|
Arsenic Lupin posted:Ooh, somebody who knows! Thank you. Your comment about needing to get FDA approval for every one of 70 tests blew my pigtails back. Golly. It hadn't occurred to me how much certification they were handwaving. I'm not an expert on the FDA approval process, but I believe it typically takes years. Also, the tests Walgreens/Theranos were offering are all off-the-shelf tests, if performed using existing (and well-validated/FDA approved) methods, so maybe it took the FDA a while to pick up on it? I mean, they shut Theranos down about two years after the announcement, and I guess a year and a half after the actual rollout? It's worth noting that Theranos did get one of their (70) tests FDA approved during this time. It may also be that Theranos misled them by saying that all the not-yet-approved tests were being carried out at the partner lab using conventional methods. Incidentally, this also happened to 23andme, although in their case the FDA only shut them down when they stopped responding, after several years of negotiations. They've since removed the more seriously health-related SNPs from their panel, at least until they've gone through FDA approval for those one by one. quote:That phase didn’t last for long, because there is much more interesting stuff in your genome than novelty items. Certain regions signal an increased risk of breast cancer, the impending onset of metabolic diseases, and sensitivity to medications. 23andMe—as well as a number of other companies—edged closer and closer to marketing their services as a way of predicting and even preventing health problems. And any kit intended to cure, mitigate, treat, prevent, or diagnose a disease is, according to federal law, a "medical device" that needs to be deemed safe and effective by the FDA. Since mid-2009, 23andMe has been negotiating with the agency, and in July 2012, the company finally began the process of getting clearance from the FDA to sell the kit that it had already been selling for five years.
|
#
?
Sep 7, 2016 02:15
|
|
|
|
| # ? Apr 29, 2024 01:50 |
|
|
I've got an effortpost on the subject in one of the pseudoscience or natural products fallacy threads, but basically the FDA is operated with a tiny fraction of its necessary budget, and is the eternal target of a really scary anti-regulation think tank that likes to set up test cases to remove whole swathes of FDA regulatory authority. This colors their approach to things- enforcement action is expensive and risky. I'm not sure that's relevant to what's happened with Theranos- I'd need to look more closely at how their stuff was actually defined under the law.
|
|
#
?
Sep 7, 2016 02:36
|
|
|
namaste faggots posted:I'm still amazed that Holmes got this far without I think I posted these six months ago, but just in case I didn't... here, have the Aug 2015 483 citations from their factories last year. Note that some of the concerns actually date from 2014. http://www.fda.gov/ucm/groups/fdagov-public/@fdagov-afda-orgs/documents/document/ucm469395.pdf http://www.fda.gov/ucm/groups/fdagov-public/@fdagov-afda-orgs/documents/document/ucm469396.pdf Anyone who had any experience in the medical device or pharma industries who took a look at these would never give a dime to this loving company, at least not after the initial pre-product hype phase. These are seriously some of the most blatant failures of GMP standards I've seen, and I've worked for some serious shitholes in pharma. These people straight up did not have a product, period.
|
|
#
?
Sep 7, 2016 02:36
|
|
|
Here's a story from the Verge detailing how they did it. I know relatively little about medical devices, so I can't speak to its accuracy without some time for research.
|
|
#
?
Sep 7, 2016 02:42
|
|
|
Lead out in cuffs posted:
Also, lol at the introduction of the article. quote:SA Forum is an invited essay from experts on topical issues in science and technology.
|
|
#
?
Sep 7, 2016 03:39
|
|
|
Sundae posted:I think I posted these six months ago, but just in case I didn't... here, have the Aug 2015 483 citations from their factories last year. Note that some of the concerns actually date from 2014. They were using an unvalidated Excel spreadsheet to store [redacted, but presumably important] data.
|
|
#
?
Sep 7, 2016 03:42
|
|
|
Sundae posted:I think I posted these six months ago, but just in case I didn't... here, have the Aug 2015 483 citations from their factories last year. Note that some of the concerns actually date from 2014. how was Theranos this stupid?? The FDA has their GMP guidelines, and yeah you have to follow them, but basically you need to setup your organization in such a way that it can follow them. I'm sure it is really complicated if you're completely new and not contracting any advisors, but it it's kinda not optional Slanderer fucked around with this message at 00:48 on Sep 8, 2022 |
|
#
?
Sep 7, 2016 03:49
|
|
|
Slanderer posted:As someone who develops software for medical devices for a living, holy poo poo how was Theranos this stupid?? The FDA has their GMP guidelines, and yeah you have to follow them, but basically you need to setup your organization in such a way that it can follow them. I'm sure it is really complicated if you're completely new and not contracting any advisors, but it it's kinda not optional Holmes had the family and personal connections to call a prestigious Marine general when someone raised questions about the validity of the tech for battle-field use and have him shut down the nay-sayers. He subsequently got a job on the board of Theranos. Go figure. Just as a reminder, Henry Kissinger was on the board of this company. That is only possible when you have money and political connections that even get you a conversation with the man, nonetheless his role on your corporate board.
|
|
#
?
Sep 7, 2016 03:57
|
|
|
General "Mad Dog" Mattis from Generation Kill was on the board. lol
|
|
#
?
Sep 7, 2016 04:04
|
|
|
Jumpingmanjim posted:General "Mad Dog" Mattis from Generation Kill was on the board. lol hahahah holy poo poo not him nooooooooooooooooo edit oh I was thining of godfather namaste friends fucked around with this message at 04:18 on Sep 7, 2016 |
|
#
?
Sep 7, 2016 04:11
|
|
|
namaste faggots posted:hahahah holy poo poo not him nooooooooooooooooo https://www.youtube.com/watch?v=fTXzcILPPp8
|
|
#
?
Sep 7, 2016 04:18
|
|
|
The fact that Theranos got as far as it did probably has as much to do with the power of old money families and old boy networks as it does VC incredulity. Under normal circumstances, someone like Holmes without her pedigree could probably get a little bit of seed capital if her song and dance was good enough, but it wouldn't be enough to sustain the company once the initial results didn't pan out and the company would probably have quietly folded by the time it got around to needing real institutional rounds However, because of her connections she was able to line up a lot more investment than was normal for a company that was essentially pre-product, and then that large chunk of cash became a self-fulfilling prophecy. VCs are basically herd animals, and FOMO is pretty real with hot companies. It also helped that she was able to use family connections to get a the appearance of US military approval, which she then traded up the chain into landing a marquee client in Walgreens. If you just look at the business side, which is all most investors are qualified to assess, it looked like the right things were happening at the right times, even if behind the scenes there still wasn't a product. Cargo cult investing is a good way to describe it. The dirty little secret about Silicon Valley's "meritocracy" is that while skill and a good idea is all it takes to succeed, a good network beats a good idea any day of the week and skill is optional if you're already rich. Baby Babbeh fucked around with this message at 04:22 on Sep 7, 2016 |
|
#
?
Sep 7, 2016 04:18
|
|
|
she didn't even have the loving science! hey guys I think it'd be a good idea if hunanity could travel to light speed can I please have 10 billion in seed money There that's what my pitch is now can someone please make me rich and famous like crazy eyes holmes
|
|
#
?
Sep 7, 2016 04:21
|
|
|
Do you think she had to sleep with VCs to get funding?
|
|
#
?
Sep 7, 2016 04:24
|
|
|
Arsenic Lupin posted:The one time I saw Larry Ellison walking around his company's headquarters, he had a bodyguard with, and nobody was allowed in the same elevator as him. On the other hand, Ellison is notoriously paranoid and afraid of death. Every SV rear end in a top hat thinks of themselves as a cop one day from retirement, and are certain that if they were to be killed, breakthrough brainwave scans granting clinical immortality would appear the next day.
|
|
#
?
Sep 7, 2016 04:35
|
|
|
On the other hand, Larry Ellison is such an rear end in a top hat that he's likely at the top of the list when it comes to tech CEOs that are likely to get shot by someone with a grudge. And incidentally, he was a major Theranos investor, so this all comes full circle.
|
|
#
?
Sep 7, 2016 05:11
|
|
|
Jumpingmanjim posted:Do you think she had to sleep with VCs to get funding? Definitely not. Then again, she did get to ride a bit of the "they're only criticizing her because she's a woman" train. I remember reading an article talking about how female CEOs in Silicon Valley are unfairly maligned in the press, and the three examples they gave were Elizabeth Holmes, Marissa Mayer, and loving Carley Fiorina.
|
|
#
?
Sep 7, 2016 05:15
|
|
|
Bushiz posted:Every SV rear end in a top hat thinks of themselves as a cop one day from retirement, and are certain that if they were to be killed, breakthrough brainwave scans granting clinical immortality would appear the next day. In a perfect world this would actually be true.
|
|
#
?
Sep 7, 2016 05:22
|
|
|
The fact that those are the only female tech CEOs anyone can think of is probably more telling than anything. I guess there's Meg Whitman? She got a bit of criticism when she took over at HP, but she's done a lot better job of managing her shitshow than Marissa Mayer did with hers. But that's probably because Meg Whitman has spent her entire career making lemonade out of lemons while it was Mayer's first time trying to turn something around. There's something to be said for experience. As for Fiorina, lol. If anything people are too nice to Fiorina because she's still taken seriously as an executive despite doing nothing in her career besides driving one of Silicon Valley's most storied companies into the ground like a rail spike.
|
|
#
?
Sep 7, 2016 05:29
|
|
|
Baby Babbeh posted:The fact that those are the only female tech CEOs anyone can think of is probably more telling than anything. Good point. I imagine Ginni Rometty sees all this and just shakes her head as she continues to keep IBM running like a well-oiled machine.
|
|
#
?
Sep 7, 2016 05:56
|
|
|
Jumpingmanjim posted:Do you think she had to sleep with VCs to get funding? No no, turn left!
|
|
#
?
Sep 7, 2016 06:51
|
|
|
namaste faggots posted:she didn't even have the loving science! Pick your parents carefully and you can succeed. It's the American Way!
|
|
#
?
Sep 7, 2016 07:26
|
|
|
Arsenic Lupin posted:Congratulations, you are now Elon Musk. Not knowing much about the current techie Jesus beyond Paypal, SpaceX and Tesla, why was his success dependent on his parent(s)?
|
|
#
?
Sep 7, 2016 14:54
|
|
|
Arsenic Lupin posted:Congratulations, you are now Elon Musk. Or the South African way at least.
|
|
#
?
Sep 7, 2016 14:58
|
|
|
I knew this wasn't Godfather but didn't remember Mattis at all. At least it wasn't a promoted Encino Man or one of his crew https://www.youtube.com/watch?v=mpjieIRRAp8
|
|
#
?
Sep 7, 2016 15:22
|
|
|
Lead out in cuffs posted:Yeah I'm in computational biology, but I work in and around pathology a fair bit and have a degree in biochemistry. The thing that got me was their claim that they can do 70 different tests on the same vial of blood. While I don't believe that this is necessarily impossible (eg you can test for mutations with ludicrously small amounts of material due to the miracle of PCR), I can see that each of those tests would need to be tailored individually to work at those volumes. And that means solving substantial engineering problems unique to each, and then proving accuracy and getting FDA approval through large enough sample clinical trials for each test. You would also need to accept that some of them just wouldn't be possible to do accurately, and work with what you could do. Again, this is not impossible, and with the amount of venture capital they raised, it might even have been doable. But they were claiming to do all 70 tests with that little input material, to the extent of selling the tests to Walgreens, after absolutely zero validation of any of the tests. That's obvious vapourware / criminal negligence territory. I was wrong about capillary size. There are differences in some cell counts in capillary vs veinous blood, but not enough to be clinically significant. There are some differences in capillary blood due to the increased amount of intracellular fluids, which can affect the results of coagulation testing, which is kinda important for determining what blood thinner dosages to provide. The basic materials of blood are all there, though. I'm curious about what gauge of needle they were using for testing, though, with all their talk of a "tiny" pinprick being used to take sample materials. If the needles are too small, they actually can shred blood cells, but that's a relatively simple issue to solve, and I haven't seen any article touch on it. I'd be really curious to see someone go through the 70 tests that Theranos was proposing and figure out the difficulties with performing them on a few microliters of blood. Adventure Pigeon fucked around with this message at 16:15 on Sep 7, 2016 |
|
#
?
Sep 7, 2016 16:12
|
|
|
namaste faggots posted:she didn't even have the loving science! You joke but an interstellar probe 'VC' project is an actual thing. https://breakthroughinitiatives.org/Initiative/3
|
|
#
?
Sep 7, 2016 16:50
|
|
|
Claverjoe posted:Not knowing much about the current techie Jesus beyond Paypal, SpaceX and Tesla, why was his success dependent on his parent(s)? His dad gave him the money he needed to start his first company after dropping out of college.
|
|
#
?
Sep 7, 2016 16:50
|
|
|
Hey, Adventure Pigeon, this is a bit of a tangent, but. I know biochemistry is a big ol' field, but are you familiar with the work of Brian Kobilka and Robert Lefkowitz on G-coupled protein receptors? I ask because I heard a long interview with Dr. Dahlia Weiss, who is one of the people building on their work to do structural-based drug discovery, in which the earliest stages are computational rather than using large-scale physical tests. Her work sounded extremely exciting and she was (as scientists tend to be) very hopeful about shortening and cheapening early-stage drug discovery. Do you know whether people who are not her lab are equally excited? (She's joining a startup in England that is researching a particular chemical that has shown very promising results in Parkinson's. The startups are supposed to compete with large drug companies in the discovery phase by using the technology above. See? Not a tangent.)
|
|
#
?
Sep 7, 2016 17:42
|
|
|
You really need some sort of high throughput experimental method to short list and validate the millions of false positive garbage hits.
|
|
#
?
Sep 7, 2016 18:07
|
|
|
Arsenic Lupin posted:Hey, Adventure Pigeon, this is a bit of a tangent, but. I know biochemistry is a big ol' field, but are you familiar with the work of Brian Kobilka and Robert Lefkowitz on G-coupled protein receptors? I ask because I heard a long interview with Dr. Dahlia Weiss, who is one of the people building on their work to do structural-based drug discovery, in which the earliest stages are computational rather than using large-scale physical tests. Her work sounded extremely exciting and she was (as scientists tend to be) very hopeful about shortening and cheapening early-stage drug discovery. Do you know whether people who are not her lab are equally excited? Unfortunately, I'm not. I haven't done anything with G-coupled protein receptors for about 8 or 9 years now. I'm involved in plant genomics these days, which is about as far as you can get from medical biochemistry within biology. Our big topic of excitement at the moment is CRISPR, and probably will be for a while. That being said, I do understand some of the basics. I think it's a case where they're not being dishonest, but they're overemphasizing one part of the problem compared to all the others. I can see how narrowing down the list of small molecules that need to be tested against a particular target before one is found that works could speed up drug discovery. Of course, if you have a target receptor, and you have an idea of what molecule it binds in nature, then creating artificial small molecules that will probably target it based on the structure of the native signal molecule isn't exactly a blind process to begin with. Her method is probably a good way to come up with drugs when the target is already known, but that means most drugs will probably be improvements on existing treatments rather than entirely new treatments. That kinda brings me to what I understand the primary challenge is, which is finding the targets in the first place. Finding targets is mixture of brute force sequencing of tens of thousands of individuals with detailed medical histories combined with a lot of statistics, and finally some molecular biology to confirm the target. Regeneron, which is probably the hottest research oriented pharma company these days, is dumping enormous amounts of money into that, and has been for a long time, and they're not the only ones. 23andMe's real business model isn't sequencing people and providing that data back to them, it's sequencing people on their dime, combining that with health information they willingly provide, putting all that sequence and phenotypic data into a big database, and then selling database access to pharma companies for tens of millions of dollars.
|
|
#
?
Sep 7, 2016 18:44
|
|
|
poemdexter posted:His dad gave him the money he needed to start his first company after dropping out of college. Admittedly it was only $28,000, which while not insubstantial kinda pales compared to what Holmes's parent's friends gave (or the small loan Donnie got).
|
|
#
?
Sep 7, 2016 18:52
|
|
|
https://en.wikipedia.org/wiki/Elon_Musk#Career posted:Zip2 You can kind of figure out things went from there
|
|
#
?
Sep 7, 2016 21:46
|
|
|
nm posted:Admittedly it was only $28,000, which while not insubstantial kinda pales compared to what Holmes's parent's friends gave (or the small loan Donnie got).
|
|
#
?
Sep 7, 2016 21:51
|
|
|
$28K is pretty standard for a seed round. The fact that Musk was able to get it all from his dad was a plus, but it's not super uncommon for startups either. There's a reason seed used to be called the "friends and family" round. In fact, the existence of a class of investors
|
|
#
?
Sep 8, 2016 01:24
|
|
|
Seed is usually more than $28K. That's more typically an angel round, or as you say friends and family. There is certainly still an element of privilege to having a circle that can produce $30K to get started, but it's not some sultan-of-Brunei rarity. That's well less than a year at most US colleges, f.e. E: most major cities have events at which you can pitch for angel money, these days
|
|
#
?
Sep 8, 2016 03:38
|
|
|
Looking into it, Musk's family is definitely wealthy; albeit not on the level of wealth and connections necessary to pull names like Kissinger: https://www.quora.com/Who-is-Elon-Musks-father His background was definitely a huge factor in his getting a chance to succeed. On the other hand, he was involved in actually producing a genuinely successful product that most people have probably used (PayPal) rather than having billions thrown at him for having a half-baked dream. Hodgepodge fucked around with this message at 03:49 on Sep 8, 2016 |
|
#
?
Sep 8, 2016 03:44
|
|
|
If you're in SF, $30K won't get you very far. Even if you are a software startup that can work out of your apartment, you still need an apartment without distracting roommates. Living expenses will eat through that $30k in a few months easily.
|
|
#
?
Sep 8, 2016 03:58
|
|
|
Sure, and many angel rounds are (much) larger. They're not meant to carry for a year though, just to get things started enough for seed-round pitches.
|
|
#
?
Sep 8, 2016 04:03
|
|
|
|
| # ? Apr 29, 2024 01:50 |
|
|
Arsenic Lupin posted:The big problem is that AFAIK no current cloud-based map is 100% reliable. Those automated trucks are going to be driving down roads that are too small for trucks (a big problem in England, a problem in residential neighborhoods everywhere), driving under bridges that are too short for the truck, driving into ponds... Blindly trusting GPS is a significant problem for human drivers, especially human truck drivers who have never driven a route before. It's going to be more of a disaster for trucks, which can't recognize signage saying they're in the wrong place. This is old, but you could always just have a single truck-driver make the trip to ensure the route is wide enough and tall enough for a truck.
|
|
#
?
Sep 8, 2016 04:38
|
|
