Episode Transcript
[00:00:00] All right, you guys, I have to talk about what just happened today in the online learning industry. And I know this industry pretty well. I used to work in it. I was one of the early employees at a company that was publicly traded in the year 2000, worth almost a billion dollars 25 years ago, called Digital Think. So I know the space fairly well, and of course, we're in it ourselves with Galileo. Well, it's collapsing, and I don't think it's a surprise to me, and hopefully it won't be a surprise to you. The company Udemy was just sold to coursera for about $930 million. Udemy is trading at about 1 to 1.2 times sales, which is extremely low. In other words, the market doesn't expect any growth out of them at all. The combined company, they claim, is worth 2.1. But I think this is just going to keep going down because the demand for online learning is slowly, acceleratingly, going, going away. Now, there's a lot of businesses that do this. There's skillsoft, there's pluralsight, there's many, many small companies that do leadership development. We obviously have a training academy for HR and for leadership. So every specialized market, whether it be sales, technology, Oracle products, companies all have training to teach you as a professional how to do something, how to use their technology, how to solve problems, how to be a better leader, et cetera. And my estimation from all the research I've done over the years is if you total that up, plus the training that is done inside of companies, it is over a $400 billion market. But the online learning courseware market is the one that is collapsing. And the reason it is collapsing is because of AI. And you can read all about this in our research on the revolution of L and D. And you'll see in the report that we're announcing in January that next year L and D is really going to be transformed. And those of you that run L and D need to be aware of this. Here's the problem. The L and D or training online learning courseware business is, for the most part, a publishing business. The way these companies work, although it's changing slowly, is they assemble subject matter experts. In the case of Udemy, they use subject matter experts from the Internet. In other words, creators who are experts. These creators or authors work with designers to build courses. Courses have chapters, courses have tests, courses have assessments, courses have simulations, courses have exercises, courses have group projects, courses have interactivities. We built all these things 25 years ago at Digital Thinking. And you can Build them pretty easily. Today, however, AI can build almost all of them automatically. Not yet, but pretty soon, all of them. In fact, I sat down with Sora, the tool that was just announced from OpenAI, I don't know, a couple months ago, and I built a five minute course on the economics of frontline work in about 10 minutes. And it created a pretty cool little course with only some prompting from me. No graphics, no instructional design, no narration, no video authoring, nothing. It did the whole thing. And we do this in Galileo Learn all the time. The courses in Galileo Learn are generated by a small team on our side and a lot of AI and they're constantly being updated with new content. They can be translated into any language and you don't even need to take them. You can literally go to the Super Tutor and ask it a question and and learn. So why would you spend millions of dollars a year, or hundreds of thousands of dollars a year on a massive course library when the content can be generated dynamically? Well, in some sense, what these companies have to become if they want to stay in business is giant teams of subject matter experts training agents like OpenAI itself. And OpenAI is already pretty doggone smart. I don't know where it gets all its content, but I've gone to OpenAI for many, many educational questions and it seems to know the answer to almost everything that I ask. And it can create the answer dynamically and it can create various different kinds of interactions to make it easier to consume. Now, that doesn't mean it's a course. A course is a linear, predictable, structured way of learning. And when you're early and new to a topic, a course is great because you don't really know what questions to ask and you need some fundamentals and some basics. However, you can also go to YouTube for that. And I learned most of what I know about AI by watching some beautifully authored YouTube videos that were explaining what an LLM was, how a neural network works, et cetera, and found very few courses on LinkedIn or anywhere else that really seemed to be that valuable. And the problem is, is really one of the paradigm, the paradigm of publishing and creating content. The linear way is too slow and it's just not relevant in a world of AI where a user wants to learn something quickly or ask a very specific question about a topic. Although I have no insights in specifics about what Coursera is up to, if these guys don't build something like that within five years, they're not going to exist. Now, that is not to say that Companies like Coursera, Pluralsight, Skillsoft, LinkedIn Learning are irrelevant. Not at all. They're very relevant because there still is a lot of demand for traditional training, but it's going to shrink, not grow. 30, 40, 50% of the 8, 900 million inquiries a week on OpenAI are about learning something. The data comes in different sources, but generally it's at least 40%.
[00:05:58] So we're talking about hundreds of millions of people per week getting the learning they need from OpenAI and they're not paying for it and they're not signing up for some complex pricing scheme where they have to buy a bunch of stuff in advance. So the business model, the technology model, the delivery model, the operating model of these training companies is just out of date. Now, in the case of Coursera, Coursera has had a pretty successful opportunity to reinvent itself multiple times. It started as a bunch of college courses taught by professors which were very exciting. In the early days they were called MOOCs because we couldn't get access to college professors without going to college and it was too expensive and a lot of those courses were free. Then they started monetizing the content based on certificates and tests and credentials. And then they went into the corporate space and built more custom curriculum for corporations, all of which are great. Pluralsight did the same thing. Skillsoft did the same thing. LinkedIn still has a more generic library, but they're going to probably go that direction too. But if you're a corporation and you want to build a bunch of expertise in, you know, maybe cloud engineering or data analytics or something like that, that's not necessarily a multi year project. You might buy access to one of these systems for a while and you may keep it at some point level of funding, but you're not going to buy this massive course library like the business models work before. So the multiples of these companies have plummeted and they really all have to reinvent themselves now without giving away all the secrets, because we do know a lot about this, the real opportunity here is to rethink the whole paradigm around continuous AI driven content. And what happens in AI driven content, as we know very well from Galileo, is once you start building a corpus of content on a topic, you have to be laser focused on the quality of the content, the accuracy of the content, the timeliness of the content, the contours of the content, the expert labeling of the content. You're still in the content business, but you're not in the publishing business anymore. And I'm not saying again that courses are going to go away overnight. But I would not be surprised if five years from now, we all learn through inquiry on tools like ChatGPT. And you know, those tools are going to be embedded in our phones, embedded in our eyeglasses, embedded in our computers, running locally on those devices, not even going online to one of these massive data centers. So the content will be assembled and collected in a strategic way by the AI.
[00:08:44] So in our particular case, we feel very strongly that this is essentially a business of being an expert and collecting information from experts. Now, getting back to Udemy for a minute, this is essentially what Udemy did in its early days. Udemy stock went to almost $31 a share when it first went public, because the original paradigm or business idea for Udemy was to collect and crowdsource a creator economy of content experts. And what Udemy sold, and I was very involved in helping them with this, was this idea that content would appear faster and more authoritatively in Udemy than in any other platform because an expert on the Internet would author a course, Udemy would share content or revenue with them, and the course would become available long before LinkedIn or Skillsoft even knew that a course was needed. And that actually worked really well for Udemy. And Udemy did quite well and probably still has a large of large number of contracts in the corporate market. So, but it quickly panned, didn't pan out. And if you look at their stock, you can see what happened is as more and more companies built more and more content, and the content market is in some sense a commodity market, because anybody can build a course, it became harder and harder for Udemy to differentiate itself. Udemy also suffered the challenge of the consumer versus the business market. And we did this at Digital Think too. The first online course that made Digital Think famous was a course on wine tasting, believe it or not. And in 1999, that was a very hot topic because the Internet was new and nobody could take anything. There really wasn't anything online like this. And the consumer market, you know, kind of panned out for a while. But, you know, tools like YouTube and even Twitter and Instagram slowly just made it harder and harder and harder for professional authors to make money in the consumer market. You can look at Masterclass, you can look at section, you can look at, I mean, in some sense, even Big Think. Lots and lots of companies who love to author content have struggled to make money in the corporate training market. So it's still a massive space. And there is demand for credentialed training, There is demand for certification there's demand for recertification, you in every role in every industry and of course in AI, there's lots and lots of new skills to learn and lots of new technologies being invented. And so these companies are not going to disappear, but the growth trajectory for them and the current business model is really dismal. In the new model, what's likely to happen in my opinion is there's going to be companies selling things like Galileo. Galileo is a very unique product and I'm not here to pitch it, but I just want to explain it.
[00:11:31] Galileo is an agent and a learning system coupled together. The agent behaves like ChatGPT. It answers any question about training and learning and HR and workforce planning and vendors and the HR market. It answers any question about leadership performance management, goal setting, succession, et cetera. It's a leadership coach, it's an HR consultant, it's an HR advisor, it's an HR business partner. But then behind it is a dynamic learning platform, also powered by sana, that allows us to build courses. So when you sign up for Galileo, you're getting 750 courses on every area of aspects of management, leadership and HR that we have authored with the help of AI, not solely by AI. That business model is actually to me, where all of these companies need to go and they need to find ways of doing IT in a unique fashion that focus on specific domains. I think IT training seems to be the one that's the most interesting to all of them because there's always a demand for IT training and there's always a lot of changes and new technologies and new products and new techno new topics in IT training. By the way, just to give you another sort of nuance on this market, one of the most successful companies in technology right now is a company called Databricks. Databricks sells a very sophisticated data platform that's used by, I don't know, thousands of companies. It's a multibillion dollar market cap company. Databricks head of Learning is a very interesting, high performing professional and she told me when I interviewed her about this that she doesn't even consider herself to be a learning professional. She told me she does not want to work for hr, she wanted to work for the CEO and she wanted to be responsible for learning and enablement of the entire ecosystem of Databricks users, including the employees. And so what she's been doing is the same thing in some sense we do with Galileo, is building an enablement platform that does formal learning certification, deep learning, deep skills development and problem solving, and Information gathering on behalf of this ecosystem. One of the companies they work with is a vendor called uplimit. I'm sure they work with others and they're very crafty people. I'm sure they'll be using other tools. That's where this market is going now. The Coursera people are not dumb. They, they know this. The Udemy people know this. The Skillsoft people know this. The LinkedIn people know this. I'm not the only one who sees where this is going. I may see it, I might see it a little more acutely because I've been doing it for so long. But the problem they all have is getting from point A to point B. When you have thousands and thousands and thousands of courses and hundreds and hundreds of videos and lots and lots of expert curriculum, it isn't an overnight shift to suddenly make everything dynamic. We actually did this. We took our original Josh Burson Academy, which was really a spectacular experience in the old paradigm, and we shifted it over to a dynamic learning platform over a period of about six months of hard work. And, you know, we're not that big. We're not trying to build training on every topic on the planet. So I think what's going to happen in these companies, and Udemy had been doing this already, is they're going to spend a lot of time re engineering what they have for this new world of online learning. On the positive side of all of this, this new world of online learning is just spectacularly fantastic. I know this because I experience it myself. Imagine you get a new job or you're in a meeting and you get a project or an assignment and you need to get involved in something that you haven't done before and there's a new technology and there's new jargon and you're trying to catch up in a hurry. You can go to a learning platform like Galileo or another AI driven platform and you can start asking questions. You can literally ask Galileo to build a course, a development plan for you as a novice. And if you've been using Galileo for a while, it knows your skills and it'll show you and build you a learning path based on the content we've already created and new dynamic content, and you can just consume it as needed from your phone, from your computer, from wherever you have access to it. Imagine that experience over time. We're all going to have personal coaches, personal tutors, personal learning experiences that will help us with our professional lives. And there's no reason why these AI platforms shouldn't. If you use one that you use for other purposes shouldn't get to know our learning needs in a more extensible way. So if you think about the opportunity for a company like Workday that owns Sana or Microsoft, who obviously has the copilot, if you're using the AI platform or an AI platform for your own daily productivity and work, the system can easily look at the content you've created, the emails that you're dealing with, the documents in your corpus, and understand what you do and recommend training based on your activity. You can even use Galileo to recommend training based on the meetings that you've been having, because the meetings are recorded in Galileo. So this in some sense is a massive new business opportunity. But I always wonder if the companies that thrived in the old paradigm can figure out and move to the new paradigm fast enough. And you know, I'm a little bit older than many of the people listening to this podcast, and I've lived through this before and it's fairly unusual for a company that thrived in one technology era to thrive in the radically changed one that replaced it. I'm not saying it can't happen, but it rarely happens. And I think what's probably going on at Coursera is they're probably buckling their seatbelt and saying we need to strap in for our very, very important ride here. And the more resources and the more customers and the more revenue and the more staff and the more content we can get right now, the better.
[00:17:32] I'm not exactly sure if that's going to work out. We will see. Having a large library of content can be a good benefit during this transformation, but if it's not the right content, it can also hold you back. Okay, so what does this mean for you as a business person, as an HR person, as an L and D person? I think you need to read our research report on the transformation and the revolution of L and D.
[00:17:55] In February, we're going to launch a massive study we've been working on for over almost two years with a maturity model and detailed examples, case studies and specifics on how to transform L and D. We have now gone through this with four or five of our largest clients and helped them move to this new world. And we're all going to get there in a big hurry. And I think 2026 is going to be a very, very exciting and important year in this new world of corporate learning. Stay tuned for more on this. Obviously, I'll keep you up to date and anyone that would like help with their strategy or understanding the market including those of you in the vendor space, just please reach out to us, and we'll be happy to talk you through your situation and help you adapt to these changes. That's it for now.
