A couple of weeks ago, I shared an equity calculator I’d built with you. Today, I want to share a lesson I learned from the experience and how it allowed me to connect the dots to help a BTG member reframe the way he looked at solvers.
While it’s not a main focus these days, I spent a good three months to start the year deep in the “vibe coding” rabbit hole, because I believe these tools will change the world and I thought it would be wise to know what’s going on. Plus, it was fun!
The equity calculator was one of the first things that actually worked – emphasis on the word "actually."
Before that, I built a lot of things that didn't.
The Graveyard
When I first started vibe coding, I was a kid in a candy store. You could just... tell the AI what you wanted, and it would build it. So I did what seemed natural. I said something like: "I want a personal app that saves ideas and tasks for me, imports my emails and calendar, and has a bunch of cool features. Suggest some features."
And it suggested features – good ones that I might not have thought of.
So I said, "Great. Build them."
And it did.
"What else could we add?"
More suggestions. More features. I kept going.
And I ended up with an app that had a ton of stuff in it. A dashboard, notifications, some kind of tagging system, analytics, a no-longer-useful daily briefing via Telegram that I still haven’t turned off. The problem was that most of these features didn't work properly, several of them I'd never use, and the app grew so big that trying to fix it was overwhelming.
The AI produced a lot, but it did it without me directing it, because I didn’t yet know what I actually needed. It built a bunch of stuff that was technically cool, but useless in practice.
A friend of mine is very into AI now, learning to build things, and making the same mistakes I made. I explained what I’d learned to him using a silly analogy that made sense in my head:
Imagine you're building a log cabin, and you have a team of infinite, super-strong and shockingly obedient 5-year-olds.
If you tell them exactly what to do, and you keep an eye on them, correcting and directing them when they make a mistake, you’ll end up with a beautiful cabin.
If, instead, you tell them, "build me a cabin" and you leave… who knows what you'll end up with. And when it leaks, or the roof is sagging, you'll probably have to take it apart and put it back together eventually. Or, more likely in my experience, abandon it. (My experience with AI, not my decades in the log-cabin industry.)
The equity calculator functions because I knew exactly what I wanted, I understood how it worked, and I was clear and detailed. I sat down and said: build me this specific thing, with these specific inputs, and this specific output.
I started with a simple version, and made sure it worked as I expected it to. When it did, I added the next thing. And the next – one piece at a time, always tested before moving on.
In the first case, I was following the tool. In the second, the tool was following me.
Check the Solver
A few days after I'd had this conversation with my friend, I was on a coaching call with a group of BTG members. Jerry, we’ll call him, had come to poker from the business world, and he was struggling with solvers. Specifically, he was struggling to learn “a full gameplan” and implement GTO strategies at the table.
By the way he spoke, I could tell that when he played, he was anxious about whether or not his plays were “right,” aka solver-approved. And, from my own experience and others’, I know how expensive that can be.
Jerry was making the same mistake with solvers that I'd made with AI: He was treating it as the one in charge.
I was trying to explain something I've explained many times and in countless ways, but this time – maybe because Jerry ran a business, or maybe because I'd just had that AI conversation – it came out very differently:
"You don't work for the solver. The solver works for you."
You’re The Boss
I share this with you because for some of you, that will be the framing that clicks. One of the main lessons I’ve taken away from teaching is that different people respond to different versions of the same lesson. I’ve talked about not trying to memorize and execute exact solver strategies before, and I’m sure I’ll talk about it again some day, because it’s one of the biggest mistakes I believe players make these days.
If you’re like Jerry, when you're sitting there, deciding whether you're "allowed" to make a play, or whether the one you're considering is a mistake, or you’re trying to recall what the solver would do in this exact spot – remember what the solver actually is. It's a reference. It’s a tool to grow your knowledge so that you can make the best decisions you know how to make.
Your job isn't to play in a way that gets a thumbs up when you report back to the solver later. Your job is to play the best you can in the moment.
Sometimes that means you go against what the solver would say.
Sometimes it means you act boldly in the face of uncertainty, when you have no idea what optimal strategy looks like. That you trust yourself to make the call (or fold!), and you live with whatever happens.
You’ll be wrong sometimes, and that’s okay. If it doesn’t feel okay when you try your best but get it wrong – that’s how you know you have a mindset issue to work on. Whether you recognize it or not, that fear of getting it wrong is causing far more losing plays than it’s preventing.
Play your hand. The solver isn't watching.