4 ms·

My question is: Did they have a prototype or at least a concept?

JumpCrisscross 8y ago
> Did they have a prototype or at least a concept? Not one that could do what they promised. "The indictment alleges that the defendants used a combination of direct communications, marketing materials, statements to the media, financial statements, models, and other information to defraud potential investors. Specifically, the defendants claimed that Theranos developed a revolutionary and proprietary analyzer that the defendants referred to by various names, including as the TSPU, Edison, or minilab. The defendants claimed the analyzer was able to perform a full range of clinical tests using small blood samples drawn from a finger stick. The defendants also represented that the analyzer could produce results that were more accurate and reliable than those yielded by conventional methods—all at a faster speed than previously possible. The indictment further alleges that Holmes and Balwani knew that many of their representations about the analyzer were false. For example, allegedly, Holmes and Balwani knew that the analyzer, in truth, had accuracy and reliability problems, performed a limited number of tests, was slower than some competing devices, and, in some respects, could not compete with existing, more conventional machines." https://www.justice.gov/usao-ndca/pr/theranos-founder-and-former-chief-operating-officer-charged-alleged-wire-fraud-schemes https://www.justice.gov/usao-ndca/pr/theranos-founder-and-fo...
riantogo 8y ago
Yes, they did. Which involved humans walking over the blood sample to a regular lab setup where they would dilute it and use regular machines to output wrong results. Now all they had to do was find a solution to replace the humans, not having to dilute the sample and output correct results. But pesky investigations got in the way.
edmundhuber 8y ago
From "Bad Blood": they had a prototype called Gluebot which used a robot arm to do standard tests involving a pipette autonomously. However, their tests (using blood collected at Walgreens) were done on "hacked" (altered to cope with much smaller amounts of blood -- by increasing dilution) Siemens blood-testing machines.
- sjm-lbm 8y ago
  IIRC from the book, Gluebot was actually the prototype that eventually became the Theranos Edison.
tptacek 8y ago
They did. They had two devices, the "Edison" and the "MiniLab". The Edison was intended to run a full panel of blood tests, both chemical and immunological assays, off a finger stick of blood. It was also intended to allow Theranos to function as its own lab. The Minilab was intended to run a full panel of blood tests I believe off a larger venous sample, and was expected to be marketed to existing labs (at this point, Theranos had been forbidden from operating its own labs). In its own labs, Theranos made minimal use of the Edison, which was unreliable and supported only a few tests. Rather, they're accused of taking finger-stick draws, diluting them, and then feeding them into hacked-up adapters for COTS lab machines.
- JumpCrisscross 8y ago
  Do you believe Theranos could have beee successful company with less delusional management? (Honest question for a HN colleague for whom I have tremendous respect.)
  - JauntTrooper 8y ago
    I’ve never seen any evidence the underlying tech ever actually worked.
  - tptacek 8y ago
    I don't know anything about this space; all I know comes from Carreyrou's book. Somewhere towards the middle, one of his sources observes that it's particularly hard to get a good sample from a finger stick --- something about damaged blood cells --- and you can find recent journal articles online suggesting that finger stick samples are inherently a lot less reliable than venous draws. So, I sort of doubt it.
  - dbt00 8y ago
    Not a domain expert, but as I recall most of the things they want to test for aren't present at a sufficient level in the blood stream to get a statistically significant sample from a few drops. Blood sugar is a relatively high percentage of your blood stream, which is why strip-based glucose monitors work.

> Did they have a prototype or at least a concept? Not one that could do what they promised. "The indictment alleges that the defendants used a combination of direct communications, marketing materials, statements to the media, financial statements, models, and other information to defraud potential investors. Specifically, the defendants claimed that Theranos developed a revolutionary and proprietary analyzer that the defendants referred to by various names, including as the TSPU, Edison, or minilab. The defendants claimed the analyzer was able to perform a full range of clinical tests using small blood samples drawn from a finger stick. The defendants also represented that the analyzer could produce results that were more accurate and reliable than those yielded by conventional methods—all at a faster speed than previously possible. The indictment further alleges that Holmes and Balwani knew that many of their representations about the analyzer were false. For example, allegedly, Holmes and Balwani knew that the analyzer, in truth, had accuracy and reliability problems, performed a limited number of tests, was slower than some competing devices, and, in some respects, could not compete with existing, more conventional machines." https://www.justice.gov/usao-ndca/pr/theranos-founder-and-former-chief-operating-officer-charged-alleged-wire-fraud-schemes https://www.justice.gov/usao-ndca/pr/theranos-founder-and-fo...

riantogo 8y ago

Yes, they did. Which involved humans walking over the blood sample to a regular lab setup where they would dilute it and use regular machines to output wrong results. Now all they had to do was find a solution to replace the humans, not having to dilute the sample and output correct results. But pesky investigations got in the way.

edmundhuber 8y ago

From "Bad Blood": they had a prototype called Gluebot which used a robot arm to do standard tests involving a pipette autonomously. However, their tests (using blood collected at Walgreens) were done on "hacked" (altered to cope with much smaller amounts of blood -- by increasing dilution) Siemens blood-testing machines.

sjm-lbm 8y ago
IIRC from the book, Gluebot was actually the prototype that eventually became the Theranos Edison.

tptacek 8y ago

They did. They had two devices, the "Edison" and the "MiniLab". The Edison was intended to run a full panel of blood tests, both chemical and immunological assays, off a finger stick of blood. It was also intended to allow Theranos to function as its own lab. The Minilab was intended to run a full panel of blood tests I believe off a larger venous sample, and was expected to be marketed to existing labs (at this point, Theranos had been forbidden from operating its own labs). In its own labs, Theranos made minimal use of the Edison, which was unreliable and supported only a few tests. Rather, they're accused of taking finger-stick draws, diluting them, and then feeding them into hacked-up adapters for COTS lab machines.

JumpCrisscross 8y ago
Do you believe Theranos could have beee successful company with less delusional management? (Honest question for a HN colleague for whom I have tremendous respect.)
- JauntTrooper 8y ago
  I’ve never seen any evidence the underlying tech ever actually worked.
- tptacek 8y ago
  I don't know anything about this space; all I know comes from Carreyrou's book. Somewhere towards the middle, one of his sources observes that it's particularly hard to get a good sample from a finger stick --- something about damaged blood cells --- and you can find recent journal articles online suggesting that finger stick samples are inherently a lot less reliable than venous draws. So, I sort of doubt it.
- dbt00 8y ago
  Not a domain expert, but as I recall most of the things they want to test for aren't present at a sufficient level in the blood stream to get a statistically significant sample from a few drops. Blood sugar is a relatively high percentage of your blood stream, which is why strip-based glucose monitors work.