The End of College? Inside a16z’s School for Tech Titans | Fortune Daily

0:00Silicon Valley is the cradle of innovation because of dropouts. College has become a cruel, expensive joke. That was a 2014 clip from the much-loved comedy series Silicon Valley, where it lampooned college degrees in tech culture. But actually today it could be just as on the nose as ever. Because just last week, Andreessen Horowitz launched a new school in San Francisco. And it's for gifted high school graduates who believe that they have what it takes to join the tech elite. There'll be no grades, no tests, no traditional teachers, no formal accreditation. The first cohort will be 50 students and they'll start in September 2027.

0:51They won't pay anything, but for everyone who comes after them, tuition could cost as much as$90 ,000 a year. So is this the future of education? Or is it a high stakes bet that only works for a tiny, uber-connected, very wealthy few? Well joining us to answer all of those questions and a few more is the CEO of the new Horowitz Andreessen Academy, Gargan Biani. From New York, I'm Ellie Austin and this is Fortune Daily.

1:35Gargan, thank you so much for joining us to talk about this new endeavor. For viewers who don't know, what is the Horowitz Andreessen Academy and who is it going to be for?

1:46Gagan Biyani: Yeah, the Horowitz Andreessen Academy is a new school for high school graduates who are looking to learn how to be builders in the new economy. Tell me a bit about when you say builders, that could be broad, what kind of thing are they going to be built? Yeah, I mean, the idea is that if you're in, you want to join a technology company or you want to start technology companies, and you're young and you're ambitious, we want to build the place for you. So these are students who, yeah, they might have gotten good grades or get good SAT scores, but actually they spent most of their high school time going and building robots or creating drones, or maybe they were building their own companies or websites.

2:28Gagan Biyani: They were super excited to go and create things and push them out into the world rather than gain accolades or do homework and tests. And when you say young, what kind of ages are we looking between? They have to have graduated high school. So our average student is probably either a senior in, you know, just graduated high school. So they're in their senior year right now, or they're one or two years in college. That's our typical student. And tell me a bit about what the course structure is going to look like at the academy, because from what you just described, it's going to be pretty exciting.

3:03Gagan Biyani: Yeah, it's extreme. We flipped the typical model on its head. So instead of having a bunch of classes and you get credits for the classes and you have to get grades, we have the opposite approach. You come to the academy and your job is to create pursuits. And so you come up with your own pursuits. And these pursuits could be like, I want to learn how to write a novel, or I want to learn how to build a drone or I want to go and study LLMs. You know, pursuits can be things that are challenges for yourself that you want to learn or they can be actual products in the real world. And for any pursuit that you take on, we expect that every week you will have a meeting with your peers and a mentor that will basically be, hey, what did I do last week and what did I do the next week?

3:53Gagan Biyani: And so the entire year and the majority of the program is designed around this idea that students should be controlling their own destiny in terms of what they should do. It's only designed for the most high agency, most ambitious students because of this. When you talk about pursuits, you mentioned writing a novel, but do their pursuits have to be skewed towards AI and tech innovation? No, we expect that probably about a third of pursuits will be completely unrelated to AI and tech innovation. And then the other two thirds will probably be closer to actually building products and services in the AI and tech sphere.

4:31And so we believe in studying things like, hey, I want to go read the five of the great books of philosophy or I want to go and learn calligraphy. I've heard you say since this was announced last week that this system, this academy, it isn't a rejection of college per se, but it is a rejection of sort of a centuries old model or an upending of that model. What is it? And you just touched on this, but tell us a bit more about what it is exactly within that traditional model of higher education that you think it's time to reinvent.

5:01Gagan Biyani: Yeah, I think that the modern ethos of higher education is oriented around learning things rather than doing things. And the big difference of the Horowitz and Dresen Academy is that in the AI era, especially more than ever, we are no longer in the world of you just need to know a lot of things to be successful. You have to not just know things, but you need to have the mindsets to go and apply what you know in the real world. And that's what makes the Horowitz and Reeson Academy far more useful for these young builder type people than a traditional experience. Now, you've been building alternatives to traditional education since around 2010.

5:46You are co-founder of Udemy, which is a massive online learning platform focused around on-demand video. There's also Maven, which is built around online cohorts. And now this in-person academy in San Francisco. What was it, Gargan, about your previous endeavors, building those two platforms that I just spoke about, that led you back to wanting to do an in-person education setting?

6:10Gagan Biyani: I think that it's just a super unique time to be young right now. And I've been meeting all these great young future entrepreneurs and executives and engineers and product managers. and I was just so excited to bring them together. And when you're at that age, you're just graduating from high school, you're thinking about what you wanna do next. You wanna go meet the other amazing young people all over the country and all over the world. And so bringing them together in person is a necessity of delivering an amazing product offering. It's not something that I was pursuing because I wanted in person or digital.

6:45Gagan Biyani: I think all learning methodologies are valuable in different purposes. and for the purpose of particularly delivering for the best and brightest. We need to be in person. In fact, we can't just be in person. We have to be in person in San Francisco because they need access to the best and most cutting edge research, technology and thinking on the planet. And tell us a bit about the companies that you've partnered with, many of them headquartered in San Francisco, who are going to be coming in to speak at the academy and then offering some of the students internship opportunities. Yeah, so we're very fortunate to be partnered with Anduril, Anthropic, Coinbase, Google, Meta, NVIDIA, OpenAI, Palantir, Replit, and Stripe.

7:29Gagan Biyani: So it's 10 companies that are our founding partners. And these partners are helping us co-create the curriculum. They are helping us design, you know, they're offering compute and credits so that students can go and build things with their software and with their hardware. And some of them are offering internship, we call them co-op opportunities as well as full-time opportunities. And then we have a whole other group. Those are our founding partners. We have another 50-plus hiring partners, which includes up-and-coming companies like Base or Decagon, and then lots of bigger companies like Databricks and Shopify.

8:05Gagan Biyani: And these companies are also looking to offer exclusively focused on co-op or full-time opportunities for our students. So we believe that it's critical for a modern institution to have connectivity to the employer ecosystem of San Francisco, as well as the compute and research ecosystem. And we're doing that through these partnerships. You mentioned a second ago that the students will be in control of choosing their pursuits. And one thing that came into my mind there is, is every 18, 19, 20-year-old well-equipped to choose what's best for them and what's best for society to be dedicating their time to, particularly if they are incredibly gifted, as you say?

8:48How do you navigate that specific tension?

8:51Gagan Biyani: Yeah, well, first of all, it's worth noting that almost every university in the country lets students basically decide most of what they're going to do. So just to be clear, like... Because I guess you're positioning the academy in counterpoint to those traditional institutions. Exactly. So is this autonomy different then? So students always have autonomy, but the autonomy is defined within an existing structure. So if you want to get a major, you have to do these 20 things. If you want to get a major in economics or computer science, the difference here is you actually make micro decisions around each of your pursuits.

9:24Gagan Biyani: So instead of having control over deciding an entire topic area you're focused on, in the AI era, being multidisciplinary is extremely valuable. So all of our students are going to have multiple pursuits that create multidisciplinary humans, and all of our pursuits are designed around action. So as, you know, let's say you're taking econ one as the first thing you're doing to become an econ or a business major in college at the academy, you would actually go and run a experiment that would allow you to assess, to learn about economics by maybe starting a exchange. You know, like maybe you build a clothing swap and you figure out how to do it in an auction method.

10:04And the first 50 students are going to go free and they're going to be in the program for a year. And then after that, it's going to become a two-year program. And reports state that it's going to cost out at about$60 ,000 to$90 ,000 a year, which is on par with tuition at many elite universities in the US. How do you foresee students paying for this going forward?

10:27Gagan Biyani: I think there will be multiple ways. First of all, we will have a scholarship program. So those who do not have the means will get scholarships. for part of their or all of their tuition. Parents are an obvious place. We think parents are going to find this to be an extremely, like one of the best ways they can possibly invest in their children's education, you know? And so hopefully people see that it's extremely high ROI to have your student, your child go to a program like this. And then finally, I think a lot of them will make money. So I think a lot of the students that we attract are people who have already made money, either in high school or will make money when they're at the academy.

11:07Do you foresee there being enough scholarships available, for example, for whatever number of lower income students require them or might they be capped? I know you're just in the early stages, but what are you thinking right now?

11:19Gagan Biyani: Yeah, we think we'll be able to provide scholarships for anyone who needs it. We also think that, you know, we expect that there will be students of every income level applying for this program. And so therefore, I'm not expecting 100 % of students to have that need. So one of your stated goals is teaching students to build, as you said, and to build the most important skills in the AI era right now. I think what's interesting and what we report on a lot in the show is that those skills are constantly changing. So how are you thinking about designing a program for an industry that is reinventing itself every few months?

11:56Gagan Biyani: I think on a very micro level, the skills are changing, but at a macro level, they have never changed. And really, they're just becoming more accentuated. So the idea that you need to be ambitious, have good systems thinking, be able to build, you know, ship things into the world, have high agency, those haven't really changed throughout this entire era. the specifics of what you're building and what you're using to build do change but that's exactly why we do not have a curriculum that is highly narrow we're not like we're going to teach you how to use cloud code or we're going to teach you how to use you know chatchiput codex no our view is that you will choose the tool that is most useful for the product and service that you're trying to build and the way in which you want to build as long as we teach you how to build things and how to build things that make a lot of impact and people love then it doesn't matter what tools you're using.

12:47And what's going to be offered in the way of personal development? Because one of the reasons that students go off to a campus for four years is yes, to get the academic credentials, but also to become a well-rounded person, be around other kids of diverse backgrounds and learn how to operate in the world. Yeah.

13:04Gagan Biyani: Well, first of all, we think that being a successful builder is one of the best ways to grow personally. And so just to be clear, like building also involves, you know, creating dinner parties or hosting art events. And the types of people who do those things tend to have better social lives, find their partners, etc. The second thing is that we have lots of courses that are about interpersonal skills. And interpersonal skills is one of the most important. It's one of our four core tenets of things we're trying to ensure. So we will consistently encourage and sort of shape students' ability to go and meet people and interact with the world.

13:43Gagan Biyani: And that's one of the things that I think makes the academy unique. It's set in San Francisco, and it forces you to go interact with adults, quote unquote, as opposed to solely focusing, still being like a campus that's in the middle of nowhere where you're really only interacting with your fellow students. We're hearing so much right now about, well, does AI innovation need to slow down? Have all these tech founders been moving too fast? And I wonder on that note, Gargum, what you were just saying about bringing together a cohort of highly capable, highly ambitious, fast-moving builders. Is there any risk associated with that for society, given that we see the Zuckerbergs and the Amodis are now wielding so much power in what humanity looks like going forward?

14:28Do some of us mere mortals need to be in the mix to A, give the builder's perspective and B, maybe say, slow it down from time to time?

14:36Gagan Biyani: Well, I mean, if we teach 50 students how to be the next Amodai or Zuckerberg, I think we'll be very happy with that result. And I think that the benefit to society is immense when people go and build great things. I'm not as much of an AI doomer. I'm an AI boomer. And I believe that this is going to be one of, we're going to look back and this is going to be one of the greatest times in human history. Who is teaching at this academy? Obviously, you'll have visiting speakers coming in. Who are the day-to-day teachers? Yeah, we have a wide variety of instructors. Our instructors are industry experts and they will have their own day jobs.

15:14Gagan Biyani: That's very important because we don't believe in having a full-time faculty. We want people who practice their trade regularly. So an example is Max Mullen. He's a co-founder of Instacart. He is going to be teaching a course on building and shaping teams. We have Logan Urey, who is a behavioral economist and behavioral scientist and is also a data scientist. And she is going to teach a course on finding long-term healthy relationships. And so our instructors are people who have, and we have dozens of others, you know, Hamal Hussain, who works at OpenAI, and some people who, and then Lannis, who works at Anthropic.

15:59Gagan Biyani: So we're picking people who have real industry, deep industry expertise that are almost impossible, honestly, to learn from if you don't attend the academy. me. And how are you going to think about diversity of applications? Because obviously, students who have gone to really elite universities may have the academic credentials and kind of may have had an environment where they had the privilege of being able to build, but there's going to be a lot of other kids who maybe didn't have those opportunities, but have the core skills that you're looking for. How are you going to seek them out? Well, I mean, that's our job as a school, you know, we have to go and recruit and find the best people.

16:33Gagan Biyani: If everyone we find is just from an elite private school, I don't think people will be as impressed by our success. So we're going to find them by going out to various extracurricular activities where the young builders are and building partnerships with organizations. And we're also putting the word out on this podcast and in other places, right? And what does this model look like in 10 years, Gargant? Do you want there to be multiple academies? Is it one enormous academy in San Francisco? Are you planning to scale it and what might that look like? We certainly want to scale it, but right now we're solely focused on just delivering an amazing experience to 50 students in San Francisco.

17:10Gagan Biyani: So we'll figure out the right way to scale it once we nail it with the first few hundred students, you know, over the first couple of cohorts. And those 50 students will walk in the door of the Academy in September, 2027. That's right. Gargan, thank you so much for your time. Good luck with the Academy and I'm sure we'll be back to you to find out how it's going. Thanks, Sally. Thanks for having me. Next up, information has leaked online about Anthropics' big IPO. So what does it tell us about the company? Senior AI reporter Bea Nolan joins me to break it down. Bea, hello. Thank you so much for joining us from London again.

17:51It's great to see you. So we weren't meant to see this information about the Anthropic IPO. Tell us what's out there now in the public domain and how you think it was leaked. Yes, so we weren't meant to see this. It was a draft of the company's S1 filing that Reuters and the FT both got a hold of. It showed us a few things. I mean, the extent of the losses the company is suffering at the moment, and also some of how it's thinking about the risk, specifically the existential risk we've been hearing about so much recently. One caveat, maybe just a flag, is that these numbers are actually a little bit out of date.

18:28So we should see more up to date figures in the actual S1 when it goes public. These are from 2025. Got it. Okay. And the numbers, just to give people context that were published, it says a$42 billion loss on$4.6 billion in revenue. I know you don't know for sure, but how does something like this get out into the public domain in the first place? Well, it's hard to say without having the documents ourselves. It could be a case of these have been handed out to some early investors. Yeah, that's a really hard one to say until we have that information ourselves. B, two other striking figures from the report.

19:04So the reported loss is$42 billion, but the operating loss is$8.6 billion. What accounts for the difference there and which figure should investors actually pay attention to? Yeah, so the short answer is that you should pay attention to the 8 billion operating loss. Most of the gap with the 42 billion is an accounting change. So it's not money that actually went out the door. When Anthropic raised money in the past, some investors got deals that they could later turn into shares, the company's value then shot up. So those future shares became worth more on paper and accounting rules count that as a loss.

19:39So I would definitely either 8 billion. And is there a precedent for a company going public in this shape? So in some ways, yes, Amazon went public while it was still making operating loss and so did Uber, but we've never really seen it to this extent before. Anthropic is still making around an 8 billion operating loss as of last year. We've also never seen a company go public while citing that its technology could pose a risk to wiping out all humanity. That's certainly a new one for the risk factors. Yeah, and a pretty scary one, I have to say. Has Anthropic responded to this in anyway? They haven't.

20:16A lot of these numbers actually have, some of them have leaked before, and we've seen some other leaks of more up-to-date numbers and things like this. Their response will be their public S1, which we could see, you know, in the next month or so. The IPO is predicted to come just after the midterms. So that means we'll see this soon. One of the things that really stuck out to me, Bea, from this information was that two unnamed customers accounted for roughly a quarter of Anthropik's revenue last year. What do you make of that in particular? Yeah, so customer concentration is another risk factor that they flagged here.

20:51Maybe you would be more commonly see that in smaller companies, but for a company that's heading towards a$2 trillion valuation, you'd certainly want to see a more diverse customer base. Anthropik, I think, also flagged that a lot of their customers weren't locked into long-term deals. So any cutback on spending from these, you know, really big spender companies can really dent its revenue. Got it. And so does this, if we're looking at this kind of as a big picture story, does this leaked information change the narrative around Anthropic right now? It's obviously a company which is consistently in the headlines in the zeitgeist.

21:26How does this impact what we know about it? I don't know if people were maybe that surprised. It's good to see it all down, but there was already, we already knew that these companies were burning a lot of money, partly due to the huge amount of compute power needed to train AI models. It's just very expensive. And we did already know about these existential risks as well. But it will definitely be interesting to see how Wall Street reacts to this when they do go public. Bea, thank you so much. Great to chat. Great. Thanks for having me. That's it for Fortune Daily. I'm Ellie Austin. A quick update.

21:58Yesterday, we discussed Fat Bear week in Alaska and since then Bear 89 has been crowned as the chunky winner. Congratulations to him or her from all of us here at Fortune. I'll see you next time.