WEBVTT - How Two AI Startup Unicorns Went Bust 0:00:04.480 --> 0:00:12.440 Welcome to tech Stuff, a production from iHeartRadio. Hey there, 0:00:12.480 --> 0:00:16.400 and welcome to tech Stuff. I'm your host Jonathan Strickland. 0:00:16.440 --> 0:00:19.919 I'm an executive producer with iHeart Podcasts and how the 0:00:20.000 --> 0:00:22.959 tech are you? You know. A couple of years ago, 0:00:23.520 --> 0:00:28.040 it seemed like AI had just exploded into the mainstream 0:00:28.520 --> 0:00:32.120 out of nowhere. Now, in reality, thousands of people have 0:00:32.200 --> 0:00:36.839 been working for decades to bring various AI implementations to 0:00:36.880 --> 0:00:40.160 a point where they could be deployed in the real world, 0:00:40.440 --> 0:00:43.279 not just you know, R and D projects. I think 0:00:43.320 --> 0:00:47.080 a lot of us are hyper aware of generative AI 0:00:47.159 --> 0:00:50.559 in particular. That's the kind of application that's easy for 0:00:50.600 --> 0:00:54.080 the average person to have experienced and perceived, you know, 0:00:54.200 --> 0:00:57.920 first hand. But there's a lot more going on in 0:00:58.040 --> 0:01:01.960 AI than just chat by and image generators, that kind 0:01:02.000 --> 0:01:07.040 of thing. There's widespread agreement that AI is definitely going 0:01:07.080 --> 0:01:12.000 to have an enormous impact on pretty much everything. There's 0:01:12.080 --> 0:01:16.479 less agreement about how this transformation is going to manifest 0:01:16.800 --> 0:01:20.400 or in what time frame it's going to happen. We've 0:01:20.400 --> 0:01:24.000 got a lot of companies that have made some fairly 0:01:24.520 --> 0:01:28.880 aggressive moves into the space, either developing AI or trying 0:01:28.920 --> 0:01:33.920 to implement AI solutions, and it hasn't always worked out. 0:01:34.400 --> 0:01:38.880 In the meanwhile, you've got this incredible environment in which entrepreneurs, 0:01:39.200 --> 0:01:43.520 computer scientists, venture capital investors, and more are all trying 0:01:43.560 --> 0:01:49.480 to leverage the moment. Like that whole idea of opportunity knocking. 0:01:49.520 --> 0:01:52.360 You've got to seize on that opportunity because who knows 0:01:52.400 --> 0:01:55.120 when it's going to come around again, Even if they're 0:01:55.160 --> 0:02:01.480 not actually prepared to execute upon that opportunity. Sometimes leveraging 0:02:01.520 --> 0:02:05.040 opportunity just doesn't go the way you want it to. Now, 0:02:05.080 --> 0:02:08.880 I've already published tech Stuff episodes in the past about 0:02:08.960 --> 0:02:13.280 failed tech startups. Some of those startups reached really high 0:02:13.360 --> 0:02:18.320 levels of valuation, like topping a billion dollars, which catapults 0:02:18.360 --> 0:02:22.680 them into the so called unicorn status. A unicorn is 0:02:22.760 --> 0:02:25.480 a startup doesn't have to be tech. It's a startup 0:02:25.520 --> 0:02:29.320 business that hits evaluation of a billion dollars at some 0:02:29.480 --> 0:02:34.359 point or another. But even with that kind of high valuation, 0:02:34.800 --> 0:02:38.760 and even with the initial excitement and support of investors 0:02:38.760 --> 0:02:43.800 who have really deep pockets, startups sometimes collapse for one 0:02:43.880 --> 0:02:46.880 reason or another. So today I thought we should talk 0:02:46.919 --> 0:02:51.680 about a couple of digital health AI startups that have 0:02:52.040 --> 0:02:57.280 followed this particular path. While AI enthusiasm has built to 0:02:57.360 --> 0:03:00.600 what I would call a frenzy, in more recent months 0:03:00.880 --> 0:03:05.720 that enthusiasm has waned somewhat. Investors have kind of backed 0:03:05.760 --> 0:03:08.120 off a little bit from AI. There have been a 0:03:08.200 --> 0:03:11.480 lot of articles that have been published that say any 0:03:11.639 --> 0:03:15.520 real gains from AI are probably years down the road, 0:03:15.880 --> 0:03:19.680 at least for most implementations. This is not necessarily universal 0:03:19.720 --> 0:03:22.280 for all AI, which is an important thing to remember. 0:03:22.360 --> 0:03:27.519 Not all AI is the same, but for many implementations 0:03:27.800 --> 0:03:32.240 it's just too early to think of AI making an 0:03:32.400 --> 0:03:38.240 enormous impact on business objectives. And so again, investments have 0:03:38.360 --> 0:03:41.400 started to taper off quite a bit. People are being 0:03:41.520 --> 0:03:46.160 more particular with their investments for lots of reasons, but 0:03:46.240 --> 0:03:48.680 a big one of those is the perception that AI 0:03:48.840 --> 0:03:52.040 is not really ready for prime time to totally transform 0:03:52.080 --> 0:03:56.960 everything right now. So a lot of AI companies or 0:03:57.000 --> 0:04:00.600 companies that have, you know, kind of leveraged AI to 0:04:01.320 --> 0:04:04.040 be their sales pitch have kind of found themselves in 0:04:04.120 --> 0:04:08.200 dire straits. The condition of dire straits, not the band. 0:04:09.080 --> 0:04:12.840 Dire straits, not the sultans of swing. Now, before I 0:04:12.880 --> 0:04:18.080 get to specifics, let's just establish some general factors, all right. 0:04:18.200 --> 0:04:23.520 So artificial intelligence is an expensive discipline. It is a 0:04:23.720 --> 0:04:29.279 resource hungry computer application. So you have to either own 0:04:29.640 --> 0:04:33.920 or have access to really powerful data centers. For any 0:04:34.080 --> 0:04:38.200 AI application that's going beyond just a proof of concept. 0:04:38.240 --> 0:04:41.479 If you plan on launching something that is meant to 0:04:41.480 --> 0:04:44.960 be a product or service that is in part or 0:04:45.080 --> 0:04:48.839 wholly dependent upon AI, you have to have access to 0:04:48.880 --> 0:04:52.560 that compute power. Now, on top of that, AI applications 0:04:52.600 --> 0:04:58.679 typically require specific types of powerful parallel processing capabilities, which 0:04:58.720 --> 0:05:02.000 means that you need a particular kind of computer chip. 0:05:02.080 --> 0:05:04.640 You can't just get any computer chip, even a really 0:05:04.640 --> 0:05:07.760 fast one. You need one that's really good at handling 0:05:07.800 --> 0:05:11.159 parallel processing. So, like in the early days of AI, 0:05:11.320 --> 0:05:15.599 GPUs were a really big part of AI processing because 0:05:15.640 --> 0:05:21.360 GPUs graphics processing units typically are designed to be parallel processors. 0:05:21.360 --> 0:05:24.239 They have multiple cores, and they can do multi threading 0:05:24.240 --> 0:05:27.239 and work on a lot of different problems all simultaneously. 0:05:27.520 --> 0:05:29.919 AI needs that capability, or at least a lot of 0:05:30.000 --> 0:05:35.000 machine learning processes require that kind of computational power. So 0:05:35.080 --> 0:05:38.280 the chips best suited to do that are not always 0:05:38.440 --> 0:05:42.039 in plentiful supply. So that means we've got this big 0:05:42.160 --> 0:05:46.000 pool of AI companies. Some of them are part of 0:05:46.160 --> 0:05:51.000 much larger organizations like Microsoft or Google, but all of 0:05:51.040 --> 0:05:54.479 them are competing for a limited supply of processors that 0:05:54.560 --> 0:05:59.240 are suitable for handling AI computational loads. Not everyone is 0:05:59.279 --> 0:06:01.520 going to come out of winner in that kind of competition. 0:06:01.960 --> 0:06:05.920 You know, big companies obviously have a huge advantage over 0:06:06.120 --> 0:06:10.480 smaller startups that are dependent upon rounds of fundraising from investors. 0:06:10.520 --> 0:06:13.920 They're going to venture capitalists to get a influx of 0:06:14.000 --> 0:06:16.640 cash in order to be able to do business. Meanwhile, 0:06:16.640 --> 0:06:20.560 you have these monoliths like Microsoft and Google that have 0:06:21.279 --> 0:06:26.160 decades of wealth that have been generated and they lean 0:06:26.200 --> 0:06:28.880 on that in order to get an advantage in the market. 0:06:29.279 --> 0:06:33.800 So that's one thing. Another is that scaling up is 0:06:33.960 --> 0:06:38.479 particularly challenging for AI companies. This is hard for any startup. Right, 0:06:38.680 --> 0:06:42.400 a startup can come up with a brilliant idea and 0:06:42.880 --> 0:06:45.839 have an idea that truly has a good place in 0:06:45.880 --> 0:06:48.360 the market. Once you get to a level of scale 0:06:48.560 --> 0:06:52.880 where you can make this a revenue generating business, but 0:06:53.000 --> 0:06:56.200 getting there is hard. Right, if it's manufacturing, then you 0:06:56.240 --> 0:06:57.880 have to figure out, well, how are we going to 0:06:58.480 --> 0:07:02.760 afford to manufac facture in the bulk we need in 0:07:02.839 --> 0:07:05.760 order to make this a viable business for services, How 0:07:05.760 --> 0:07:09.400 do we make sure that our business can provide the 0:07:09.480 --> 0:07:13.200 services to the customer base we're going to need in 0:07:13.320 --> 0:07:17.520 order to have a viable business. These are non trivial challenges. 0:07:18.000 --> 0:07:20.440 You have to figure out how do you actually meet 0:07:20.480 --> 0:07:24.320 the demand that you're hoping to create. Now, first, there 0:07:24.360 --> 0:07:25.920 has to be a demand there in the first place. 0:07:25.960 --> 0:07:28.280 And if there's no existing demand, you have to create 0:07:28.280 --> 0:07:30.480 that demand. Then you have to be able to deliver 0:07:31.360 --> 0:07:35.800 value for that demand. These are really hard problems, so 0:07:35.880 --> 0:07:38.160 a lot of startups do fail, like whether they're in 0:07:38.200 --> 0:07:41.880 the tech sector or not, and for AI companies it 0:07:41.960 --> 0:07:46.640 is particularly difficult. Scientists might develop a truly intriguing use 0:07:46.680 --> 0:07:50.280 for artificial intelligence. The work's great on a small scale, 0:07:50.600 --> 0:07:53.360 like yeah, they can prove it works for a small 0:07:54.000 --> 0:07:58.200 test group or maybe a region or a very specific industry. 0:07:58.400 --> 0:08:00.920 But then to grow that so that you can meet 0:08:00.960 --> 0:08:04.120 the needs of customers around the world that could require 0:08:04.320 --> 0:08:08.120 way more resources than you can actually afford even as 0:08:08.120 --> 0:08:12.120 a unicorn. So AI startups can end up burning through 0:08:12.240 --> 0:08:17.520 cash really quickly, not through terrible mismanagement, though of course 0:08:17.600 --> 0:08:22.160 that can happen too, but just because HEYI is expensive. 0:08:22.520 --> 0:08:24.360 So I wanted to clear that up at the start 0:08:24.400 --> 0:08:26.760 because I don't want to get the impression that the 0:08:26.800 --> 0:08:30.679 folks that are behind these businesses necessarily did something wrong. 0:08:30.840 --> 0:08:34.040 Although in one case we'll see that there is at 0:08:34.120 --> 0:08:38.040 least one news outlet that very much feels like the 0:08:38.160 --> 0:08:41.200 founder of a company did many things wrong. I don't 0:08:41.200 --> 0:08:44.040 want to say that they were necessarily bad at managing money. 0:08:44.640 --> 0:08:47.080 That could be a factor as well, but I'm trying 0:08:47.080 --> 0:08:51.280 to separate this from the dot com bubble of the 0:08:51.360 --> 0:08:55.160 late nineties. So with the dot com bubble, you had 0:08:55.200 --> 0:09:00.280 all these startups that got enormous investments in cash, and 0:09:00.360 --> 0:09:02.400 you know, some of them went public on the stock 0:09:02.440 --> 0:09:06.160 market and their stocks inflated to ridiculous levels. And you 0:09:06.240 --> 0:09:10.400 had these companies that didn't have fully baked business plans 0:09:10.440 --> 0:09:15.000 in place. They were absolutely swimming in cash, and a 0:09:15.000 --> 0:09:19.640 lot of times people were making extravagant purchases like crazy 0:09:19.760 --> 0:09:23.920 office things like like a full bar or whatever, without 0:09:24.080 --> 0:09:27.720 actually being able to put those assets to work to 0:09:27.840 --> 0:09:30.560 create a business that could stand on its own. And 0:09:30.679 --> 0:09:36.400 ultimately the bubble burst and the entire industry collapsed. You know, 0:09:36.520 --> 0:09:39.760 some companies survived, but a lot of them didn't. Well, 0:09:39.760 --> 0:09:43.160 I want to draw a line between those and what 0:09:43.400 --> 0:09:46.360 might be an AI tech bubble. I think it's fair 0:09:46.400 --> 0:09:51.160 to call it an AI tech bubble because one almost 0:09:51.240 --> 0:09:56.000 universal issue for all the AI startups is this challenge 0:09:56.000 --> 0:10:01.800 that artificial intelligence is inherently expensive. They could also fall 0:10:01.920 --> 0:10:05.240 victim to the same problems that we solve with dot 0:10:05.280 --> 0:10:08.719 com businesses. That's still a possibility. I'm not saying that 0:10:08.960 --> 0:10:13.559 AI companies are somehow immune to human frailties of going 0:10:13.640 --> 0:10:17.200 overboard and like everybody gets a new car or whatever 0:10:17.240 --> 0:10:21.080 it might be, but that the very nature of AI 0:10:21.120 --> 0:10:25.520 itself becomes a massive risk as far as seeing a 0:10:25.559 --> 0:10:30.560 return on investment. So with all that set, we're ready 0:10:30.600 --> 0:10:34.840 to start diving into discussions of a pair of different 0:10:35.040 --> 0:10:40.520 digital health companies that were largely centered around the idea 0:10:40.640 --> 0:10:46.239 of artificial intelligence revolutionizing the way we do certain things 0:10:46.240 --> 0:10:50.680 in healthcare. Whether or not artificial intelligence was actually playing 0:10:50.720 --> 0:10:53.440 a part in that, that's more of an open question. 0:10:53.679 --> 0:10:55.960 We're going to get into that in just a moment. 0:10:56.080 --> 0:10:58.840 Before we do that, let's take a quick break to 0:10:58.920 --> 0:11:11.560 thank our sponsors. Okay, we're back. Let's talk about our 0:11:11.679 --> 0:11:17.160 first digital health company and what happened. And in these cases, 0:11:17.200 --> 0:11:19.439 I think it's safe to say that the AI component 0:11:19.679 --> 0:11:23.520 was really just one contributing factor to how these two 0:11:24.240 --> 0:11:27.920 health company startups failed over time. I think it was 0:11:27.960 --> 0:11:32.079 a major contributing factor, but just one of multiple factors. 0:11:32.320 --> 0:11:35.359 But another is that there was an understandable but arguably 0:11:35.480 --> 0:11:41.120 foolish rush of cash influence in the health space following 0:11:41.160 --> 0:11:45.440 the twenty twenty COVID outbreak. Now that rush of cash 0:11:45.640 --> 0:11:50.000 was understandable because obviously the pandemic had an enormous impact 0:11:50.040 --> 0:11:54.040 on people all around the world. There were tons of regions, 0:11:54.280 --> 0:11:59.720 entire countries that were operating under lockdown conditions, sometimes for 0:11:59.800 --> 0:12:03.600 so several months at a stretch. Now, obviously that would 0:12:03.679 --> 0:12:07.160 change how we do pretty much everything from how we 0:12:07.240 --> 0:12:12.640 work to how students were attending lessons, to how you 0:12:12.920 --> 0:12:15.800 actually got to see a physician if you needed one, 0:12:16.120 --> 0:12:19.880 and so out of necessity, health companies new and old 0:12:20.120 --> 0:12:23.880 attempted to adapt to this new reality. Now, it could 0:12:23.920 --> 0:12:28.120 be really hard for large established companies to adapt to 0:12:28.280 --> 0:12:33.040 rapidly changing conditions. Being nimble isn't exactly a common trait 0:12:33.280 --> 0:12:38.640 for legacy organizations. That opened up opportunities for younger startups 0:12:38.640 --> 0:12:41.959 to innovate in the space and to attempt to serve 0:12:42.080 --> 0:12:47.320 customers in ways that larger organizations simply couldn't replicate. And 0:12:47.440 --> 0:12:53.240 one of those companies was called Babylon. Now, Babylon was 0:12:53.320 --> 0:12:56.840 founded years before the pandemic. It was founded way back 0:12:56.880 --> 0:13:00.800 in twenty thirteen. Technically it's one year year younger than 0:13:00.840 --> 0:13:03.679 the other digital health company we'll talk about in a 0:13:03.679 --> 0:13:08.439 little bit. Babylon launched in the United Kingdom. Ultimately it 0:13:08.440 --> 0:13:12.040 would extend services to other parts of the world, primarily 0:13:12.160 --> 0:13:15.000 in Asia, but also the United States and a couple 0:13:15.040 --> 0:13:19.319 of places in Africa. It was a subscription based healthcare 0:13:19.400 --> 0:13:23.120 services company, and initially it was one that had a 0:13:23.160 --> 0:13:27.640 fairly simple approach. Customers or patients in other words, would 0:13:27.640 --> 0:13:33.000 communicate with healthcare professionals via text messages and video conferences. 0:13:33.240 --> 0:13:38.520 So it was a telehealth solutions company, which again not 0:13:38.720 --> 0:13:43.160 that groundbreaking, right, there were other telehealth solutions out there, 0:13:43.559 --> 0:13:46.800 and at this stage there was no real AI component. 0:13:47.240 --> 0:13:51.559 This was all let's put patients in touch with doctors 0:13:52.000 --> 0:13:54.200 and do it in a way where the patient doesn't 0:13:54.240 --> 0:13:56.440 necessarily have to take time out of his or her 0:13:56.640 --> 0:14:00.880 or their day to go and meet with physician. They 0:14:00.880 --> 0:14:04.319 could do it through this app. So you might say, well, 0:14:04.320 --> 0:14:07.240 where is the AI component if this episode is about 0:14:07.320 --> 0:14:10.880 AI health startups. Well, that took the form of a 0:14:11.040 --> 0:14:14.600 chat bot that was developed later on in Babylon. It 0:14:14.640 --> 0:14:18.040 was an idea that I think was present from the beginning, 0:14:18.120 --> 0:14:21.080 Like this was a goal early on was to develop 0:14:21.160 --> 0:14:24.520 a chat bot that would be able to interact with patients, 0:14:24.960 --> 0:14:27.680 and the chatbot would be able to answer questions, and 0:14:27.880 --> 0:14:31.440 ultimately Babylon claimed it would be able to do things 0:14:31.480 --> 0:14:35.960 like diagnose patients. So not just like answer questions about 0:14:36.120 --> 0:14:41.080 physician availability or simple questions that might lead to a 0:14:41.120 --> 0:14:45.440 way to alleviate symptoms that are acutely bothering a patient, 0:14:45.640 --> 0:14:50.400 but actually diagnosing the underlying cause of those symptoms. That 0:14:50.520 --> 0:14:54.480 is a huge claim, I mean as a remarkable claim, 0:14:54.520 --> 0:14:58.360 and it requires remarkable evidence to support it. When you're 0:14:58.400 --> 0:15:02.240 talking about healthcare, there's a high bar you need to meet, 0:15:02.600 --> 0:15:05.960 a high bar of confidence. If you do not meet that, 0:15:06.440 --> 0:15:09.920 then that means you probably should not not Probably you 0:15:10.000 --> 0:15:13.760 should not offer these services to patients who were talking 0:15:13.920 --> 0:15:19.200 literally matters of life and death. So understandably, a lot 0:15:19.240 --> 0:15:22.840 of critics were raising concerns about how reliable this chatbot 0:15:23.040 --> 0:15:26.520 actually was and whether or not it was ethical to 0:15:26.640 --> 0:15:32.480 even suggest that a chatbot could accurately diagnose someone's ailments 0:15:32.560 --> 0:15:35.480 through interacting with that patient. And again, all of this 0:15:35.640 --> 0:15:39.720 was happening before the pandemic and well before Open Ai 0:15:39.880 --> 0:15:43.760 really opened the floodgates with chat GPT in twenty twenty two, 0:15:44.280 --> 0:15:49.520 So a very aggressive approach toward positioning AI as a 0:15:49.560 --> 0:15:53.440 solution to a complicated problem. Now, in the past, I've 0:15:53.480 --> 0:15:58.040 talked about how AI generated answers can sometimes be incomplete 0:15:58.360 --> 0:16:02.680 or unreliable. That's you using today's AI chatbots, which are 0:16:03.320 --> 0:16:06.040 miles ahead of the stuff that was being developed in 0:16:06.120 --> 0:16:08.720 the twenty tens. So imagine putting your life in the 0:16:08.800 --> 0:16:15.680 virtual hands of a fallible chatbot back in twenty eighteen. Wolf. However, 0:16:16.000 --> 0:16:19.680 according to a great piece that appeared in Wired, it 0:16:19.720 --> 0:16:23.360 was written by Grace Brown. It's titled The Fall of 0:16:23.440 --> 0:16:27.680 Babylon is a Warning for AI Unicorns. Well, according to 0:16:27.680 --> 0:16:31.640 that piece, the AI bit of Babylon might have been 0:16:31.640 --> 0:16:33.520 a stretch in the first place, it might have been 0:16:33.520 --> 0:16:38.640 disingenuous to reference this as artificial intelligence. So Brown cites 0:16:38.680 --> 0:16:41.840 a consulting doctor by the name of Hugh Harvey, who 0:16:41.920 --> 0:16:45.560 at one time worked for Babylon. Now. According to Harvey, 0:16:45.840 --> 0:16:49.600 the AI decision tree that it would follow when interacting 0:16:49.640 --> 0:16:53.800 with patients was essentially an Excel spreadsheet that correlated to 0:16:53.880 --> 0:16:57.320 different parts of the human body. So a patient using 0:16:57.440 --> 0:17:01.480 Babylon's app would indicate, you know, where their symptoms were 0:17:01.520 --> 0:17:04.600 affecting them, like oh, my leg is itching or something 0:17:04.640 --> 0:17:07.960 like that. The app would essentially hone in on possible 0:17:08.000 --> 0:17:13.440 diagnoses by eliminating all the stuff that wasn't a potential 0:17:13.520 --> 0:17:17.360 candidate for an explanation, which is not a very sophisticated 0:17:17.480 --> 0:17:22.520 method of determining what the underlying cause is. And as 0:17:22.640 --> 0:17:27.240 Harvey told Brown, quote, I was like, well, this isn't 0:17:27.359 --> 0:17:31.520 really artificial intelligence, end quote. But whether we should classify 0:17:31.560 --> 0:17:35.159 the inner workings of Babylon as AI or not AI 0:17:35.400 --> 0:17:39.280 was definitely part of the company's messaging to investors. So 0:17:39.680 --> 0:17:42.719 you could say, well, this isn't really artificial intelligence. This 0:17:42.800 --> 0:17:46.639 is a very simplistic decision tree. There's no artificial intelligence 0:17:46.680 --> 0:17:49.919 going on here. There's no decision making, but that's not 0:17:50.000 --> 0:17:55.000 how the company was marketing their capabilities to potential customers. 0:17:55.359 --> 0:17:59.040 Babylon was saying, we use artificial intelligence to help treat 0:17:59.119 --> 0:18:02.280 patients to die noose and treat them. So, whether you 0:18:02.320 --> 0:18:04.760 want to argue that AI was happening or not, the 0:18:04.800 --> 0:18:08.080 company certainly was claiming that to be the case. And 0:18:08.160 --> 0:18:11.560 Babylon initially did pretty well when it came to raising investments. 0:18:11.600 --> 0:18:15.520 So by twenty nineteen, before the pandemic, Babylon had raised 0:18:15.520 --> 0:18:19.119 more than half a billion dollars in funding over the years. 0:18:19.359 --> 0:18:23.000 So remember it was founded in twenty thirteen. By twenty nineteen, 0:18:23.200 --> 0:18:26.320 more than half a billion dollars in various investment rounds. 0:18:26.640 --> 0:18:29.679 In twenty twenty one, the company played the risky maneuver 0:18:29.800 --> 0:18:32.920 of going public through the use of a special purpose 0:18:33.000 --> 0:18:39.080 Acquisition company or SPAC SPAC aka a blank check company. 0:18:39.359 --> 0:18:42.879 Now I have talked about spacks before, but let's have 0:18:42.960 --> 0:18:47.760 a quick refresher. Typically, when a private company is preparing 0:18:47.760 --> 0:18:52.440 to transition into a publicly traded company where the average 0:18:52.440 --> 0:18:56.240 citizen can buy stock in the company, it first has 0:18:56.280 --> 0:18:59.240 to go through an extensive set of steps in order 0:18:59.359 --> 0:19:03.159 to get to the IPO or initial public offering. This 0:19:03.280 --> 0:19:06.440 involves a ton of scrutiny from regulators. Here in the 0:19:06.560 --> 0:19:10.840 United States, it's the Securities and Exchange Commission, or SEC. 0:19:11.400 --> 0:19:16.720 As Kate Ashford wrote in Forbes Advisor quote, going public 0:19:17.119 --> 0:19:21.240 is a challenging, time consuming process that's difficult for most 0:19:21.240 --> 0:19:25.400 companies to navigate alone. A private company planning an IPO 0:19:25.600 --> 0:19:29.280 needs not only to prepare itself for an exponential increase 0:19:29.280 --> 0:19:32.280 in public scrutiny, but it also has to file a 0:19:32.440 --> 0:19:36.640 ton of paperwork and financial disclosures to meet the requirements 0:19:36.680 --> 0:19:41.040 of the Securities and Exchange Commission SEC, which oversees public 0:19:41.200 --> 0:19:44.639 companies end quote. So if a startup is looking to 0:19:44.640 --> 0:19:47.760 get access to a ton of cash through going public 0:19:47.880 --> 0:19:51.439 and it's a bit strapped for time, an alternative to 0:19:51.600 --> 0:19:56.480 the IPO is the SPAC. So with a SPACK you 0:19:56.600 --> 0:20:00.199 have a holding company. So this company doesn't really make 0:20:00.359 --> 0:20:04.679 or do anything. It's kind of like an empty envelope. 0:20:04.960 --> 0:20:07.120 So the one thing it can do is it can 0:20:07.160 --> 0:20:10.520 go through all the regulatory processes required to go public 0:20:10.640 --> 0:20:13.520 and become a publicly traded company. So now you've got 0:20:13.560 --> 0:20:16.520 a publicly traded company that doesn't actually do anything else. 0:20:16.840 --> 0:20:20.320 So using this empty shell of a company, you then 0:20:20.480 --> 0:20:24.920 can acquire a private company that's just itching to go public, 0:20:25.240 --> 0:20:28.000 but it doesn't have the time or the ability to 0:20:28.040 --> 0:20:30.840 do this through the IPO method, or if they did 0:20:30.880 --> 0:20:34.679 do an IPO, the value of stock that would be 0:20:34.680 --> 0:20:37.040 determined through that process would be much lower than what 0:20:37.080 --> 0:20:40.600 they actually want it to be. So your SPAC, your 0:20:40.640 --> 0:20:46.640 SPAC acquires this private startup. Now through the transitive property 0:20:46.640 --> 0:20:50.119 of ownership, that startup is a publicly traded company, or 0:20:50.119 --> 0:20:53.399 at least it's part of a publicly traded company, And 0:20:53.440 --> 0:20:55.600 it's like the startup got a chance to skip all 0:20:55.600 --> 0:20:58.440 that boring paperwork and get straight to the part where 0:20:58.440 --> 0:21:02.120 people throw money at it. However, if it turns out 0:21:02.160 --> 0:21:05.119 the startup doesn't have the ability to succeed in the 0:21:05.160 --> 0:21:08.960 public marketplace, while all of this can then come crashing down. 0:21:09.080 --> 0:21:12.480 Shareholders can lose confidence in the company, They can sell 0:21:12.480 --> 0:21:16.320 off their shares, share prices can fall, that big old 0:21:16.400 --> 0:21:19.120 pile of money can start to shrink, and it's almost 0:21:19.160 --> 0:21:21.879 like skipping all those steps that are intended to make 0:21:21.920 --> 0:21:24.600 sure that companies can make the transition from private to 0:21:24.640 --> 0:21:28.080 public in a sustainable way might be a bad idea. 0:21:28.480 --> 0:21:31.320 That's what happened with Babylon, at least to some extent. 0:21:31.640 --> 0:21:36.640 A SPAC called al Kourie Global Acquisition Corporation acquired Babylon 0:21:36.760 --> 0:21:40.040 in October twenty twenty one. The SPAC, in turn had 0:21:40.080 --> 0:21:45.160 the backing of palanteer Is, Peter Thiel's big data analytics company. 0:21:45.480 --> 0:21:48.600 It was one of many investors that were part of this. 0:21:49.080 --> 0:21:54.439 Babylon's valuation was estimated at four point two billion with 0:21:54.560 --> 0:21:58.800 a B dollars, and it is wild to think that 0:21:59.080 --> 0:22:02.879 just two years later Babylon would get sold off for 0:22:03.040 --> 0:22:06.800 parts as the value of the company had totally collapsed. 0:22:06.920 --> 0:22:10.880 And by collapsed, I mean that eighteen months after being 0:22:10.920 --> 0:22:13.919 listed in the New York Stock Exchange, the stock price 0:22:14.119 --> 0:22:17.879 was ninety nine percent lower than where it had started off. 0:22:18.400 --> 0:22:21.640 How did that happen? I'll explain more, but first let's 0:22:21.640 --> 0:22:35.320 take another quick break. Okay, we have Babylon, a startup 0:22:35.359 --> 0:22:39.920 from twenty thirteen that reaches incredible heights through this reverse 0:22:40.000 --> 0:22:44.720 merger process with a SPACK and is worth or value 0:22:44.800 --> 0:22:47.960 that I should say, four point two billion dollars. Well, 0:22:47.960 --> 0:22:51.720 according to Brown's Piece and Wired, Babylon was actually already 0:22:51.760 --> 0:22:54.520 in trouble. By the time it joined the stock exchange, 0:22:54.680 --> 0:22:58.280 the company was running through cash very quickly in an 0:22:58.359 --> 0:23:02.199 attempt to scale the business, to grow it beyond what 0:23:02.320 --> 0:23:05.119 it already was. Now. As I mentioned at the top, 0:23:05.320 --> 0:23:09.560 AI businesses in particular are really costly to scale, So 0:23:09.720 --> 0:23:12.679 unless you've got really deep pockets, like the pockets of 0:23:12.680 --> 0:23:16.479 one of the big five tech companies Microsoft, Google, Meta, Apple, 0:23:16.600 --> 0:23:20.359 or Amazon, well, you'll likely find challenges in making your 0:23:20.400 --> 0:23:23.560 money last long enough for you to scale properly and 0:23:24.000 --> 0:23:28.000 be self sufficient. Babylon was spending way more money than 0:23:28.040 --> 0:23:30.560 it was bringing in, so it was losing money year 0:23:30.600 --> 0:23:33.960 over year, and as a publicly traded company, Babylon had 0:23:34.000 --> 0:23:36.560 to share this information with the sec. You know, if 0:23:36.600 --> 0:23:39.280 you're a privately held company, you don't have to talk 0:23:39.320 --> 0:23:43.200 about how much money you lost. The public remains uninformed, 0:23:43.240 --> 0:23:47.679 But with publicly traded ones, that information gets filed and 0:23:47.720 --> 0:23:51.000 it becomes available to the public, so you can see 0:23:51.040 --> 0:23:55.440 how much money the company is losing year over year. Well, clearly, 0:23:55.560 --> 0:23:59.600 shareholders lost confidence. The stock price crashed, and just a 0:23:59.600 --> 0:24:03.000 couple of years after having going public with that SPAC transaction, 0:24:03.200 --> 0:24:07.480 Babylon went into administration. In the UK and bankruptcy in 0:24:07.520 --> 0:24:10.480 the US. So administration in the UK is kind of 0:24:10.520 --> 0:24:13.800 similar to bankruptcy here in the United States. They're not identical, 0:24:14.040 --> 0:24:18.680 but they are similar processes. It's meant to try and 0:24:19.040 --> 0:24:22.240 return as much value to investors as possible while a 0:24:22.560 --> 0:24:27.800 business effectively shuts down. So Brown's piece gives more details 0:24:27.960 --> 0:24:32.159 about what was actually going on within Babylon, but in general, 0:24:32.400 --> 0:24:34.679 it was a case of a company spending money it 0:24:34.760 --> 0:24:37.560 had not yet raised in the hopes of hitting that 0:24:37.680 --> 0:24:41.359 sweet spot and delivering upon the company's value proposition. The 0:24:41.400 --> 0:24:44.320 stories of Babylon sound kind of similar to what I 0:24:44.440 --> 0:24:48.600 heard about Aharranos Now Farrannose was that infamous high tech 0:24:49.000 --> 0:24:53.639 health company that absolutely imploded after an expose revealed that 0:24:53.680 --> 0:24:57.080 the company's flagship product, a device that was meant to 0:24:57.400 --> 0:25:00.960 analyze a tiny micro drop of blood and potentially run 0:25:01.160 --> 0:25:04.240 hundreds of different medical tests on It turned out that 0:25:04.280 --> 0:25:07.760 product just did not work as advertised, and in fact, 0:25:08.280 --> 0:25:10.639 it might not ever work at all, at least not 0:25:10.800 --> 0:25:13.600 to the extent that was being promised by the company. 0:25:14.040 --> 0:25:15.960 No matter how much effort was put into it, it 0:25:16.000 --> 0:25:19.159 was going to run up against some fundamental limitations that 0:25:19.280 --> 0:25:23.160 meant it just could not work the way it was envisioned, 0:25:23.359 --> 0:25:25.840 and that the whole company was essentially a house of 0:25:25.880 --> 0:25:29.480 cards built upon this belief that ultimately tech can do 0:25:29.600 --> 0:25:32.520 anything if you just work at it hard enough, and 0:25:32.880 --> 0:25:35.800 it turned out that just wasn't true. Well, it sounds 0:25:35.800 --> 0:25:40.080 like Babylon suffered a kind of similar fate. Now check 0:25:40.080 --> 0:25:42.480 out Grace Brown's article on Wired to read a more 0:25:42.520 --> 0:25:45.399 detailed story about that. But we need to move on 0:25:45.600 --> 0:25:48.960 at this point. So Babylon ultimately goes out of business, 0:25:49.240 --> 0:25:53.360 sells its various business divisions and assets off to other 0:25:53.400 --> 0:25:56.240 companies to return as much value to investors as possible, 0:25:56.359 --> 0:25:58.760 and goes by by Now we're going to talk about 0:25:58.800 --> 0:26:03.080 a different digital health company that had AI aspirations, this 0:26:03.119 --> 0:26:07.760 one called Olive, sometimes called Olive AI. So for this bit, 0:26:08.000 --> 0:26:12.040 I'm referencing a few different articles that I found particularly 0:26:12.119 --> 0:26:15.000 helpful while reading up on the company. One of those 0:26:15.400 --> 0:26:19.359 is an article by Emily Olsen in healthcare drive dot com. 0:26:19.400 --> 0:26:21.840 It was written back in November twenty twenty three. It 0:26:21.920 --> 0:26:26.280 is titled health AI startup Olive to shut down. So 0:26:26.400 --> 0:26:29.640 spoiler alert there, except you know, that's what this episode's 0:26:29.640 --> 0:26:32.520