Impactism
日本語 Join Impactism
Manifesto The Impact Standard Foundations Objections Read For organizations 日本語
Join Impactism
A row of yellow rating stars on a colorful surface
Photo by Towfiqu barbhuiya on Unsplash

Essay

Whuffie Is a Warning

June 23, 202611 min read

We are building a movement that wants impact, not money, to be the measure of a life. Someone already wrote that idea as fiction, the world found it beautiful for a decade, and then the author himself called it a dystopia. The most famous attempt at the thing we are reaching toward did not just fail in practice. It failed in a way its own creator could name. We would be fools not to study the autopsy.

The fiction is Whuffie, the reputation currency at the center of Cory Doctorow’s 2003 novel Down and Out in the Magic Kingdom. The novel imagines a post-scarcity, post-death future — the “Bitchun Society” — where energy is free, work is optional, and money has become pointless because nothing is scarce. Into that void steps Whuffie: a running tally of how much respect other people hold for you, replacing money as the thing everyone chases (Doctorow 2003). Do something admirable and your Whuffie rises; fall out of favor and it drops. It reads, at first, exactly like our pitch. A world that honors contribution, not just capital.

A careless movement would claim Whuffie as a forerunner and move on. We do the opposite: we take it as the single most important warning we have, because the person best placed to judge it — the man who invented it — looked back after thirteen years and recanted.

The author’s own autopsy

In a March 2016 column for Locus, Doctorow opened with a sentence we have to quote in full, because it is the whole problem in miniature: “I need to confess something: ‘Whuffie’ would make a terrible currency” (Doctorow 2016). He did not soften it. The column’s title is “Wealth Inequality Is Even Worse in Reputation Economies,” and its argument is that a single, legible, convertible reputation score does not escape the pathologies of money. It imports them and adds new ones. “Whuffie has all the problems of money,” he wrote, “and then a bunch more that are unique to it.”

Take the failure he names most sharply. A reputation score sold as meritocratic, he argued, ends up pooling “around sociopathic jerks who know how to flatter, cajole, or terrorize their way to the top.” The mechanism is grimly familiar: once you have a lot of standing, people compete to hand you the chances that generate still more standing, and you are positioned to take credit for what goes well and assign blame for what doesn’t. That, Doctorow noted, is exactly how money already works. Making the unit “reputation” instead of “dollars” did not break the feedback loop by which advantage compounds. It just renamed the currency.

And there is a second wound, deeper, in how Whuffie is computed. Doctorow describes it as “a score that a never-explained set of network services calculate by directly polling the minds of the people who know about you and your works, reducing their private views to a number.” Sit with that phrase — reducing their private views to a number. The richness of what people actually think about your contribution, the texture of respect, gratitude, disagreement, all of it flattened into one scalar that can be ranked, compared, and spent. He came to see this not as a utopia he had drawn but as a dystopia he had failed to recognize while drawing it.

We did not have to go hunting for our critics. The strongest one signed the original.

Steelmanning Whuffie before we bury it

Honesty cuts both ways. Before we draw lessons from the failure, we should state the case for Whuffie at full strength, because parts of it are genuinely right.

Doctorow’s instinct was sound: in a world where survival is handled, the thing worth organizing a society around is not accumulation but contribution and regard. Whuffie correctly identifies that something inherits the throne money vacates — that abundance does not end the game, it changes what the game measures. We agree with that completely; it is the premise of our whole project. Whuffie also correctly intuits that this successor should be social — conferred by others, not seized — rather than a thing you mine alone. A person broke but respected should not starve; a person rich but despised should not buy peace. As a moral aspiration, Whuffie is almost exactly ours.

So the failure is not in the dream. The dream is shared. The failure is in four specific design choices — and that distinction is the entire essay. Whuffie shows that you can want the right destination and still build a machine that delivers you to the wrong one. The lesson is not “abandon the goal.” It is “the goal does not forgive bad mechanism design, and here is precisely where the mechanism breaks.”

The descendants, and what corrupted each

The strongest reason to trust Doctorow’s recantation is that we do not have to take it on faith. We have run the experiment, repeatedly, in the real world — every time we built a number to capture human worth and let it buy real things. Each descendant failed, and each failed through one of the same four faults: convertibility, central control, a single legible number, gameability.

Klout — convertibility. From 2009, Klout reduced your entire social presence to one number from 1 to 100, meant to score your influence. It mattered for a while; at its peak it scored hundreds of millions of profiles, and Lithium Technologies bought it in 2014 for a reported $200 million (TechCrunch 2014). The corrupting move was making the score spend. Through “Klout Perks,” high scorers were handed free products, hotel-room upgrades, and airline-lounge access — the score converted into material advantage (Econsultancy). The instant a reputation number buys things, it stops measuring esteem and starts measuring whatever games the number, and Klout was eminently gameable — an opaque, unaudited algorithm that rewarded volume and could be inflated with bots and engagement pods. Lithium shut it down on May 25, 2018 (Search Engine Journal 2018). A legible, convertible, gameable score is precisely Whuffie, and it died the way Doctorow predicted.

Sesame Credit — central control. Ant Group’s Zhima (“Sesame”) Credit is an opt-in commercial score inside Alipay, rating users from 350 to 950, and it is frequently — wrongly — conflated with the Chinese state’s Social Credit System; it is in fact a private loyalty program, and the People’s Bank of China declined to extend the licenses of the eight private pilots it grew out of (Zhima Credit, Wikipedia). We cite it carefully and only for what it actually is: a commercial reputation score that converts into real life. A high score waives deposits, speeds you through processes, and — through a partnership with the dating site Baihe — raises your visibility to potential partners. The fault here is not, in the commercial case, the dystopia of the popular imagination; it is subtler and more instructive. A single operator defines what counts, and so the citizen optimizes for the operator’s definition rather than for the thing the definition was a proxy for. Central control turns a measure of trustworthiness into an instruction set.

Uber and Airbnb — the single legible number, inflating. Two-sided ratings were supposed to make strangers trustworthy to one another, and at first they did. Then they rotted, in a way that is now precisely documented. Across more than a decade of data from an online marketplace, Apostolos Filippas, John Horton, and Joseph Golden showed that ratings inflate over time — not because service improved, but because leaving an honest low rating is socially costly when you are face-to-face with the person it will hurt. Raters feel pressure to give “above average” marks, which drags the average up until the scale is useless: on many platforms a 4.6 out of 5 is a failing grade. Their conclusion is the line we keep returning to — reputation systems “as currently designed, sow the seeds of their own irrelevance” (Filippas, Horton & Golden 2019). A single number everyone can read becomes a number everyone is too polite to lower, and an unfalsifiable score measures nothing.

Reddit, Hacker News, and Stack Overflow — gameability. Karma and reputation points are the most ordinary reputation currencies alive, and the most openly gamed. Joel Spolsky, who co-founded Stack Overflow, is candid that its points and badges were “a dusting of gamification, most of it around reputation” — a designed incentive meant to steer behavior toward the actions the site wanted (Spolsky 2018). The trouble is that any designed incentive built on a number becomes a target. A 2021 study of Stack Overflow catalogued four distinct reputation-fraud patterns — including voting rings, where users coordinate to upvote each other — and built tools to catch them; the platform reduced the scores of most flagged accounts (Mazloomzadeh et al. 2021). This is Goodhart’s law wearing a karma badge: the moment the number is worth having, optimizing the number displaces doing the work the number was supposed to reward.

Four systems, four faults — and notice they are not four different lessons. They are four faces of one lesson. Convertibility, central control, a single legible number, and gameability are the load-bearing pillars of Whuffie’s design, and each real descendant collapsed when one of them gave way.

The turn: every failure is a rule

Here is where the warning becomes useful, and where we part from Whuffie deliberately, choice by choice. We did not arrive at our design rules and then go looking for validation. We read the wreckage first, and let it dictate the rules. Each property we adopt is the photographic negative of a failure above.

Whuffie was convertible — it bought things, like Klout’s perks. So the unit is non-convertible: conferred, carried, never cashed. It is not spendable, not redeemable, not a balance you draw down for advantage. The instant standing buys a discount, it stops being honor and becomes a price, and a price corrodes the very motive it was meant to celebrate — the case we made at length in Honor Is Not a Price.

Whuffie was centrally computed by “a never-explained set of network services,” and Sesame Credit is defined by a single operator. So recognition must be peer-conferred, not issued by a central authority that decides what counts. No headquarters holds the formula. The texture stays in the conferring, not in a definition someone optimizes against.

Whuffie was a single legible number, and so were Klout and the inflated star ratings — one scalar, rankable and spendable. So the unit is plural: many measures, many communities, never one global score. A single number invites exactly the comparison, hoarding, and inflation that killed every example here. We do not build a leaderboard for human worth, because a leaderboard is money’s deepest pathology, not its cure.

Whuffie and every karma system were gameable and permanent — bank the number, then defend and inflate the balance forever. So recognition must decay: it cannot be hoarded, because what cannot be banked cannot be gamed into a fortress of compounding advantage. Standing that fades keeps asking, what are you contributing now? — not what did you once accumulate?

Non-convertible. Peer-conferred. Plural. Decaying. Read down that list and then back up at the four failures, and you will see they are the same list, inverted. Whuffie failed for reasons we can name in advance, and we have tried to design against every one of them by name.

The honesty we owe you

We will not end on triumph, because we have not earned it. Locating the failures is not the same as proving the cure. We can say with confidence that Whuffie’s faults were real, that its author disowned them, and that every real-world descendant repeated at least one of them. We cannot say we have built something that escapes them all. No one has run our design — a non-convertible, peer-conferred, plural, decaying recognition system, operating at the scale of a culture, with the firewalls built in from the start. That experiment is unrun. We are proposing to run it, carefully, and to report honestly when a piece of it breaks.

There is one objection we have to hold open rather than wave away. Doctorow’s deepest charge was that reputation economies concentrate advantage around those who flatter and maneuver — and a fair reader will ask whether any system of regard can fully escape that, ours included. Peer-conferral can become a popularity contest; decay can be outpaced by someone relentless; plurality can fragment into cliques that each crown their own insiders. We do not have a proof that our four rules defeat the concentration Doctorow described. We have a hypothesis that removing convertibility and the single number removes its sharpest teeth, and an obligation to watch for it returning in a new disguise. That watch is part of the design, not an afterthought to it.

What we are sure of is narrower and firmer. The dream Whuffie chased is worth chasing: a world that honors contribution over accumulation. The machine Whuffie built to chase it was wrong in four specific, nameable ways. And the difference between a dystopia and a movement may come down to exactly this — whether you let the number be spent, let one authority define it, let it collapse into a single score, and let it be banked forever. We are betting it does not have to. The most famous reputation currency ever imagined is not our model. It is our warning, and we have written its lessons into our rules.

Make impact, not money, the measure of a life — conferred, carried, never cashed. Whuffie is the cautionary tale that taught us the difference.

Sources

← All essays