Two how-to books from the early years promised to walk beginners through digital money, step by step. The steps changed completely. The two questions underneath them — what’s this worth, and who’s holding yours — never moved an inch.
The whole bitcoin question, separated into what changed and what didn’t.
Left is the pitch as it stands in 2026. Right is the part a planner has to say out loud. Owning it’s now easy. Sizing it’s still the entire game.
The Machinery Was Real. The Steps Went Obsolete. The Questions Kept.
Give the early how-to books their due: the thing they were explaining was, and remains, genuinely novel. Bitcoin is a public ledger maintained by thousands of independent computers, none of which is in charge, secured by a global network of specialized machines paid in new coins for settling transactions — with the total supply capped, by the protocol itself, at 21 million. No board can dilute it, no government can print more of it, and no bank sits in the middle of a transfer. Whatever one thinks of the price, the engineering solved a problem computer science had considered unsolvable: strangers agreeing on who owns what, with no referee.
The step-by-step chapters, though, aged like milk — and instructively. The 2012 reader generated cryptographic keys, backed up wallet files, and wired money to unregulated exchanges run by hobbyists. Nearly every specific step has since been replaced, most decisively in January 2024, when U.S. regulators approved spot bitcoin ETFs — putting the asset inside an ordinary, regulated brokerage account alongside the index funds. The on-ramp that once required an evening of homework now requires a ticker symbol.
What survived both books untouched are the two questions the rest of this letter is about. First: what’s a claim on nothing but scarcity actually worth? Bitcoin produces no earnings, pays no interest, and rents to no tenant; there’s no cash flow against which to compute a fair price, which puts it in the same valuation family as gold — worth what the next person will pay, forever. Second: who’s actually holding yours? That question sounds procedural. Section two is about the year it became the whole story.
Both extremes in this debate argue against a version that stopped existing. The skeptics’ “funny internet money” now clears through the same custodians as your index funds; the believers’ “replacement for the dollar” has spent fifteen years behaving like a volatile speculative asset instead. What actually exists in 2026 is something more ordinary than either camp wants: a scarce, liquid, cash-flowless asset that some portfolios hold a sliver of, the way portfolios have long held gold.
And one distinction does the work of a whole bookshelf: novel engineering isn’t the same thing as a claim on future cash. The first is real and impressive. Only the second gives an asset a floor.
Down 80%, Four Separate Times — And That Is The Base Case, Not The Disaster
Any honest sizing conversation starts with the asset’s actual ride, because nothing else in a household portfolio behaves remotely like it. Bitcoin has collapsed roughly 80% or more on four separate occasions — 2011, 2013–15, 2017–18 (a peak-to-trough of about 83%), and 2021–22 (about 77%). Each time, obituaries were published in serious outlets. Each time, it eventually made new highs. Both halves of that sentence are true, and both halves must be priced in advance: the survival record is why reasonable people hold a sliver; the drawdown record is why the sliver must be sized before the fact, not rationalized after.
The volatility also breaks the tools people bring to it. Dollar-cost averaging, rebalancing bands, “buy the dip” — all of it works mechanically, but the emotional load is calibrated to stock-market drawdowns of 30–50%, not 80%. In practice the failure mode is always the same one, and readers of this month’s letters will recognize it: the position was sized for the ride up, and the household became a forced seller on the ride down — sometimes by margin, more often by simple panic at a number that was always in the historical record, staring at them the whole time.
And a hard-earned line for the enthusiasts: never on leverage. An asset that routinely moves double digits in a day, borrowed against or bought on margin, isn’t a conviction trade; it’s an appointment with the liquidation desk. Every crypto cycle’s worst stories — the 2022 lender failures above all — were leverage stories wearing bitcoin costumes.
We size speculations by their history, not their thesis. The test is mechanical: take the worst drawdown the asset has actually delivered, apply it to the proposed position tomorrow morning, and ask whether any plan number moves. If the answer is yes, the position is too big — whatever the thesis says.
The behavioral clause matters as much as the math: a correctly sized position is one you can watch fall 80% without doing anything — because the plan never depended on it. Sized that way, bitcoin becomes what it honestly is at a household scale: a small, interesting, violently volatile diversifier. Sized any other way, it becomes the plan’s single point of failure.
850,000 Coins Vanished With One Exchange. The Asset Was Fine. The Customers Were Not.
In early 2014, the largest bitcoin exchange on the planet — Mt. Gox, a Tokyo operation that had begun life as a trading-card marketplace — halted withdrawals and went dark. Roughly 850,000 bitcoins belonging to the exchange and its customers were gone, lost to some still-disputed blend of theft and mismanagement. The detail that makes it the perfect teaching case: bitcoin itself worked flawlessly through the whole affair. The ledger never missed a block. What failed was the middleman — and the customers’ claims on it — which is precisely the distinction the 2012 books were straining to teach with their wallet-file homework: the asset and the IOU for the asset are different things with different failure modes.
The intervening decade offered a remedial course for anyone who skipped it — culminating in FTX in 2022, where the confusion between customer property and house money produced one of the largest frauds in financial history. Between the two bookends, the industry’s own slogan — not your keys, not your coins — hardened from ideology into actuarial observation.
Which is why the 2024 ETF approval, easy to dismiss as mere packaging, was the decade’s most consequential event for households. It replaced the weakest link — the unregulated middleman — with the same regulated custody, segregation rules, auditing and insurance infrastructure that already protects every index fund you own. Purists object, correctly, that an ETF holder owns a claim rather than keys; self-custody remains the sovereign option for those willing to be their own bank, with the responsibilities that phrase implies (keys can be lost, stolen, or inherited by no one — an estate-planning wrinkle the obituary files are full of). For an ordinary household, the honest ranking is now straightforward: a regulated ETF first, self-custody for the technically serious, and an unregulated offshore exchange never.
One last unglamorous item the how-to books barely mention: taxes. The IRS treats bitcoin as property — every sale, swap, or purchase made with it’s a taxable event with a gain or loss to compute. The ETF wrapper simplifies this too, reducing a ledger-keeping hobby to an ordinary 1099. Unsexy; decisive.
The custody story is the whole fifteen years in one lesson: the failures that actually cost people money were never failures of the novel part. The cryptography held. What broke, repeatedly, was the oldest thing in finance — other people holding your property badly. The remedy is equally old: regulated custody, boring wrappers, audited middlemen.
So the house position, in full: if a client wants bitcoin, we’d rather it be small, in a regulated wrapper, inside the plan where we can see it — than large, on an app, outside it. The fifteen-minute review covers all three adjectives: the size against the −83% test, the wrapper against the custody lesson, and the tax treatment against the property rules.
Holding coins on an app? Curious but never bought? Same meeting.
Fifteen minutes: the position sized against its worst historical year, the custody checked against the 850,000-coin lesson, and the tax treatment nobody mentions at the barbecue.