Hey there 👋,
There's a line from Tropic Thunder.
Ben Stiller's character is getting a lecture about method acting from Robert Downey Jr. The rule: you can go full character, you can commit completely, but you never go full retard.
The joke lands because the failure mode is real. Full commitment without calibration is how you end up somewhere you can't get back from.
I've done this twice now. Both times I thought it was conviction. Both times it was just concentration risk.
The first time was crypto. Last cycle I went basically all in on Solana early enough (sub $20) to do well. But the obsession was total. I wasn't managing it, I was consumed by it. Everything else became a distraction from the chart.
The second time was poker. I wrote about folding on that full time a few weeks back. The problem wasn't poker itself. The problem was the same: everything else got subordinated to it. I wasn't running a portfolio of effort, I was running a single position.
Both experiments taught me the same thing. When you go full, you lose perspective on whether it's actually working. You're too in it to see it clearly.
So here's where I've landed.
I'm back in crypto this cycle. SUI ecosystem specifically. I bought under $1 and it’s just had it’s first proper pump. I think this has real legs and I expect to make at least as much of a return as I did with SOL. That's a genuine conviction trade, not a gamble. But it's one third of my attention, not all of it.
Poker is still in the mix too. I haven't walked away from it, I've just right-sized it. The results have dramatically improved since I stopped treating it like a job. I identified two specific leaks: my preflop 3-bet frequency was too low and my turn continuation bet was too low. Both are aggression problems. Fixed them. The results are dramatic, I’m now terrorizing the pool. When I wasn't consumed by it, I could actually see what was broken.
That leaves the AI agency Why Not Bot where I’m now fully focussed on UK charities (rebuilt the site in Claude Design). I believe you really do need to niche down with this. The systems compound. Every skill I build gets sharper.
But selling Claude Cowork implementations is time-limited in the same way selling custom GPTs was. Right now there's a window where companies need someone to come in and build this for them. That window won't stay open forever. The tools will get easier, the competition will arrive, and the margin will compress.
Not going all in sounds obvious on paper. But the pull is real, especially if you've built things before. You got somewhere by going all in once, so you decided that's the strategy. It isn't. The commitment was fine. The concentration was the problem.
The other thing about one-third is it forces you to stay honest about each part. When poker is your whole identity, you'll rationalise the bad runs. When it's one leg of a three-legged stool, you have to actually look at whether it's working. Same with crypto. Same with the business. You can't hide a broken position inside a portfolio the way you can when it's the only thing you're doing.
That's the actual skill. Not picking the right thing to go full on. Finding what percentage of yourself to put where.
I'm still figuring mine out. But the one-third framework feels right for where I am right now.
SUI to $10. WhyNot to its next ten UK charity clients. Poker bankroll to taking shots at $500nl again. That's the scorecard.
If you want to follow how the AI agency side plays out in real time, including the skills I'm building and what I'm learning, that's what the Skool community is for.
Cheers, Richard (@richardpatey)