I've long felt that the tech industry is somewhat blinded by decades of Moore's Law and its equivalent in storage, networking, etc. 2-10x improvements are so normal for Silicon Valley that they begin expecting to find them everywhere.
In just about every other field or industry, a 10% improvement every decade is amazing and represents the combined efforts of tens of thousands of researchers. A 2x improvement is a once-in-a-lifetime revolution that will be backed my mountains of joint; and 10x (like Theranos promised) completely unheard of.
This also functions as a great BS detector - breakthroughs of the scale Theranos claimed just don't happen out of nowhere - there would be thousands of peer reviewed papers leading up to it.
"Extraordinary claims require extraordinary evidence", and Theranos didn't even offer an ounce of it.
I tried to search. Sadly can't find it at the moment.
But, there was/is a reddit thread that, without naming the company, originally asked is it normal for a company be "pumping numbers" in the medical field. It was an an engineer/scientist who asked it years before anything public. A number of commenters guessed it was Theranos, eventually the poster came back to confirm the suspicions while expounding on the whole experience.
this is the reddit that I miss, and this is the ONE example where it kind of sucks that we can comment on old posts. legendary thread; thank you for finding it.
Nah, that's doubtful. I've seen plenty of posts on HN and elsewhere that held the same sentiment. MANY people said that what they were trying to do was very difficult and they did not do anything to prove they had any technical advances to do it. The company was shady and overly secretive in a really strange way. The board was inappropriate for a biotech company.
Everyone with a background in biology or microfluidics said these things. It wasn't hard to find.
That's why I was so surprised when they partnered with Walgreens. I thought at the time that Walgreens must know something that the public didn't.
Al Capone went to jail for tax evasion. She got convicted for fraud because it was easy to prosecute, she got sentenced at least partially for cynically endangering peoples lives. There's clearly an argument that this is an abuse of the legal system, but she's not being singled out here.
People don't get convicted for doing bad things. They get convicted for breaking specific laws. In this case, the laws against defrauding investors were clearly broken, to the extent that a jury unanimously agreed they were broken. Whatever laws might have applied to the patients did not meet that bar.
I keep seeing a lot of "throw them in jail" posts here on HN and it's bad. It should be very hard for the state to jail someone. It should require extremely solid proof that they committed a crime which was defined in law. If we lose that principle, we revert from rule of law to rule of judges and mob rule. Some people like Holmes don't get punished in a way that satisfies everyone they hurt, and people like SBF might never get punished, and that's fine because lowering the burden of proof would only result in more injustice.
The fact that the law punishes defrauding investors, but not defrauding patients or employees, just shows that those laws are written by rich people, for rich people.
because the prosecutors don't have good evidence of anyone actually harmed by the bad tests, and the testing agreements probably got the patients to agree if it didn't work they wouldn't rely on it.
prosecutors don't have good evidence of anyone actually harmed by the bad tests
I think it's more nuanced than that. Even if the prosecutors had mountains of evidence of harm to patients, defense attorneys have a lot of experience casting doubt on that kind of evidence making trail very tricky. My wife works in the medical/legal overlap and constantly sees stuff like "yes the patient spent 30 years working in an asbestos plant next to an industrial incinerator, but they smoked a cigarette once in the bathroom in high school so you can't say their cancer came from any one cause". And sometimes juries buy it.
But when you accumulate mountains of evidence that basically says "yeah, we know our stuff doesn't work...fake it till you make it, yo", you've got counterarguments like "we were just kidding" and "it's all fake" that isn't as resonate with the jury. So, the prosecution goes with the high-confidence charges.
The whole thread is an interesting read in retrospect, quite a mix of effusiveness and skepticism: https://news.ycombinator.com/item?id=6349349