Gwern's Reviews > Bad Blood: Secrets and Lies in a Silicon Valley Startup

Bad Blood by John Carreyrou
Rate this book
Clear rating

by
11004626
's review

really liked it

Bad Blood is a straightforward read about the rise and fall of Theranos, done in chronological order in third-person up until Carreyrou becomes personally involved, at which point things accelerate to the SEC civil settlement. Carreyrou doesn't end too strongly but says that the criminal investigation may well end up charging Holmes & Sunny. This means that it lacks a really conclusive 'ending': Theranos was continuing to limp on, having received funding from a vulture on the strength of its patent portolio, ironically enough, which apparently was valued at $1b, and Carreyrou mentions in one interview that Holmes was reportedly scouting VCs for a new startup. (After reading BB, I had to think: maybe a second Holmes startup isn't a bad idea - after all, if she could get this far with no working product at all, what could she do with an actual product? It may look bad, but it'd probably work better than most startups.) Coincidentally, I began reading this just hours before Holmes & Sunny were criminally indicted (vindicating what I had been telling people - the SEC civil settlement didn't mean they were going to get off scot-free). Good timing on my part. This puts more of a period on reading BB, although the story is far from over. There's a quip that the most American character is the conman, because America is the land of second chances - Elizabeth Holmes is only 34 years old, after all, and even having aggravated the DoJ by persisting with Theranos, it's hard to imagine her being sentenced (as a woman and without a lot of bodies and without Shkreli's autistic genius for infuriating judges) to more than a few years at worst, so I wonder if we've seen the last of her?

In any case, BB is good for resolving a lot of details about Theranos.

For example, I was perplexed at the time by the large Walgreens deal: Walgreens is a large, competent, sophisticated provider of pharmacy services, well capable of thorough testing; if Theranos was not what it was hyped up to be, how could Walgreens fail to notice? My assumption was that Theranos had done something clever to produce fake results (if not perhaps as clever as the FSB at Sochi). BB provides the answer, which is dismayingly mundane: Theranos bluntly refused to provide any kind of real validation or access to its machines, and some Walgreens execs were furious about it and correctly convinced Theranos was a fraud, but others were seduced by the vision, and the doubters signed on because they were terrified of forcing Theranos into the arms of CVS, which is a rivalry I had no idea about. ("Van den Hooff listened with a pained look on his face. 'We can't not pursue this,'' he said. 'We can't risk a scenario where CVS has a deal with them in six months and it ends up being real.' Walgreens's rivalry with CVS, which was based in Rhode Island and one-third bigger in terms of revenues, colored virtually everything the drugstore chain did. It was a myopic view of the world that was hard to understand for an outsider like Hunter who wasn't a Walgreens company man. Theranos had cleverly played on this insecurity. As a result, Walgreens suffered from a severe case of FOMO - the fear of missing out." Who knew?) A similar desperation appears to have animated Safeway's ill-fated Theranos commitment. And the general coverup appears to have owed much to the realities of lawfare in the USA: Theranos had enough cash to wield legal threats against the justly-terrified whistleblowers, costing Tyler Shultz a staggering $400,000+ and gaslighting suspects with constant PI surveillance, and possibly tactics that went beyond the legal (Theranos/Holmes appear suspiciously well-informed at times). It's no surprise it took a major newspaper like the WSJ to investigate it.

It's also interesting for the unexpected details. For example, dressing like Steve Jobs wasn't Holmes's idea! She was told to do it by one of her ex-Applers. And her family connections were dangerous as much as they were helpful: the shiny board of directors, for everyone it impressed, put other people off and made them suspicious, and without her family connections, the family friend Richard Fuisz would never have tried to patent-troll her out of peevish spite which directly fed into the first Fortune article and eventually Carreyrou's own investigation. (With 'family friends' like these, who needs enemies?)

And Carreyrou is good about considering to what extent Theranos really reflects on SV: as he points out, a lot of the actual investors were 'dumb money' (my phrase) who did minimal real due diligence and ignored red flags, like Rupert Murdoch who put in $125m on the basis of 2 meetings with Holmes and a phone call to someone else, while the usual life-sciences VCs were unimpressed with Holmes's bluster & ignorance and took a total pass on her. (Google Ventures took a hard pass when their guy walked into a Walgreens and Theranos couldn't do the test using just a nanotainer of his blood - a simple test that many others also did but then ignored the excuses and failures.) Culturally, Theranos was barely SV: yes, Apple may have fanatical internal secrecy, but they are the exception that proves the SV rule and have suffered for it (in machine learning especially), while everyone else adopts considerably more internal transparency for precisely the reasons that Theranos employees cite - how do you sanely do R&D if no one is allowed to talk to each other? (Again, Apple has suffered for this in trying to keep up in non-materials-science and non-manufacturing R&D, like machine learning: what's the last impressive new tech you can think of which was developed inside Apple?) It's not really that easy to draw a novel lesson here. Was Theranos initially too ambitious? Perhaps, but lots of startups scale back or pivot to new ideas based on their trial-and-error; reality cannot be planned out. Did it get too much money? It raised $6m initially, which is not that much for their purpose. Should new startups not be funded at all or not allowed a decade+ to work out ideas, or Walgreens blamed for seizing on a new opportunity as fast as possible? But people already complain about investors being too risk-averse and short-term (despite Theranos being 17 years old now!) and companies being bloated slow bureaucracies. Was the problem lack of 'peer review'? Except peer review doesn't work and isn't scientific, works the worst in cases of fraud (think of all the cases of people fabricating scores or hundreds of papers which slide through 'peer review' only to finally be exposed not by 'peer review' but when the results failed to replicate), and would've been inferior to simply seeing if the tests worked or not, and that's how all the smart money like Google Ventures took a pass on Theranos. Should we outlaw investing millions of dollars based on a phonecall? Hard to imagine that working out well. Should we criticize VCs for being gullible? But most of the VCs (not) involved weren't gullible! Should we criticize the board for letting her accumulate so much stock and then letting her talk them out of firing her in 2008? Probably, yes, but hindsight is 20/20 and the worst problems hadn't happened yet. Should blood testing in general be verboten to investors? But Holmes is very, very, far from the first person to try to improve on existing blood tests and fail, much like the perennially fruitless quest for a Alzheimer's disease cure - a good book on this topic is John Smith's The Pursuit of Noninvasive Glucose Blood Tests: "Hunting the Deceitful Turkey" documenting the endless failure of people trying to improve on finger-stick blood glucose tests for diabetics - and people keep trying because anyone who succeeds will make so much money because the human costs of failing to succeed is measured in hundreds of millions or billions of lives over the coming centuries, and failure is simply not an acceptable option.

Carreyrou suggests toward the end that Holmes might have psychopathic traits:

A sociopath is often described as someone with little or no conscience. I'll leave it to the psychologists to decide whether Holmes fits the clinical profile, but there's no question that her moral compass was badly askew. I'm fairly certain she didn't initially set out to defraud investors and put patients in harm's way when she dropped out of Stanford fifteen years ago. By all accounts, she had a vision that she genuinely believed in and threw herself into realizing. But in her all-consuming quest to be the second coming of Steve Jobs amid the gold rush of the "unicorn" boom, there came a point when she stopped listening to sound advice and began to cut corners. Her ambition was voracious and it brooked no interference. If there was collateral damage on her way to riches and fame, so be it.


I think this is wide of the mark and he gets closest in the final lines. What is the stereotypical profile of psychopathy? One might put it as: someone who is unable to make or commit to plans, who acts spontaneously on selfish and often self-destructive impulses, covering up for it with manipulation of others or with even more brazen deceptions often so ill-thought-out & easily falsified as to beggar belief, with a history of violence (often unreported) and especially sadistic cruelty (often emerging during childhood and focusing on animals), unable to maintain long-term relationships, sexually promiscuous and often impregnating or pregnant at an early age, often below average intelligence, greedy and covetous of money or rewards, apt to embezzle or steal from employers, typically racing from employer to employer to outrun immune systems etc.

The portrait of Holmes in BB is very far from this. There is no hint of tendencies towards sadism or violence in her childhood, merely a mention of competitiveness. Holmes is, at least initially, quite bad at self-presentation: One quoted VC paraphrased describes her early pitches as unimpressive: "she'd come off as a dowdy young scientist back then, wearing Coke-bottle glasses and no makeup, speaking nervously to an audience of men two to three times her age" and Carreyrou points out (to my surprise) that her Jobsian wardrobe wasn't even her idea - but that of an Apple designer she hired:

Ana felt that Elizabeth could use a makeover herself. The way she dressed was decidedly unfashionable. She wore wide gray pantsuits and Christmas sweaters that made her look like a frumpy accountant. People in her entourage like Channing Robertson and Don Lucas were beginning to compare her to Steve Jobs. If so, she should dress the part, she told her. Elizabeth took the suggestion to heart. From that point on, she came to work in a black turtleneck and black slacks most days.


An additional interesting thread throughout BB (although Carreyrou puts no emphasis on this and I wonder if he missed the connection) is how Holmes continuously sought to amass more stocks or voting control of Theranos: one oddity in the end of the Theranos saga was that Holmes was never, and could not be, fired because she continued to own so much stock and voting power. Rather than selling out early and retiring to a life of leisure, she held on to the bitter end. This is particularly striking because, if I'm reading the timeline and indictment right, Theranos reached valuations of $50m+ long before Holmes/Sunny ever did anything that was truly fraud and irreversible; as far as I can tell, Holmes could have sold millions of dollars of stock and left at many points, entirely safely, and when Theranos ran out of runway, it would be regrettable but nothing she could go to prison for. Instead, she invested considerable efforts into clawing back the large, near co-founder-level stake of her first employee, to the point of threatening to sue an extremely wealthy director who wanted to buy some of it himself rather than giving it to her at a huge discount; she further proposed in 2007 allotting a block of stock to a nonprofit foundation in perpetuity (controlled by, of course, herself); and whenever an employee was fired, Theranos practice seems to have been to carefully hunt using coworkers & laptops & files for any reason, no matter how spurious, to clawback stock options.

And in Theranos's mismanagement, we don't see much that could be described as sadistic beyond ordinary bounds - indeed, the 'disappearing' is about separating people from Theranos as quickly and totally as possible, rather than toying with their prey. The disappearing served a useful role in enforcing compartmentalization, risk-aversion, and covering up information, but might there not be another reason?

Ian Gibbons puts his finger on it exactly when he said that "It's a folie à deux ." Or perhaps it would be more precise to invoke narcissistic personality disorder and compare Elizabeth Holmes to Donald Trump.

Holmes did not start off as a psychopath determined to rip off VC and SV by using her cunningly honed social skills and sexuality to manipulate horny old white men, as one narrative goes. She was a normal ambitious Stanford undergrad (having met a dozen or so Stanford undergrads recently, Holmes now seems much more understandable to me), perhaps a little too eager to launch a startup, with delusions of grandeur about a entrepreneurial destiny and a bit of a chip on her shoulder; for reasons which cannot be known (as counterfactuals are not observable), she got lucky or was female or had family connections or something and she got some VC and support from her professors for what was a more feasible sort of idea which might've been workable, dropped out for a startup, was mentored by the likes of Larry Ellison (surely a red flag if ever there was one), hooked up with an entrepreneur even luckier & more delusional in a remarkably long-term monogamous relationship, selected for employees who initially offered helpful advice in fitting into SV tropes & self-presentation but gradually were recycled into sycophants and slaves, and developed her reality-distortion field abilities through practice and self-persuasion and a cultivated paranoia/martyr complex, and mutual narcissistic feedback loops with true-believer employees and Sunny and eventually the media, 'vanishing' anyone who threatened to damage her narcissistic supply and punishing them for being wretched hateful human beings and endangering the mission, all of which lasted for many years (while Theranos was only truly in the public eye from 2014-2017, the first version was founded in 2003, fully 11 years before!). That's very different, even if the end game, where criminal fraud and blatant lies are necessary to keep the show going, looks similar.
68 likes · flag

Sign into Goodreads to see if any of your friends have read Bad Blood.
Sign In »

Reading Progress

June 15, 2018 – Started Reading
June 15, 2018 – Finished Reading
June 17, 2018 – Shelved

Comments Showing 1-7 of 7 (7 new)

dateDown arrow    newest »

Peter Tillman A remarkably good review/analysis/essay. Thank you!


Paul One quibble - you're confusing sociopathy with psychopathy. These are different (though psychopaths are also sociopaths), and Carreyrou is right to say Holmes has some sociopathic traits, I think


Gwern Paul: that's an old argument in the literature (eg comes up tediously often in Handbook of Psychopathy). I find it unconvincing either way and don't bother to police the distinction in my writing - they can't agree on the difference, so any attempt on my part is a waste of time and lost on readers anyway. Holmes may have some -pathy symptoms, but of course symptoms can overlap a great deal between disorders, and I think the picture as a whole is much more consistent with narcissism.


Paul yeah, I agree that narcissism is probably a closer fit


Amanda Great review and it helped fill in some of the holes. As someone who really doesn't have any experience with sv startups I had no idea what is normal and wasn't isn't normal for that area.


message 6: by Gwern (last edited Dec 30, 2018 06:22AM) (new) - rated it 4 stars

Gwern There was a lot that was abnormal. At least of the startups I know about, most of them are actually quite forthcoming, in person if not online. Taking secrecy to the level Theranos did for so many years (decades!) is, outside of Apple where it's a way of life (and an exception to many rules), and outside of temporary 'stealth mode', quite unusual. Usually, startups are 'default dead' and threatened far more by obscurity than by publicity... They need customers more than they need stealth, because the incumbents can't or won't take their idea (either because they are too slow, think it's unprofitable, or that it's a lousy idea), in part for recruiting etc, and publicity is vital to them.


Megan Great review! I'm reading the book now and it really feels like narcissism to me as well. Especially how she turns on anyone to criticize the company in any way, such as when she fire Mosley.


back to top