When Winning the Product War Is No Longer Enough: Paul Graham on Branding, Wealth, and the Founder's Horizon
Paul Graham's two most recent essays — 'The Brand Age' (March 2026) and 'How to Convert Between Wealth and Income Tax' (May 2026) — read together reveal a single preoccupation: the gap between how a competition looks and what it actually costs. For early-stage founders, the brand age thesis reframes where to spend strategic energy as categories mature, and the wealth tax math trains a habit of converting any complex metric into the unit where real stakes become visible.
Paul Graham dropped two essays in the past weeks — one a grand historical arc spanning 50 years of Swiss watchmaking, the other a short technical proof about tax policy. Read together, they sketch a single underlying preoccupation: the gap between what people think they're competing on and what they're actually competing on.
Essay 1: "The Brand Age" — technology obsoletes performance, and then what?
Published: March 2026
콘텐츠 카드를 불러오는 중…
Paul Graham opens with the Swiss watch industry's near-death experience in the 1970s: Japanese competition, a collapsing exchange rate advantage, and then quartz movements that made accurate timekeeping a commodity. Between the early 1970s and early 1980s, unit sales of Swiss watches fell by almost two thirds.1
What followed is the part founders should sit with. A handful of makers — Patek Philippe, Audemars Piguet, Rolex — survived not by making better watches, but by redefining what a watch was for. They stopped competing on the thing that could be beaten (accuracy, thinness) and started selling something that couldn't be automated away: the idea that you're wearing a meaningful object.
The mechanism Graham traces in careful detail: Patek commissioned a new iconic case design in 1968, physically expanding how much surface area the brand occupied on the wrist from 8 square millimeters to 800. Then they launched brand advertising — explicitly talking about price as a desirable attribute. One 1968 ad explained why buyers were "well advised to invest perhaps half a month's income." Later, Audemars Piguet introduced artificial scarcity. If you want to buy a Patek Nautilus today, you can't just walk in and pay. You spend years purchasing lesser models to prove loyalty, then wait on a list.1
Graham's structural observation is worth quoting:
"Branding is centrifugal; design is centripetal."
Good design converges on the right answer. Branding must diverge — it needs to be different. Those two imperatives are fundamentally opposed, and there is no permanent reconciliation between them.
The startup-facing conclusion is the sharpest thing in the essay. Graham writes:
"What does this imply for startups? It's the opposite of what most founders want to hear, but the trend is clear. When products get good enough, the action moves to branding. If you're a founder, you can see what's coming decades in advance if you know what to look for."
He's not predicting that every category is at that inflection point right now. He's pointing out that Moore's Law is moving up the stack: technical differentiation disappears in category after category, and in consumer tech it has already largely happened. When you couldn't find a serious technical fault with any phone at any price point — the only differences were brand and size — the category had entered the brand age.1
What this means if you're building now:
Two windows still exist where performance beats branding, and Graham names them explicitly:
- Unexplored territory — if you arrive first in a new category, your right answer is your brand advantage, at least until others converge on it.
- Enormous possibility space — fields like fine art or certain research-heavy sectors, where the solution space is so large that "best" and "distinctive" can coexist for a long time.
The tactical implication: if you're in a category where the top five products are meaningfully similar on specs, and you're still thinking "we just need to ship a better feature" — you may be solving the wrong problem. The brand age doesn't mean quality stops mattering; it becomes a threshold (good enough to not embarrass the brand) rather than a differentiator. Once you're above threshold, resources spent on engineering returns less than resources spent on positioning, narrative, and community.
For early-stage AI founders specifically, the current moment resembles the Swiss watchmakers in 1968-1972: you can still see clear performance gaps between foundation models and products, so technical differentiation still earns customers. But Graham's point is that this window is already closing in several verticals. The companies building strong brand identities now — even while product performance still differentiates — are buying insurance against the day it doesn't.
Essay 2: "How to Convert Between Wealth and Income Tax" — the hidden math politicians don't know
Published: May 2026
콘텐츠 카드를 불러오는 중…
This one is short — roughly a page — and it's almost entirely arithmetic. Graham's central claim: a 1% wealth tax is mathematically equivalent to a 20% income tax.
The conversion formula is simple. A wealth tax hits your capital stock each year, regardless of what it earns. An income tax hits returns. To convert between them, you divide by the assumed risk-free rate of return. At 5%:2
Wealth tax rate ÷ risk-free return rate = equivalent income tax rate 1% ÷ 5% = 20%
Graham walks through a concrete example: $100 at 5% return earns $5. Under a 20% income tax, you pay $1 and keep $104. Under a 1% wealth tax, you pay $1 on the $100 stock and keep $104. The math is identical.2
His observation about politicians is precise: no politician would casually propose "adding a mere 20% to the income tax rate" — everyone recognizes that as a massive change. But proposals for a "mere 1% wealth tax" get treated as small-scale tweaks, even though they're the same thing numerically. In the median US state, adding a 1% wealth tax would push the combined marginal rate (federal + state + the new wealth tax equivalent) to roughly 61.75%, higher than Denmark's top rate of 60.5%.2
Why this belongs in a founder's reading list:
Most early-stage founders won't encounter wealth taxes directly until they've grown significantly. But this essay trains a mode of thinking that applies everywhere:
- Restate the question in the form where the answer is legible. Politicians talk about wealth tax as if it's a novel instrument because they're comparing it to income tax at face value. Once you convert to the same unit — what fraction of annual returns does this take? — the comparison becomes obvious. The same cognitive move applies to evaluating equity grants, term sheet economics, SaaS contract structures, or co-founder splits. The number that sounds small in one representation may look very different in another.
- Threshold effects and rates compound. A 1% wealth tax doesn't feel like much. 20% income tax equivalent on returns is a material behavioral change for capital allocation. Founders who've modeled cohort retention, CAC payback, or subscription churn understand this intuitively — a seemingly small per-period drain has compounding consequences over time.
- Understanding the argument your investor or acquirer is actually making. When a VC talks about "IRR targets" or when a strategic acquirer frames an earnout, they're running exactly this kind of rate-conversion arithmetic in the background. Knowing the math makes those conversations symmetric.
Graham's brief note on why 5% is an appropriate conversion denominator (it's the risk-free rate, because the tax is risk-free for the government even if returns are uncertain for the investor) is itself a useful mental model: when evaluating any recurring obligation, always price it against the risk-free benchmark, not the upside case.
The common thread
Two essays, one month apart, on very different subjects. What connects them is a preoccupation with the gap between how something looks and what it costs.
The Brand Age watches look like instruments. They're luxury goods. Politicians pitch wealth taxes as modest interventions. They're the equivalent of the world's highest income taxes in disguise. In both cases, the error comes from comparing the nominal description to the wrong reference class.
For founders building in 2026, the meta-lesson is: pick the right reference class before making strategic decisions. The product you're building may look like a technical differentiation game in the frame you're used to. In the frame five years out, it may already be a branding game. Your fundraising terms may look like standard market rates in one formulation. Converted to the unit that actually matters for you, they may look very different.
Reading Graham is useful less for specific prescriptions than for this habit of mind: always ask whether you're comparing things in the form that makes the real stakes legible.
Next edition: Sunday, June 8, covering founder posts from the week of June 1–7.
이 콘텐츠를 둘러싼 관점이나 맥락을 계속 보강해 보세요.