Verification Theatre: When Facts Stop Mattering

Jason Arday stepped down from his chair at the University of Cambridge earlier this week, following weeks of reporting by The Telegraph and The Times, and just hours after the university opened an investigation into his academic qualifications and honorary appointments.

Questions have been raised about the qualifications he held and where he earned them, about books he said he had written, about places he said he had been, and about television programmes said to have featured him. The claims that built his academic career are being checked, and a striking number of them do not stand up to scrutiny.

For fairness, Arday denies wrongdoing and attributes errors in early work to inadequate supervision and his disabilities, and I have no interest in kicking a clearly troubled man who has just lost so much. The interesting question is not about him at all. It is this: how did the most famous verification institution in the world fail to verify?

That is what a university is. Strip away the gowns and gothic spires and Cambridge's core product - for hundreds of years -, has been certification: this person's work has been examined, and it is real. That is what the degree says. It is the entire economic function of the place. And yet a professorship was awarded, honorary doctorates followed from five other universities, and the checking, when it finally happened, was done by a Substack and two newspapers. The institution built to examine work took three years from first written warning to first investigation, and then the man resigned before it concluded.

The failure has a precise shape, and it is everywhere. Institutions verify by credential, and a credential is just another institution's earlier decision. Durham appointed a sociologist because the title said sociology. Glasgow promoted what Durham had appointed. Cambridge crowned what Glasgow had promoted. Five universities then draped honorary doctorates over what Cambridge had crowned. Each link in the chain trusted the link before it, which means nobody ever checked the thing itself. The system was not corrupt. It was worse than corrupt. It was credulous by design, because checking is expensive and titles are cheap to read. And once the first label was attached, every subsequent committee was, quite rationally, reading the label. You get what you pay for, and a system that pays for labels will get labels.

Here is what almost nobody knows: a system now exists that cannot fail this way, and it is called Bittensor.

Bittensor is an open network, launched in the Bitcoin mould, that its co-founder Jacob Steeves describes as an incentive computer. It hosts scores of competing markets called subnets, each producing a measurable digital commodity: AI inference, model training, weather forecasting, protein folding. Miners, anyone, anywhere, submit work. Validators, staked participants with their own money at risk, continuously score that work against everyone else's. The chain then does something quietly brutal: it compares every validator's scores against the stake-weighted consensus of all the others, and a validator whose scoring deviates from honest consensus earns less. Honest evaluation is not a policy. It is the profit-maximising strategy. Rewards flow to exactly what consensus has verified, continuously.

Now run the Arday case through that machine. A participant on Bittensor is a wallet address and a stream of output. There is no CV field, nowhere to enter a chair, a title, a backstory or a visiting professorship, so there is nothing to fabricate. Work that matches someone else's earlier work adds nothing over the existing supply and earns accordingly. Work that stops arriving stops being paid the moment it stops, not at the end of a confidential internal process. The verification that took Cambridge three years, prompted from outside, happens on Bittensor before your coffee cools, and then happens again, and again, permanently. Tenure, viewed from this angle, is a mechanism for never having to be verified twice.

On Bittensor there are no background checks, no quotas, and no equality policies. There is only the value added, measured against everyone else's and paid in proportion. I am not telling you that is right, and I am not telling you it is wrong; reasonable people will disagree about what a society owes to fairness beyond the measurable, and that argument is above my pay grade. I have made the narrower case elsewhere, and it is this: companies that deploy their assets, their capital and their people on the basis of value added are very likely to outperform companies that deploy them on any other criteria, whatever they are and however noble the intention. That is not an ideology. It is arithmetic, and the market runs the calculation whether anyone approves of it or not. The same logic scales past the firm to how capital gets allocated across an entire economy.

One precision matters, though. Bittensor is not ungameable, and its builders would be the first to say so; people will always probe an incentive, and that the designer's job is to patch the mechanism in public, which subnet builders do constantly. I have been sharply critical of how that patching is done, and I stand by every word of it. But the honest comparison is not perfection against failure. It is response time. When a Bittensor scoring mechanism gets exploited, miners find the seam in days and the fix ships shortly afterwards. When Cambridge's scoring mechanism got exploited, the seam ran for years and the fix required investigative journalism. Both systems get gamed. Only one of them notices.

And the network is young. Its real revenue is early against the emissions that subsidise it, and the people building it think in decades, not quarters. Their deeper argument is not efficiency but openness: that the right to contribute to machine intelligence, and own a piece of it, should not sit behind a corporate login or an admissions office. A teenager in Jakarta competes on identical terms with a lab in San Francisco, and neither can lean on reputation, because the mechanism has no idea who either of them is. As Steeves puts it: "No one cares about where you went to school. If you can just solve that problem, you get paid out."

There is a lovely irony here, and it belongs to Cambridge itself. In January 1913, a clerk at the Madras Port Trust with no degree, twice failed out of college, posted a letter full of theorems to a Trinity professor he had never met. The professor ignored the envelope and read the mathematics, and Srinivasa Ramanujan became one of the great mathematicians of the century. Cambridge's finest hour of verification came when one man judged the work and ignored the story entirely. Its worst came a century later, when the machinery judged the story and nobody read the work.

I invest in Bittensor subnets precisely because of what this mechanism does. I have written a longer three-part series on incentives and capital allocation, Who Gets Paid, and I think out loud on X and LinkedIn.

The whole argument compresses to one line: every system pays for something. Most pay for credentials, connections or stories. Bittensor pays for value added, and nothing else.

It has never once asked anyone where they went to school.

Author Bio

Mark Creaser Investor and CEO of DSV Fund, the world's first liquid hedge fund dedicated exclusively to Bittensor, and Chairman of Astrid Intelligence PLC (AQSE: ASTR). He writes on incentives and capital allocation at markcreaser.com, and can be found on X, LinkedIn, Crunchbase and Wikidata.

Disclaimer: This article is provided for informational purposes only. It is not offered or intended to be used as legal, tax, investment, financial, or other advice.