There's Nobody in the Forecasters Hall of Fame

Show Notes

This is the first recording of the year, and I opened it by telling Dan something I'd been dreading saying out loud: I've decided to wind down fig. The compliance and bookkeeping costs had each jumped about fifty percent, I wasn't growing the top line fast enough to cover them, and the business really needs a large pool of assets under management before it works as an ongoing thing. I'd given myself a signpost when I started — by forty, either pour my heart into it full time or move on — and at forty, after some honest introspection, I gave myself permission to follow through without shame. It wasn't a success, but it wasn't a failure either. I put risk on the table, I scratched the itch, and people who've been in the arena tend to respect the people still in it. I feel very little regret, and I'm happy about that.

From there Dan and I wandered into the strange economics of selling online, and it's where the episode got its name. I was marveling at how anyone runs a seasonal business — buying all your inventory in one quarter to sell it in another, betting on demand you can't actually see — and Dan dropped the line that titled the episode: there's nobody in the forecasters hall of fame. It's impossible to do accurately. Procter and Gamble can probably nail the toothpaste numbers, but who knew Peloton would catch fire and then cool off just as fast? We got into how Amazon takes roughly forty cents on every dollar just for operations through fulfillment, and once you account for the pay-to-play ads machine you're effectively handing over closer to sixty. I told Dan that more than half of all third-party sellers on the platform are now Chinese, and that the system seems to reward foreign sellers who skirt tariffs and income tax — money made off American consumers that never gets taxed here. My take, and I said plainly it was an opinion, is that we mostly don't need new regulation; we need to enforce the laws already on the books.

That tariff thread folded neatly into the book we've been reading, The Sovereign Individual. The authors argue that as the digital economy makes it easy to escape a country's tax burden, people will simply do it — and that the nation-state will hollow out and give way to something more like city-states and modern feudal lords. I told Dan I think the core idea is bunk: I don't believe governments will sit idly by, and I pointed at how quickly the country moved on TikTok as evidence we can exert control when we decide to. What frustrates me about the book is the feeling that two very smart guys keep grabbing historical details that don't quite fit, exploring the same idea over and over without giving me anything new. Dan put it better than I could — they're the kind of thinkers who, at a dinner table, would leave you impressed by their intelligence but wouldn't make you feel a single thing about yourself. We contrasted that with the old Disraeli story, where the gift wasn't seeming brilliant but making the person across from you feel brilliant.

We also spent real time on where AI is headed, and I think Dan laid out the history cleanly. Google researchers published a paper back in 2017 that kicked off the whole large-language-model revolution — the T in GPT stands for transformer — and the open AI tools we all use now are built on that research. He'd just read about a new Google paper, which he thought was called Titans, all about memory, and the gist is that the context-and-memory problem these models have today may get a lot smaller. We riffed on what a model would be like if it actually knew you — your email, your search history, your revealed preferences, the things people aren't always honest with themselves about. Dan's been deep in a Nat Eliason course on coding with AI, building little agents at night with Parks and Rec on in the background, and he described the loop everyone hits: you fix one thing, five others break, and the whole skill is learning to find the right spotlight and feed the right context.

Then Dan pulled out a couple of Carl Jung quotes that have been finding him lately, and the conversation went quiet and serious for a stretch — the greatest burden a child must bear is the unlived life of the parents, which we turned over from a few angles, and until you make the unconscious conscious it will rule your life and you will call it fate. We tied that back to the Bloomberg quote I'd sent over, which itself was about Henry Singleton: never let planning get in the way of doing, don't devise a five-year plan, central planning didn't work for Stalin or Mao and it won't work for an entrepreneur either. Singleton famously said he just liked to steer the boat each day. It all rhymes — so much planning is built on forecasting, and again, there's nobody in the forecasters hall of fame. You'll miss the plan because the world changed on you, and then you'll blame yourself for the world's lousy luck. So focus on the inputs you actually control, the number of bike rides instead of the number of pounds, and go with God on the rest.

Until you make the unconscious conscious, it will rule your life — and you will call it fate.

— Sean