A Value-First Summer · Solstice to Equinox
I built a 20-sided die for my summer. Here’s the game, played in public.
Twenty initiatives in motion at once — on purpose. The advice was always “pick one thing.” This summer is the opposite experiment, lived in the open: breadth made safe by an engine that keeps every thread visible and closes the ones that matter. Watch the value accumulate, all season.
The thesis
Threads don’t go stale because there are too many of them.
The thing I’ve heard my whole career — from consultants, coaches, founders — is: pick one thing. Focus is the path. Too many threads and you’re scattered, ineffective, about to drop everything.
Here’s what I’ve actually found to be true. Threads don’t go stale because there are too many of them. They go stale because they’re invisible. The constraint was never the number of things in motion — it was the cost of keeping each one alive: the overhead of remembering where you left off, deciding what gets attention today, noticing when something’s been quiet too long.
That’s a tooling problem. Not a focus problem. And this summer, I’m proving it in public.
The engine
Anything you make has three layers.
The surface is what everyone sees — the website, the post, the deck. For most of my career, this is where the days went. It’s also exactly the layer AI now produces for almost nothing.
Underneath is the substrate — the structured record of what was said, decided, and learned. The memory. Kept well, you never reconstruct where a thread stands. That used to be expensive too. It isn’t anymore.
And under that is the irreducible part: the curation. The human read — the judgment that looks at a finished-looking draft and says, “No, the real next move is this.” It’s the one thing AI can’t do, because it can’t be the person the work is for. The curation layer was the whole product the entire time; we just couldn’t see it, too busy hand-making the other two layers to notice.
Once you see that, the old math on focus expires. When the surface and the substrate cost real human hours, breadth was reckless. It isn’t anymore.
The reframe
The real constraint was never the number of threads.
My own CFO raised the objection before I could: you cannot give twenty things your full attention every day. He’s right — and the die was never about doing twenty things a day.
A thread doesn’t go dark because you have too many — it goes dark because it goes invisible. The cost of picking it back up climbs so high you quietly never do. So I built the thing that pays that tax for me: one structured memory that indexes every conversation and decision, a leadership team and an organization of agents doing the surface work and the structuring while I sleep, and a cockpit whose only job is to walk up each morning with the next decision on each thread already framed.
The agents keep every thread visible and warm. I show up and curate — a couple of clicks, ten seconds of voice — and the thread moves. Fifteen minutes of judgment, multiplied across a team that never sleeps, is not fifteen minutes of output. That’s the whole trick.
The ritual
A twenty-sided die, pointed at the one failure mode that kills good work.
Two things happen each morning before I roll. First, a relationship sweep — anyone I’ve gone quiet on comes before anything new. People before product, always. Second, anything genuinely time-bound gets handled. Those are the rules of the house, and the roll never overrules them.
Then I roll. The die picks which star of the constellation gets real attention that day. Twenty are in motion across the summer; only a handful are ever live in the queue at once. Because the roll is random, it eventually visits the quiet ones I’d never choose on my own — the threads my attention naturally starves. That’s the entire design: randomness aimed precisely at the one failure mode that kills good work. Going dark.
All twenty faces are real, named things. For most of the season the twentieth was held back — you only reached it on a natural 20. It came out on September 2: it is Value-First Markets, the Value Market, and it has been public since. The fog is off; the die is the same.
The live value-accumulation
The value accumulates here, all summer.
Five strands of value rise and gather across the season — energy, attention, frequency, serendipity, and momentum. This is the shape they pour into. Two of the five carry real readings across the whole season — energy and frequency, one point for every cycle the morning pass actually ran, from the solstice onward. Attention and serendipity stay empty because both fill only when a person confirms something, and that confirming step never ran. Momentum is the counted one, and it sits below. We hold the scaffold rather than draw a curve for a strand with nothing in it. Published September 20, 2026.
- Energy
The spark a thread carries when it is alive and worth tending.
- Attention
Where the day's focus actually landed — and which threads it reached.
- Frequency
How often a thread comes back into the light, so it never goes dark.
- Serendipity
The unplanned moment that mattered — the kind only human judgment settles.
- Momentum
The one countable strand: how many threads actually cross the line.
Four of these strands are felt, not counted — energy, attention, frequency, serendipity. Only momentum is a number, and you can see it just below. Energy and frequency read from every cycle the morning pass ran. Attention and serendipity read empty, and empty is the true reading — they fill only when a person confirms a move, and nobody did.
The one honest number
The season set itself one hard rule.
Four strands you feel. This one you count. The rule, written in June and unflattering on purpose: Each month has to realize more than the last. Not how much got started — how much actually reached someone.
The baseline it was set against
18
June, to the 20th
May realized 4. Read live on 2026-06-20, the day the season opened.
Where the count stands at the close
No month after June carries a confirmed figure. That is an absence, not a zero — it does not say nothing was realized. It says the confirming step never ran, so the season cannot tell you whether it met its own rule.
15 realizations are on the record waiting for that confirmation. Confirming one is a person’s judgment, never an automatic mark — which is the whole design, and also why the number stopped.
What that adds up to is the retrospective’s to say, and it gets written after the close, not before it. The findings live here.
Baseline as of 2026-06-20 · standing as of 2026-09-20, derived from the board’s own record rather than typed in.
Where it’s headed
Not a pitch I’m preparing. A proof I’m trying to earn.
It all comes to a head in late September, in a room full of moonshot-minded builders. My posture there is not a deck. The framing that gave me permission for all of this: agency is the point, and you have to be willing to create value first.
The whole summer is the work; the gathering is just the day it gets read by people whose judgment I trust. Everything between now and then is me getting honest about whether the engine actually carries the load — or whether “focus on one thing” was right all along. I genuinely don’t know yet. That’s what makes it worth writing down.
The close
The season runs solstice to equinox. The equinox is September 22.
A season with an end date is a season that can be wrong about itself. This one closes on the equinox with a live stream and a published retrospective, and an autumn season on the same model is declared at that close.
In June the summer wrote down what it believed, in three claims, each with the condition that would prove it wrong. Those conditions are the whole point: a claim you cannot lose is not a claim. The page below holds the frame and the evidence as it comes in — and it holds no verdict, because the season has not closed yet and deciding the answer early would make the test worthless.
Follow along
The season closes on the equinox. Be there for the read-out.
Here is the honest part, since this page is about being honest: this said “a weekly dispatch” all summer and the weekly dispatch never shipped. The season got written, not mailed. So the promise now is the one that is actually being kept — the retrospective when the season closes on September 22, and the next season’s board when it opens. Not a highlight reel. The findings, including the parts that did not work.