenty-one new tests behind them.
I go full-time on the other side of the company: writing, client relations, sales, and marketing.
Writing the daily BIP. Running prospecting calls and discovery. Working the AI marketing employee I am building with Ainoa on Upwork, live in seven days. Picking up the phone when an inbound lands. Closing. The job between now and September, ideally sooner, is to find the next two enterprise clients and the two after that. The job between now and December is to have four to five specialists underneath me, doing fulfilment, while I write and sell.
The switch is uncomfortable. I have been the person in the code for the last few months. Pricing the work. Building the agents. Testing. Deploying. The reason it is uncomfortable is that I know how to do it and I trust myself on it. Going back to sales and marketing as the day job means trusting Jan with the build, trusting the marketing AI automations we are building for prospecting, and trusting the writing to do the rest. So far, despite spending hours per day writing, I have not had a single lead from it. I know BIP writing is a long game and the long game needs faith. With AEO, GEO, and SEO shifting daily, I also know the rules that built my faith in organic writing could be unfounded. They were based on the rules of 2013, and things are changing fast.
Three trusts to hold for the next six months. This is where the MCT oil, L-Theanine, and creatine come in.
Sales and marketing as the full-time job is a different kind of work than building. More conversations. More cold reads of inbound. More public-facing writing. The afternoons get harder, not easier, when the morning is already two prospecting calls and a 500-word post. The brain at 2pm matters more in this work than it did in the build work, where the code told me when I was tired.
The supplements are not the magic. The sleep, the daily walk, the meditation streak (2,985 days or so in a row) are. The supplements are a small bet on the same feeling I had a few months back, when the blood was reading 70 fasting glucose. If three things from Amazon move me a few percent toward the head I had then, the bet pays.
The bigger experiment is the job swap.
If you are a C-suite buyer reading this and thinking about fractional AI ops or AI deployed as a service, what you are watching here is the part of the engagement that almost never gets shown. A two-person team rearranging itself to be ready for clients three and four. The build side staffs up first, then the sales side. The order is not arbitrary. Without a strong builder you cannot promise an enterprise client a working AI agent in production in six weeks. Without a strong seller you cannot find that enterprise client in the first place. We are doing both, in this order, in public, with the scoreboard updated at the bottom of every post.
Tomorrow I sit down with the ai employee we are building and lock the prospecting cadence on the OpenClaw fork. Tomorrow night I sit down with Kirsten and re-do the household budget through the back half of 2026, both cases: the year-end goal formalises, the year-end goal does not.
I may be buying clarity at the supplement aisle, switching my own job from the build side to the sales side, betting on a CTO I have solid experience with. It feels right.
Day 55 of 365.
Monthly Revenues $11,800 | Clients 2 | Prospects (AI marketing employee live in 7 days) | Team: Me + Jan (CTO)