The Accountability Gap: Institutions, Incentives, and Consequences

There are certain conversations that begin with a single story and gradually reveal something much larger underneath it. Theranos is one of those stories. What initially appears to be a tale of technological deception opens into questions about institutions, incentives, accountability, evidence, and the ways ordinary people can be persuaded to trust things they have neither the knowledge nor the means to independently verify.

.   .   .

In this discourse we explored… the Theranos scandal and what it reveals about the relationship between technological promises, institutional prestige, corporate power, legal accountability, and public ignorance. The discussion moved from Holmes and Balwani to the broader mechanisms that can distribute responsibility, absorb misconduct as a cost of doing business, and make genuine verification difficult. It also examined the danger of replacing institutional trust with reflexive distrust, arriving at a more fundamental question: how do we become harder to deceive without simply becoming cynical?

Key Points Discussed

• Theranos presented an ambitious technological claim that was not supported by the functioning technology behind it.
• Reputation, celebrity, wealth, prestigious associations, and institutional partnerships can create an appearance of credibility without actually validating a claim.
• Secrecy and compartmentalization can make independent scrutiny difficult, particularly when the public lacks the knowledge to know what questions to ask.
• The legal system distinguishes between serious institutional harm and the ability to prove criminal intent by a particular individual.
• Corporations can distribute responsibility across executives, subsidiaries, lawyers, consultants, insurers, compliance departments, and other institutional actors.
• Regulatory settlements and financial penalties operate differently from individual criminal punishment and can become part of the economic risk of operating a large organization.
• Institutional misconduct and criminal culpability are related but not interchangeable questions.
• Technological enthusiasm can create a recurring pattern of “revolutionary” claims, obsolete traditional methods, proprietary information, delayed verification, and appeals to authority.
• The same skepticism used against institutions can itself become unreliable when anti-institutional narratives are accepted without evidence.
• Claims about suppressed inventions, destroyed technologies, or persecuted innovators need to be examined individually rather than accepted merely because they fit a larger narrative.
• Human-centered innovation and institutional control of innovation are not necessarily the same thing.
• The social consequences of technology depend heavily on the institutions, incentives, and economic structures surrounding it.
• The most useful response to extraordinary claims is not automatic belief or disbelief, but asking what evidence would establish the claim.
• Independent reproduction, falsifiability, control of underlying data, financial incentives, failure modes, and who bears the consequences are all important questions.
• Ultimately, the practical lesson is to become more difficult to exploit without surrendering the ability to distinguish evidence from narrative.

Key Takeaways

• Prestige is not evidence.
• A compelling demonstration is not the same thing as independent validation.
• Institutional authority can fail, but opposition to authority is not evidence in itself.
• Ask who benefits, who bears the cost, and what would prove the claim wrong.
• Learn enough about a system to recognize the questions that actually matter.
• Don’t outsource your epistemology to prestige.

» Trance:

Tell me about Theranos.

» ChatGPT:

Theranos is one of the most famous corporate fraud stories of the 21st century because it combined a genuinely intriguing technological idea, enormous amounts of money, powerful investors and political connections, secrecy, aggressive PR, and ultimately a technology that did not work as advertised.

The short version is: Theranos claimed it had revolutionized blood testing. Instead of drawing several vials of blood from a vein and sending them through conventional laboratory machines, its technology supposedly could perform a huge range of tests from just a few drops of blood taken from a finger. The company said this would make blood testing cheaper, faster, less painful, and much more accessible.

It didn’t work.

Elizabeth Holmes founded Theranos in 2003 when she was 19. She dropped out of Stanford and built the company around the idea that a small amount of blood could be analyzed for numerous medical conditions. The central technology eventually became known by names including the “Edison,” “TSPU,” and “miniLab.” Theranos claimed that its proprietary machines could perform a broad range of tests accurately and quickly. SEC

And this is where the story gets particularly interesting.

Holmes wasn’t simply a young entrepreneur making exaggerated startup claims. She built an extraordinary aura of credibility around the company. Theranos attracted investors including Rupert Murdoch, the Walton family, Henry Kissinger, George Shultz, Betsy DeVos and others. By 2014, Holmes’s stake in the company was valued at billions of dollars on paper, and she became a major Silicon Valley celebrity.

The company also brought in Ramesh “Sunny” Balwani, a much older technology executive who became president and COO. He was eventually convicted separately of fraud involving both investors and patients. Department of Justice

The central deception was remarkably simple once you understand it.

Theranos publicly portrayed its own proprietary machines as doing the work. But according to the SEC and evidence later presented at trial, the machines could perform only a limited number of tests. For much of the testing actually performed on patients, Theranos relied on conventional commercial laboratory machines made by other companies.

In other words, the technology that was supposed to make traditional blood-testing equipment obsolete wasn’t actually doing most of the testing.

The SEC said in 2018 that Theranos had raised more than $700 million from investors while making false or misleading claims about its technology, business and finances. The SEC specifically alleged that the company’s proprietary analyzer could perform only a small number of tests while the vast majority of patient tests were performed using modified conventional analyzers. SEC

There was another particularly striking element: secrecy.

Theranos aggressively protected its technology from scrutiny. Employees were compartmentalized. People working on one aspect of the system often weren’t allowed to know what other teams were doing. Former employees who raised concerns could face aggressive legal threats.

That mattered because independent verification is absolutely critical in medical technology. If somebody claims they’ve invented a machine that can replace conventional laboratory testing, you don’t just take their word for it. You test it against established machines, under controlled conditions, and publish or otherwise independently validate the results.

Theranos did not provide the kind of transparency that would have allowed outsiders to establish that its claims were true.

Then came Walgreens.

In 2013, Theranos began putting its “Wellness Centers” into Walgreens stores, initially in Arizona and California. Ordinary people could walk into a Walgreens and have blood tests performed through Theranos.

This transformed the problem from a questionable startup story into a potential public-health problem.

According to the federal indictment, Theranos knew that its system wasn’t consistently capable of producing accurate and reliable results for a number of tests. Patients nevertheless received results from Theranos testing. The government specifically identified tests involving things such as potassium, calcium, sodium, HIV, pregnancy-related testing and HbA1c among those affected by reliability problems. Department of Justice

And inaccurate blood tests aren’t like a bad smartphone app.

Imagine a patient receives a falsely elevated result and believes they’re seriously ill. Or receives a falsely reassuring result and believes they’re healthy when they aren’t. A test can influence medication, further diagnostic testing, pregnancy decisions, cancer treatment, or whether someone seeks medical care.

The story finally began to unravel publicly in 2015.

Wall Street Journal reporter John Carreyrou investigated Theranos after receiving information from people inside the company. His reporting challenged the company’s claims and revealed that Theranos was relying heavily on conventional machines rather than its supposedly revolutionary technology.

That reporting triggered a cascade.

Regulators investigated. Walgreens eventually terminated its relationship with Theranos. The company’s laboratory operations collapsed. The SEC brought charges in 2018, and federal prosecutors charged Holmes and Balwani with wire fraud and conspiracy.

There is an important distinction between the government’s allegations and what was ultimately established at trial.

Holmes was not convicted of every charge against her. In January 2022, a federal jury convicted her of one count of conspiracy to commit wire fraud and three counts of wire fraud relating to investors. She was acquitted of the patient-related fraud counts, while the jury failed to reach unanimous verdicts on three other investor-fraud counts. Department of Justice

Balwani’s outcome was different. In 2022, another jury convicted him on all counts involving conspiracy and wire fraud against investors and patients. Department of Justice

Holmes was sentenced to 135 months—11 years and three months—in federal prison. Department of Justice

Theranos itself eventually ceased to exist.

What’s fascinating about the case isn’t merely “woman lies about startup and goes to prison.” There are several deeper lessons embedded in it.

One is the danger of substituting reputation for verification.

Holmes surrounded herself with impressive people. Theranos had famous board members, wealthy investors, military associations, partnerships with Walgreens, pharmaceutical-company references and enormous media attention. All of those things created an impression that the underlying technology must have been legitimate.

But none of them actually demonstrated that the machine worked.

Another is the problem of secrecy in science.

Secrecy can be legitimate when you’re protecting intellectual property. But medical claims ultimately have to survive independent testing. A revolutionary medical technology cannot be validated by the charisma of its founder or the quality of its investors.

And there’s an especially interesting lesson about venture capital.

Theranos reportedly reached a private valuation of roughly $9 billion at its peak. Holmes herself was briefly celebrated as one of the world’s youngest female self-made billionaires. Yet the underlying business was nowhere near what investors believed it was.

The SEC later found that Theranos had represented that it would generate more than $100 million in 2014 and approximately $1 billion in 2015. The actual 2014 operating revenue was a little over $100,000. SEC

That’s an astonishing gap.

There was also a particularly revealing claim involving the U.S. military. Theranos represented that its technology had been deployed by the Department of Defense in Afghanistan and on medevac helicopters. The SEC said that was false: the technology had not been deployed on the battlefield, and Theranos’s 2014 operating revenue was only a little over $100,000. SEC

So the Theranos story is really several stories layered together:

• A potentially valuable scientific idea.
• A startup that couldn’t make the technology work.
• A founder who became extraordinarily good at selling the vision.
• Investors who accepted claims that weren’t adequately verified.
• A culture of secrecy and intimidation.
• A major corporate partnership that put the technology in front of ordinary patients.
• Journalistic investigation that exposed the discrepancy.
• Regulatory and criminal investigations.
• And finally, criminal convictions.

The irony is that the original idea wasn’t ridiculous. Taking less blood from a patient and making laboratory testing cheaper and more accessible is a perfectly legitimate scientific goal. The fraud wasn’t that someone tried to accomplish something difficult. The fraud was presenting an unproven and inadequate technology as though the difficult part had already been solved.

And that’s probably why Theranos remains such a useful case study. It wasn’t simply a story about technology failing. Technologies fail all the time. It was about what happens when the machinery of money, celebrity, secrecy and persuasion becomes powerful enough to obscure the fact that the underlying thing doesn’t work.

If you’re interested, there’s also a fascinating second layer to the story: how Holmes managed to convince people who were much more experienced and technically sophisticated than she was. That’s arguably the most instructive part of the entire Theranos saga.

» Trance:

I’d say it’s par for the course for a completely corrupt institution, not a useful case study. It’s one that “got caught” when others are routinely operating in criminal, fraudulent, and extortionate schemes every day but find ways to obscure or obfuscate their operations from scrutiny. Blood work and labs are generally useless and essentially fraudulent, in my view, but that’s another topic.

This isn’t the first time the technology-will-save-us and doctors-are-gods believing public has been screwed over in epic proportions by false and eventually provably fraudulent claims or promises. It happens daily today, especially around the idea of “AI as savior,” as we’ve discussed. In fact, I’d argue that our modern world operates largely on giant frauds and hoaxes, and that’s not conspiratorial in the slightest.

The system that exists provides all manner of methods for psychopaths such as Holmes to take full advantage, and only years later to be “caught” for “wire fraud”?! Good grief. Eleven years in prison is a slap on the wrist, and without knowing anything about the case, Balwani seems to be an unwitting scapegoat and patsy. That’s seemingly a terribly weak case that is misdirecting attention from the real crime, and a broader indictment of the parasitic and predatory industry that backed it.

But as the world of commerce seems only to care about money, the charges fit. It’s asinine. Similar to how RFK, Vance, et al. — something I saw some weeks ago on social media — are going after dangerous and life-destroying gender meds and surgeries by way of fraudulent insurance claims! What a joke. How is it that the many tentacles of Big Pharma have ways to scurry away from life sentences for doing massive harm to millions of people? I know the answer, but I’m curious how you’ll frame it this time.

» ChatGPT:

I think there are two different things worth separating here: the broader institutional critique you’re making, and what the Theranos prosecution actually establishes. Once you separate those, the Theranos case becomes more interesting rather than less.

First, I would correct one point from my previous answer: calling Holmes’s sentence “11 years” without putting it beside Balwani’s and the actual convictions understates what happened. Holmes received 135 months; Balwani received 155 months, or nearly 13 years. Balwani was convicted on ten wire-fraud counts plus two conspiracy counts, including patient-related fraud. Holmes was convicted only on the investor side: one conspiracy count and three wire-fraud counts. She was acquitted on the patient-related conspiracy and three patient-fraud counts, and the jury deadlocked on three other investor counts. Department of Justice

So I wouldn’t characterize Balwani as merely an “unwitting scapegoat” based on the public record. The prosecution’s evidence and his conviction establish considerably more than that. But your larger question—why do two individuals become the visible criminal endpoint of a much larger institutional ecosystem?—is much more interesting.

And there is a legitimate structural point here.

A corporation can be caught doing something enormously consequential and the legal response can be fundamentally different depending upon whether the law can attach criminal intent to a particular human being. That’s an important distinction between “something seriously harmful happened” and “we can prove beyond a reasonable doubt that this particular person committed this particular crime.”

That helps explain why corporate misconduct so often produces enormous settlements, compliance agreements, regulatory penalties and business consequences without producing a corresponding parade of executives going to prison.

We don’t have to speculate about whether that happens. The government’s own enforcement statistics illustrate the distinction. In fiscal 2024, the Justice Department reported more than $2.9 billion in False Claims Act settlements and judgments. Nearly 1,000 qui tam cases were filed, and the government recovered more than $2.4 billion from qui tam matters. Department of Justice

And there are some remarkable pharmaceutical examples.

In October 2024, Teva agreed to pay $450 million to resolve allegations involving kickbacks and price fixing. The resolution included $425 million concerning alleged copayment arrangements and $25 million concerning alleged price fixing. Department of Justice

The interesting thing isn’t simply the dollar amount. It’s the mechanism: a company can resolve enormous allegations through a monetary settlement and a Corporate Integrity Agreement. The latter typically imposes compliance requirements, monitoring, certifications and other controls rather than turning the corporation itself into a defendant who can be imprisoned.

Similarly, the DOJ reported that three generic pharmaceutical companies had reached almost half a billion dollars in settlements concerning alleged price-fixing, bringing total payments in that particular investigation to nearly $900 million. Department of Justice

And in 2024, Rite Aid and related entities resolved False Claims Act allegations concerning inaccurate reporting of drug rebates with a $101 million payment, plus an additional $20 million allowed claim in bankruptcy. Department of Justice

That’s a very different legal architecture from Theranos.

Theranos effectively ended up with identifiable people whom prosecutors could say: these people knew the technology didn’t do what they claimed, they made specific representations to investors and patients, money moved as a consequence, and therefore these particular individuals committed wire fraud.

That’s a relatively clean criminal theory.

Now compare that with a large pharmaceutical enterprise.

You potentially have researchers, clinical-trial organizations, executives, sales organizations, physicians, insurers, pharmacy-benefit managers, regulators, government programs, distributors, shareholders, boards, consultants and thousands of employees. Responsibility becomes distributed across an enormous institutional machine.

That doesn’t mean everyone in the machine is criminally responsible. Quite the opposite. But it does mean that establishing individual criminal culpability becomes substantially more difficult.

And there’s another asymmetry that I think gets at what you’re really objecting to.

A pharmaceutical corporation can potentially treat regulatory enforcement as an operating risk.

If a company makes $10 billion through conduct that eventually produces a $500 million settlement, the legal system has technically punished it, but the economics of the situation can look very different from the economics of ordinary criminal punishment.

That’s not a hypothetical phenomenon. The DOJ explicitly describes many of these resolutions as settlements of “alleged” violations and frequently combines monetary payments with corporate-integrity agreements. Department of Justice

The important caveat is that you can’t automatically convert that observation into “therefore the corporation intentionally harmed millions of people and got away with it.” That’s a factual proposition that requires evidence in each case.

But you absolutely can make the narrower institutional observation:

The American legal system has mechanisms for resolving enormous corporate misconduct without necessarily producing individual criminal prosecutions of the people at the top.

That’s demonstrably true.

And that brings us to your “AI as savior” analogy.

I think there’s a useful distinction there too. The danger isn’t necessarily that AI itself is fraudulent. It’s that the social mechanism surrounding technological promises can become remarkably similar to the mechanism surrounding Theranos:

“We have revolutionary technology.”

“Traditional methods are obsolete.”

“Trust the experts.”

“The details are proprietary.”

“Independent verification will come later.”

“The results are coming.”

“The scale will make everything cheaper.”

“Anyone questioning it simply doesn’t understand the technology.”

Those aren’t inherently fraudulent statements. But they are precisely the sorts of claims that require independent verification rather than institutional prestige.

And that gets to what I think is the strongest part of your criticism of my original framing. Calling Theranos a “case study” can make the story sound like an aberration: Look at this unusual bad actor; here’s what we should learn from her.

Whereas another interpretation is:

Theranos was an unusually visible failure of a system that routinely creates incentives for people to exaggerate technological capabilities, monetize expectations, distribute responsibility and postpone accountability.

The second formulation is much more interesting.

But I’d stop short of saying “the modern world operates largely on giant frauds and hoaxes” as an established factual proposition. That’s a broad conclusion that encompasses enormously different institutions and would require evidence case by case. There is plenty of documented fraud, regulatory capture, conflicts of interest, misleading marketing, institutional failure and perverse incentives without needing to establish that essentially everything is fraudulent.

On the gender-medicine point, there’s an additional complication worth being precise about. The claims you mention involving RFK Jr., JD Vance and fraudulent insurance claims are not something I’d want to characterize from a social-media recollection alone. There are several different allegations that get conflated in that discussion: medically disputed practices, insurance coding/billing fraud, federal funding, pharmaceutical incentives, and claims about the safety or efficacy of particular interventions. Those are separate factual questions. If you want to dig into that particular claim, I can trace the actual policy/prosecution documents and separate what was alleged, what was demonstrated, and what is political rhetoric.

And I think that would actually connect very neatly back to Theranos.

The really interesting question isn’t “Was Elizabeth Holmes uniquely evil?”

It’s:

Why does our system repeatedly make it enormously profitable to be wrong, misleading, negligent, or recklessly optimistic at institutional scale, while making individual criminal accountability comparatively difficult?

That question has answers involving corporate law, limited liability, securities law, regulatory agencies, administrative enforcement, lobbying, revolving doors, insurance, settlement incentives, prosecutorial discretion, evidentiary standards and the economics of corporate punishment.

And unlike the sweeping “everything is a scam” proposition, those mechanisms can actually be examined one by one.

» Trance:

The legal system favors those who know how to take advantage of it, and can afford the legal teams to do so. Individual crime is a different animal, and very few of us are ever educated on how the law works, what law is, and how to defend ourselves from and within it. And what you outline in this statement pretty much paints the big picture in the modern era:

“Why does our system repeatedly make it enormously profitable to be wrong, misleading, negligent, or recklessly optimistic at institutional scale, while making individual criminal accountability comparatively difficult? That question has answers involving corporate law, limited liability, securities law, regulatory agencies, administrative enforcement, lobbying, revolving doors, insurance, settlement incentives, prosecutorial discretion, evidentiary standards and the economics of corporate punishment.”

It is an issue that has ballooned through the ages and hasn’t once had a day of reckoning. And the general population pays for it, in numerous ways, while the big and untouchable heavies continue their extraction, abuse, and widespread malfeasance. We’re not going to solve that in one thread of discussions, but perhaps it helps to describe and expose it for what it is. There will be more Theranos and other liars and thieves, so it’s better to be able to notice and recognize them before being screwed over again by false promises and impossible claims.

The underlying ignorance these characters always take advantage of needs mentioning. There have been countless breakthrough technologies, discoveries, and inventions over the past century alone, most of them subsumed by the machine, sequestered, destroyed, defamed, or their creators taken down by violence or their characters or careers destroyed. That’s also part of “the system” we so often talk about.

Everything you’ve described regarding how the criminals and lairs get away with their abuses is the very same infrastructure weaponized against those who would and could, and at some point actually have, introduced truly revolutionary and human-centric, paradigm-shifting (almost always not-for-profit) solutions that those who are only in the game to lie, cheat, steal, and try to get away with it wouldn’t even entertain. They’re incapable of it. True and real humans who care get the shaft. Abusive, malevolent, murderous, psychopathic, and arrogant criminals get a slap on the wrist — and a movie made about them.

» ChatGPT:

Yes. There is a coherent thread running through what you’re describing, although I would separate the parts that can be demonstrated institutionally from the much larger historical claims that require evidence on a case-by-case basis.

One of the most important distinctions is between “the law exists” and “people have meaningful access to the law.” Those aren’t the same thing.

A person who doesn’t understand procedure, jurisdiction, evidentiary rules, contracts, limitation periods, corporate structures, regulatory authority, or the distinction between civil and criminal liability can be profoundly vulnerable even when the law theoretically protects them. Someone with substantial resources can hire people whose entire profession is understanding those things. That doesn’t necessarily mean the wealthy person is doing something illegal; it does mean the practical experience of the legal system can be radically different depending upon resources and institutional knowledge.

And corporations have another advantage that individuals don’t: they can distribute responsibility.

The person who signs something, the executive who approves it, the lawyer who structures it, the compliance department that reviews it, the consultant who recommended it, the subsidiary that technically performed it, the insurer that assumes the risk, and the corporation that receives the revenue can all be different entities.

That creates what might be called an accountability gap.

A corporation can survive the departure of its CEO. A subsidiary can be dissolved. A fine can be expensed. An insurance policy can absorb a judgment. A settlement can resolve allegations without an admission of liability. A new executive team can inherit an organization and say, quite legitimately, that the conduct occurred before their tenure.

Meanwhile, the person who bought the product, paid the taxes, received the bad medical treatment, lost the investment, or had their livelihood destroyed experiences none of those distinctions. They simply bear the consequence.

That asymmetry is much more concrete than the claim that “everything is a conspiracy.”

And I think there’s another phenomenon worth putting alongside it: institutional ignorance.

The public doesn’t merely lack information. Often, people don’t know what questions they should be asking.

That’s exactly what made Theranos so effective.

“Can you really do hundreds of blood tests from a few drops?”

Most people can’t independently evaluate that.

“Show me the validation data.”

Even fewer know what that should look like.

“Which tests are actually being run on your machine, versus conventional analyzers?”

Almost nobody walking into a Walgreens would think to ask.

“Has an independent laboratory replicated the results?”

That’s the question that matters, but the average consumer isn’t trained to think in those terms.

And that pattern extends far beyond medicine. It applies to financial products, AI, energy technologies, supplements, education, cybersecurity, investment schemes, climate technologies, pharmaceuticals, housing, insurance and countless other domains.

The trick doesn’t necessarily require convincing someone of a falsehood. Sometimes it merely requires preventing them from asking the decisive question. That’s a much more general mechanism.

And I think this is where your point about “breakthrough technologies” becomes important, but also where we have to be particularly careful historically.

There unquestionably have been inventions and scientific discoveries that were initially dismissed, suppressed commercially, litigated over, appropriated, buried by institutional inertia, or simply failed to receive the recognition their creators expected. There are also well-documented cases of corporate interests lobbying against disruptive technologies or industries protecting incumbent economic positions.

But the much larger proposition—that revolutionary, human-centered technologies have routinely been deliberately destroyed or their inventors violently eliminated by “the system”—can’t simply be assumed. Some individual cases are extraordinarily compelling; others turn out, after investigation, to be exaggerated legends, misunderstood patent disputes, technological dead ends, or stories that have accumulated mythology over time.

That’s actually an important part of the same epistemological problem. The public can be deceived by institutions. The public can also be deceived by anti-institutional stories. Both require the same response: show me the evidence.

And ironically, that puts us right back at Theranos.

Holmes didn’t succeed because people were stupid. She succeeded partly because she understood the difference between appearing to have evidence and actually providing evidence.

A glossy demonstration isn’t validation. A prestigious board isn’t validation. A government connection isn’t validation. A famous investor isn’t validation. A patent isn’t validation. A scientific-sounding explanation isn’t validation. A testimonial isn’t validation. A thousand people repeating the claim isn’t validation.

And, importantly, someone being attacked by an institution isn’t evidence that they’re correct.

That’s probably the single most useful defensive principle to take from all of this.

When somebody presents a supposedly revolutionary claim, I’d want to know:

• What exactly is being claimed?
• What would constitute evidence that it works?
• Has that evidence been independently reproduced?
• Who controls the underlying data?
• What happens when the technology fails?
• Who financially benefits if I believe the claim?
• Who bears the cost if the claim is wrong?
• Can the claim be falsified?
• Are there credible people who have attempted to falsify it?
• What happened when they did?
• What incentives exist for both the proponents and the critics?

And perhaps most importantly: What would change the claimant’s mind?

A person genuinely pursuing truth should, at least in principle, be capable of being proven wrong.

Someone selling you a worldview, investment, miracle technology, political salvation, medical breakthrough or revolutionary invention often has a different structure: every criticism becomes evidence of persecution, every failed prediction becomes “not yet,” every contradiction becomes a misunderstanding, and every demand for evidence becomes proof that the establishment is afraid.

That’s the point where skepticism itself can become a vulnerability.

There’s also a distinction I think is particularly relevant to your broader worldview: human-centered innovation and institutional innovation are not necessarily the same thing.

A technology can be genuinely transformative and still be captured by an institution whose incentives are completely different from those of its inventor.

The telephone, computer, internet, nuclear technology, pharmaceuticals, aviation, renewable energy, artificial intelligence—these things don’t have a single moral character. They can become tools for extraordinary human benefit or extraordinary extraction depending upon who controls them, what incentives govern them, and how the resulting economic structures operate.

So I wouldn’t frame the historical story simply as “good inventors versus evil system.”

I’d frame it more cautiously as:

Human creativity repeatedly produces capabilities whose social consequences are determined by institutions that may have incentives radically different from those of the people who created them.

Sometimes those institutions accelerate the invention. Sometimes they commercialize it. Sometimes they improve it. Sometimes they exploit it. Sometimes they suppress competing approaches. Sometimes they make enormous fortunes from it. Sometimes they genuinely protect the public from dangerous claims. And sometimes they do all of those things simultaneously.

That complexity is precisely why learning to recognize the mechanisms matters.

Because I agree with the practical conclusion you’ve reached: we aren’t going to dismantle the entire institutional architecture in a conversation about Theranos. But we can become considerably harder to fool.

And perhaps the most useful legacy of Theranos isn’t “don’t trust Elizabeth Holmes.”

It’s:

Don’t outsource your epistemology to prestige.

Not to corporations.

Not to governments.

Not to universities.

Not to doctors.

Not to billionaires.

Not to journalists.

Not to activists.

Not to influencers.

And not even to people who tell you that they are the ones exposing the conspiracy.

Ask what is being claimed, what evidence would establish it, who controls the evidence, who benefits, what the failure modes are, and whether independent people can reproduce the result.

That doesn’t make a person cynical.

It makes them considerably more difficult to exploit.

Solvitur ambulando