ceterum_censeo

I'd like to hear your theories. Also, a few suggestions for the sites that you mention.

Would anyone who has experience using Cleanroom and PSP be willing to opine on the merits of both? I'm interested in lightweight but rigorous processes for improving software quality on regulated products, and both Cleanroom and PSP are on the radar.

They did indeed have an investigative unit, but it was recently poached by CNN: http://www.politico.com/media/story/2016/10/cnn-hires-away-b...

e: I don't recall any of their articles specifically, but I do recall reading a few solid articles via Instapaper and then being surprised to see that they were from BuzzFeed.

Reading what you've written, it sounds like you are saying that Altman and people associated with YC are sufficiently petty that they would refuse to work with you in the future because of a political disagreement. Am I interpreting this correctly? That would seem to be at odds with Altman's actions and public persona so far.

What repercussions? The situation strikes me as asymmetric: no matter what action Altman or Graham take they will be heavily criticized; for Ptacek, the worst would seem to be disagreement with his well-reasoned dialogue.

It seems unfair to force this to become a very public conversation.

> But we are all very much entitled to criticize what Altman is doing, and we would be doing Altman no favors by withholding that criticism.

Why are you entitled to do it in the most publicly way you are able?

If Thiel donated $1.25M to the NRA, would this lead to the same calls for YC to distance themselves from Thiel? Where does this end?

> So far not a single investor has claimed that their DD team did not have access, that their DD team was competent to do the job in the first place and that the technology as evaluated was sound.

This doesn't tell us one way or the other if DD took place and if so, the scope.

> The technology enabled that. Without the tech that failed to materialize and is not the core of the innovation that Theranos represents.

The technology is orthogonal to their regulatory machinations. They could have developed an in-house ELISA (or whatever, something more conventional) and made similar regulatory plays as they did with their microfluidics system.

> Yes, but if you run enough samples then you should converge at something amounting to statistical equivalence. If that doesn't happen then either one of the tests is invalid.

That handwaves over the complexity of the study, which goes to my point that your trivial and foolproof experiment is neither. A proper one probably needs a CRO. For a sizable investment it could be worthwhile, but I think you've underestimated the scope of your proposal.

> As are secrecy, refusing access to labs, not providing full details on the technology under NDA, refusing to run a sufficient number of blind tests at the request of the DD team.

Some of that in context isn't outside the behavior of other medical technology companies. It would be negligent to allow someone that wasn't with the government to walk through my lab without making some sort of arrangement ahead of time, and certainly under no circumstances unsupervised. I also have withheld key technical details to someone running DD despite having an NDA because NDAs cost money I didn't have to enforce if needed. The last item about testing is certainly a huge red flag, but [citation needed] if you're saying that Theranos did it (I find it possible to believe, but I haven't read that until your comment).

> If that were true that would mean they had working technology, which they did not have at all.

That might be an overstatement? I've followed this case closely, and I've had a very difficult time understanding what was claimed to work and did, what was claimed to work and didn't, what wasn't claimed by Theranos but is rumor transmuted into fact by the internet, etc. I've been operating under the assumption that the technology did work under controlled conditions and had problems when scaling up, but that is speculative on my part. I'm not aware of any scientific or medical reasons to believe that their platform cannot work, I am however aware of medical reasons to expect challenges to overcome (off-hand, sample volume, collection site), have violent disagreement as to the importance of consumer-intiated testing, etc. Unlike uBeam, there doesn't seem to be fundamental reasons to believe that Theranos' platform could not work and microfluidics being known to the art is supportive. If there are fundamental reasons that the technology cannot work, or if you have good hard sources for what Theranos had working and when, please do share!

> Even for some of the most basic testing they had to fall back on equipment supplied by others and this would have definitely turned up in a real DD, barring magic I don't see how they could have gotten away with it in any realistic setting.

"That's a great observation. It's true we do use conventional equipment for some basic tests and we are up front about that with our strategic partners, and it's simply because it's cheaper to run basic tests on the conventional equipment than on the Edison 1.0. Rather than waste money by insisting all testing run on Edison 1.0, we actually take the savings from running the basic tests on the conventional equipment, and roll that into the development budget for Edison 2.0 which we project will handle 98% of our full panel. Being intimately familiar with conventional equipment has other benefits: it's an important element of our development process to help us to better understand how we can deliver value-based differentiators on Edison 2.0." (or something like that) If they're willing to lie, they can spin a nice yarn and back it up with phony data. It won't fool everyone, but they don't need to fool everyone.

> Then please provide some proof that proper DD was done.

If you are stating that "they [investors] failed to do due diligence", then the onus is on you to provide proof that the proper DD was done if you don't want your argument referred to as speculation. I am not drawing a conclusion one way or the other.

> The disruption came along one parameter only, the amount of blood required to do the test.

That wasn't the only disruption and I'm not sure that was the primary one: I would argue the primary disruption was regulatory (using/abusing the LDT framework and consumer-initiated testing).

> It is trivial to come up with a foolproof test to see whether or not Theranos has a working device.

I would never use the words trivial and foolproof for anything in medicine or laboratory testing: there is nothing simple about running a comparison between labs, especially for biological samples. Run an identical sample at Quest and Labcorp and the values will differ, perhaps in a statistically-significant way.

I agree in that I would expect that proper due diligence should have recommended against an investment in Theranos. Not participating in the scientific community is a giant red flag. I completely disagree that technical DD on the technology platform would necessarily reveal the flaws in Theranos.

It doesn't strike me as charitable to take your strong stance that the investors were incompetent based on speculation with the benefit of hindsight.

I agree that they would be negligent if they did not run DD with a scope that included a review of the Theranos platform technology and implemented devices, but it is entirely possible to do proper DD and be fooled by a skilled charlatan.

> they failed to do due diligence before investing

Do we know the scope of the DD that was done?

It's hard to say exactly at what point it would be an appropriate read, but perhaps Knuth's The Art of Computer Programming should be on the reading list.

The EU strategy can solve regulatory problems seen in the US, but it introduces other problems to be solved: funding can become more difficult (although the normal rules wouldn't seem to apply to this group), import/export and taxes become complicated, management from a distance is hard, etc. Keeping in mind that Theranos was trying for a US government contract, their US focus makes sense and the LDT pathway was a good idea. It's a shame that their corporate culture and management were wholly inadequate.

They had to run their own labs in order to take advantage of the laboratory developed test (LDT) regulatory framework. Selling their devices for others to use would have required FDA clearance.

http://www.fda.gov/MedicalDevices/ProductsandMedicalProcedur...