One night I walked into an Indie Hackers meetup in Atlanta. I knew no one. By the end of the night, I’d met the person who’d become my accountability partner for the next six years and the group that would become my mastermind for two years. Some of the most important relationships in my life came from that single event.

Months later, one of those same people told me to pivot. That advice broke a three-year streak of zero income.

I had zero control over when any of that would happen. The only thing I controlled was walking through the door.

We talk about dark games as crises. Things that show up and you don’t control when:

• A market crash
• A pandemic
• A platform ban

But breakthroughs work the same way:

• Viral moments
• Unexpected partnerships
• The call that changes everything

The asymmetry is what makes this interesting.

	Finds you anyway	Requires you to play
Negative dark games	A recession still hits. A health scare still arrives.	—
Positive dark games	—	You have to be building, shipping, showing up.

Positive dark games require you to be in the game. You have to be playing to catch the viral moment. You have to be building to have the breakthrough. You have to be shipping to get the inbound that changes your trajectory.

You can try to avoid the game entirely. Negative dark games will still find you. But you’ll miss all the positive ones.

The strategy is two-sided. Build buffers so negative dark games can’t take you out. Low fixed costs, savings, health. Things that absorb shocks you can’t predict. Then stay in the game so positive dark games can find you.

🎲

What’s one room you could walk into this week?

I play speed chess with a three-minute clock. I lose more games by running out of time than by making bad moves. I’m winning the game. The clock hits zero and it’s over.

The best move is worthless at zero seconds.

I’ll see a strong move but spend 40 seconds thinking through whether there’s a stronger one. Sometimes there is. I find it with two seconds left. The search for the “best” move cost me the game.

Speed chess beat the perfection out of me. Hundreds of losses. Losing a game you were winning changes you faster than any mantra about perfectionism.

In a longer game you can think three moves ahead. You have slack. In a three-minute game you play by instinct. You trust your preparation. You move.

The same thing happens in work. A post that takes 45 minutes to write and 3 hours to polish. A product that works but isn’t yet “beautiful.” An email that communicates the point but could be tighter.

“Perfection” is expensive. Every hour you spend polishing one thing is an hour everything else sits waiting.

⏱️

What’s one thing you could ship this week by trusting your first instinct?

I just made a documentary about prediction markets. The thesis was clean: people bet real money on what they think will happen and the crowd’s price tells you what the informed money actually believes.

Polymarket called the 2024 election at 95% while the networks were still hedging. The case for accuracy is strong.

Something kept nagging me while I was editing the documentary.

The model goes back to Friedrich Hayek’s 1945 essay “The Use of Knowledge in Society.” The information relevant to a complex decision is scattered across millions of people. Market prices aggregate those fragments.

When traders put money behind their beliefs, the price becomes a live probability estimate. A thermometer reading the temperature of collective knowledge.

The model assumes the thermometer is passive. It reads. It reports. It stays out of the way.

But Polymarket’s odds shaped the conversation long before election night. Months before the vote:

• Pundits cited them
• Donors saw them and reallocated
• Volunteers recalculated whether it’s worth knocking on doors

The price shaped fundraising, enthusiasm and momentum. By the time it hit 95%, the loop had been running for months.

The price becomes a force in the outcome it claims to measure.

The thermometer changed the temperature.

The documentary celebrates prediction markets going mainstream. ICE, the parent company of the NYSE, invested two billion dollars. Kalshi won its CFTC battle. Robinhood embedded contracts in its app.

Mainstream adoption means more participants, more liquidity, more information. That’s what makes the price accurate.

Mainstream adoption also means more people see the price. More media outlets report it. More decision-makers respond to it. The price starts influencing the thing it claims to measure.

The same force that makes the price accurate moves the outcome.

This pattern shows up in other markets. In the stock market, a rising price gives the company cheaper capital and easier hiring. The stock price becomes part of the company’s value. The counterargument for elections: a ballot is a ballot regardless of what Polymarket says.

True. But voters see the odds and update their behavior.

• Donors give less to the candidate trailing at 35%
• Undecided voters reconsider
• Volunteers skip the rally

The ballot count runs on votes. The voters run on information, including the odds.

George Soros understood this pattern decades before prediction markets went mainstream. In 1992, he saw that the British pound was locked into the European Exchange Rate Mechanism at a rate the economy couldn’t sustain. He shorted ten billion dollars in pounds. His bet created downward pressure. Other traders saw the pressure and followed. The Bank of England couldn’t defend the peg. The pound collapsed. Soros made over a billion dollars in a single day.

He called the pattern reflexivity. His bet predicted the pound would fall. It also helped make it fall.

Prediction markets work the same way. The more people trust the price, the more the price shapes the outcome it claims to track.

I covered prediction markets for Trends.vc and made the documentary.

I believe the core thesis: crowds with skin in the game produce better forecasts than “experts.” I also think it has a hole: what people believe will happen influences what actually happens. The more people who trust the price, the more the price shapes what happens next. Soros made a billion dollars because he understood this before everyone else did.

🌡️

Which of your beliefs is already shaping what happens next?

I asked how great products are measured. The answer is retention. Products with high churn die.

Then I thought about my own story.

Trends.vc has retained me for six years. Writing reports, synthesizing trends, compressing signal. That work is aligned with how I think. I could do it indefinitely.

I built HeadsUp because I wanted it. Competitive intelligence for my own decisions. The product retains me because I use it. If every customer left tomorrow, I’d still run it for myself.

Will customers stay? Ask yourself first. Will I stay?

If the answer is no, you’re building because the market looks good. The opportunity seems big. The TAM is attractive. You’re on a countdown. Three months of enthusiasm. Five months if you’re disciplined. Then burnout. Then you stop iterating. Then the product stalls. Then customers leave.

Customer retention requires founder retention. The founder has to stay in the game long enough to iterate until it works. And “long enough” is usually measured in years.

The products that last are the ones where the founder and the work are in harmony. The work is the thing they’d do anyway. The product is a side effect of who they are.

That is the real retention metric. Not “will users come back?” but “will you come back?”

🫀

What’s one thing you’d work on this week purely because you want to?

I set up AI agents to run tasks while I step away from my desk.

The idea was simple. Let the agents work. Go cook. Clean up. Listen to a podcast. But each time an agent finishes a task, a chime plays. And each time I heard that chime from the kitchen, I’d walk back to my desk to check the output. Two minutes later I’m back at my desk working on something I told myself I’d step away from.

The chime is supposed to help. It tells me when progress happened. But the same sound that keeps me informed during work hours pulls me back during off hours. It trained me to return.

Then I noticed the screen. The display was just there in my peripheral vision. A lit screen in my line of sight is an always-on invitation. I’d look over, see it and walk right back.

So I hacked it. I figured out how to turn off the display while keeping the laptop lid open. The agents keep running. The chime still plays. But there’s nothing to see when I walk over. And turning the screen back on takes an extra step. That small friction is enough.

You can manage attention inside a well-designed environment. You can’t manage it inside one that hijacks it. Every strategy that asks you to outmuscle a trigger that’s always present is a losing bet.

Removing the trigger means the fight doesn’t start. Remove the stimulus. Remove the glow. Remove the chime. Make “checking” boring because there’s nothing to check.

When you want something, remove friction. When you don’t, add it.

The easiest way to stop doing something is to make it harder to start.

🔇

What habit would change if it were harder for you to start it?

I launched Twin to sixty impressions and one like. I moved on.

That was one of thirty-something experiments I ran that day.

If you run one experiment at a time, the result can feel definitive. If you run thirty in a day, each one is a blip.

Same math as poker. Play a perfect hand and still lose. That’s variance. Play a hundred hands and the flukes start to wash out.

Most founders launch once, then stare at the analytics like one result just told them what it means about them. It didn’t. It gave them one data point.

Twin launched to sixty impressions and one like. I also shared it in a few WhatsApp groups and had one real conversation with a devil’s advocate. He gave me two legitimate objections. Those objections became the next two features I built.

So I don’t read that launch as “one like.” I read it as one like, one useful conversation, and two better features.

Running many experiments makes it harder to turn one weak result into a verdict on yourself. Maybe that experiment was bad. Fine. You still have other experiments in motion.

🃏

Which result would feel different if you had 30 others running beside it?

I just caught something I’ve been missing.

When your experimentation rate is high enough, the question stops being whether you’ll succeed. It becomes when.

It’s math.

An experiment is a bet with fast feedback. A tweet, a feature, a cold email.

One experiment a month means one outcome carries all the weight. Thirty experiments means each one carries three percent. You have no idea which one will hit. Twenty-eight fail and you still got 2 wins. Those 2 wins compound. They teach you where to point the next 30.

The confidence comes from knowing you’ll run enough experiments to find out what’s right.

You stop clinging to individual outcomes. Each experiment is just a data point. The attachment drops because the volume makes any single experiment less precious. When each experiment is one of 30, the stakes on each individual one approach zero.

Experiments produce their own confidence.

This is why “be more confident” is empty advice. Confidence is a side effect. Change the input (experimentation rate) and it changes on its own.

This is the law of large numbers. Throw enough experiments at anything. It works out.

⚗️

How many experiments will you run this week?

One guest lost weight on red meat. The next cleared his skin. A third got his energy back.

I gave it a shot. Within days I had headaches so sharp they felt life-threatening. My body handles cholesterol differently.

It’s the same with sales. “Cold-call your way to revenue. Push through the discomfort.” That advice works for people who get energy from direct outreach. I can still convert and persuade. I learn how from people whose brains work like mine.

I call it grain. My cholesterol response is grain. Cold calls drain me. That’s grain.

• Stage
• Energy
• Skills
• Capital
• Location
• Personality
• Risk tolerance

You can work with your grain or against it. Two founders chasing the same outcome with different grains need different playbooks.

Most advice is correct in a vacuum. It becomes destructive when you install someone else’s playbook without checking who wrote it. A bootstrapped introvert studying a venture-backed extrovert’s go-to-market is like running the wrong code on the wrong hardware.

Swipe files work best when they’re filtered by grain. Pricing strategies from solo founders in their first year. Sales playbooks from someone who hates cold calls and still closed deals.

Study success through the lens of people who share your grain.

🧩

What’s a strategy that works for you precisely because of how your brain operates?

Last week I caught myself about to make a one-time fix.

I was building Multi with AI agents. One of them added a high-priority task to the product backlog. It landed at the bottom of the list. The backlog sorted by recency, so the newest task always sat last regardless of importance.

My instinct was to drag it to the top. Five seconds. Done.

But tomorrow another agent adds another task and it lands at the bottom again. I’m back to manually sorting. Each fix is a 1x return.

So I wrote a protocol at the top of the backlog file. Priority tiers. Ordering rules. Instructions for evaluating new items against existing ones. Now when any agent touches that backlog, it reads the protocol and places the task where it belongs.

I did the work once. It compounds across sessions.

I’ve started seeing this across my whole workflow. I work with AI agents on writing, product development, research and distribution. Each conversation is a chance to build something disposable or something reusable. The gap between the two over time is enormous.

Three examples from the last month.

Skills over instructions. I write for my newsletter and publish through WordPress. The first time I asked an agent to publish a post, I gave it step-by-step instructions. The second time, I wrote those instructions into a skill file. Now any agent in any session can take a draft from my editor to a live post on the site: formatted, linked and indexed. I explained the process once. It runs on autopilot.

Process design over process execution. My daily input processing used to be a conversation. “Here’s my voice note, extract the drafts, update my context.” Now it’s a single trigger. Voice notes go in, finished drafts come out, my context stays current. The entire pipeline runs without me repeating a word.

Structured backlogs over ad hoc lists. The backlog story I opened with goes deeper than sorting. The protocol forces each agent to compare new items against existing ones in the same tier, check for overlap and re-evaluate whether current priorities still hold. The backlog gets smarter each time something is added to it.

Each time you solve a problem with an AI agent, ask: am I making a withdrawal or a deposit?

• A withdrawal is a one-time action. You get the result. The knowledge stays in the conversation and dies when the session ends.
• A deposit is a reusable component. You get the result and you get it again next time without repeating yourself.

Skills, protocols, structured documents, management rules embedded in files agents read. These are deposits. They accumulate.

Raw conversations, one-off instructions, manual corrections. These are withdrawals. They evaporate.

The compounding math is brutal. If you make 50 decisions a week about how agents operate and each one is a withdrawal, you make 50 decisions again next week. If 20% of those become deposits, by month three you’re making 10 decisions a week. The system absorbed the other 40.

Most people using AI right now are making withdrawals. They ask a question, get an answer, move on. The next session starts from zero. The AI gets more capable with each model upgrade but the system around it stays the same.

The leverage is in the assets you build around the model. A skill file is a tiny asset. A backlog protocol is a tiny asset. A style guide agents consult before writing is a tiny asset. Stack enough of them and the system operates with less friction each week. The context got richer. The model stayed the same.

That’s the compounding return.

🔮

What’s one process you repeat with AI that could become a reusable component?

One line of code taught me more about brand than a year of positioning work.

I was debugging why HeadsUp emails showed a generic “H” in Gmail instead of our logo. The fix seemed straightforward. Get the logo into the inbox.

Three options.

• Google Workspace accounts for each sender address. $168/year to host profile photos.
• BIMI, the official email logo standard. Needs a $1,000+ certificate, a trademarked logo and DMARC enforcement.
• Send all emails from my personal account. Set the logo as my profile photo. $0.

The cheapest option was also the best one.

At this stage, emails from a person outperform emails from a brand address. Gmail itself prioritizes “people” mail over system notifications. Open rates favor real names. Trust favors founders.

“Dru Riley from HeadsUp” with a brand avatar handles both signals. The name builds trust. The logo builds recognition.

The switching point comes later. You move to a brand sender when customers no longer know or care who the founder is. That’s a graduation, not a starting point.

I spent an hour evaluating infrastructure I didn’t need. The answer was the thing I already had.

🔬

What’s a case where the simplest option turned out to be the best one?