Changing the economics of an industry (Austen Allred & Ben Grynol)

Technology is disrupting almost every facet of life – including foundational practices such as education and health. Austen Allred is the CEO of Bloom Institute of Technology, formerly known as Lambda School. The premise is a unique one: it is an online technical coding education program with an income share agreement model in which students pay a portion of tuition up front, and a portion after obtaining a job. In this episode, Austen chats with Levels’ Head of Growth, Ben Grynol, about the school’s unique approach, the future of adaptive education, and the similarities to onboarding and educating Levels’ customers.

9:38 – The early days of Lambda School

Austen started Lambda as a series of free online classes with a home base in Slack.

The way we started Lambda School was actually just by teaching free classes. It was basically, in the early days, we threw everybody in Slack and then we would have a Zoom that was scheduled and we would @channel everybody, “Hey, the Zoom is going live right now.” We would have thousands of people. At the time, Zoom didn’t have a big enough overflow so we’d also live stream it to YouTube. So you had Slack, there was a chat there. You had a Zoom chat and then you had YouTube chat. Slack was our early crutch where it’s not built to be a place for a school to happen necessarily, but it was something that we could pull off the shelf and your people are divided up into cohorts and they have channels and they can send messages and they can drop code snippets and they can message with each other, they can message with an instructor. We can create help channels. So it was super chaotic in the early days, but it was just like, let’s just get people in the same digital environment and let them interact with each other.

11:27 – Democratizing the Silicon Valley experience

One of the catalysts for Lambda School was the widespread desire to learn skills that were primarily only accessible in places like the Bay Area.

Before COVID, we started doing something that was really cool. We had in-person meetups, which in the early days part of the thesis of Lambda School was, A, we wanted to reach people who aren’t in a position to move to San Francisco necessarily. And so, specifically me, I moved out to Silicon Valley and lived in a car and I moved there from small-town Utah with 4,000 people. There literally was not a code bootcamp within a several-hour drive of that location. And when I would go back, everybody would say, “Hey, I want to do what you did. I want to figure out…” And I didn’t attend a bootcamp. I taught myself everything, but, “How can I follow that same path?” And there wasn’t a great thing for me to recommend. I couldn’t point to something and say “Do that.” So the original thesis was just, “Let’s reach all the people in hard-to-reach places and let’s do it using technology.”

12:24 – The shortcomings of in-person bootcamps

Compared to traditional coding bootcamps, Lambda has financial advantages from being 100% online.

If you look at code bootcamps financials, a non-trivial amount of their spend goes to real estate and physical location. And then they base everything off of how many students can I get in that room at that time, which then means I have to turn that room over four times a year if I want to break even which then means I have to have my school be three months long because I need to get people out of that classroom. So it’s when you break those barriers down, the cost structure is different, the accessibility is different. So really, actually the earliest focus was just building something online, which feels obvious now in a way that was not obvious in 2016, 2017. And from there you realize, “Oh, here are the advantages we have. Here are the disadvantages. We can build this to overcome disadvantages. Now that we have this advantage, let’s lever up on that.”

13:55 – Why class size doesn’t matter

Traditionalists would argue that a 1,000-person class isn’t feasible. But when you have sub-groups and a community of peer support, it works.

Your help is coming from a giant team of other people who are there instantly. And that’s something that’s very difficult to do in a physical environment, but it’s obvious in an online environment. And so we can break up and subdivide into classrooms instantly of people who are stuck on Section 3 of Unit 2 in the fourth problem set. Now they’re instantly together with someone who is focused just on that and working with them. And it’s way better than a single teacher running around to all the students trying to handle everything individually.

15:07 – The value of adaptive education

When you have flexible online programs, students can learn at their own pace.

What if a student wants to speed up? What if they know this thing already? How can they fast forward that?” Sometimes you can cut a month or you can cut two months off of the learning because they’ve already done that. Or “What if I move at a faster pace or what if I move at a slower pace? How can the school and the software automatically flex around that and just use the scheduling of instruction time and help in real-time to adjust to all those people?” So what’s kind of the next step is, I’m sure in the future we’ll have people that takes three months to get through the same content that it takes somebody else a year to get through. And then you can accommodate schedules.

21:15 – Customization at scale is the next frontier

If Amazon can customize your shopping dashboard and experience, Austen thinks the next step is customizing other things – including education and health.

I think at scale customization is really, really interesting. And I think some products have been doing that already where if I log into am Amazon, it’s a completely different experience than if you log into Amazon because it knows our history. I think that level, instead of shopping or consuming content, combining that with “What should I be doing?” I think about “What is my personal dashboard? What habits should I be forming? What behaviors should I be shifting?”, that’s incredibly powerful. I think health, to me, is one of the most exciting places to apply that because it is so customized. It has to be customized. And education is another one where my health is completely different than my wife’s health even if we’re eating the same thing every day. The way I experience an educational product is completely different than the way my wife experiences an educational product even if the content is exactly the same. So I think that’s kind of the frontier.

24:19 – Adaptive education software

One way to customize education is by using adaptive software. Austen cited test schools that are now outperforming expensive private schools.

Think about who the smartest 1% or 2% of people are in a traditional classroom, and you can get half the class to that level. That’s a massive, massive shift. So what is the response in the education community to that? The response is, “Well, that’s great, but we can’t afford to hire a personal tutor for every single student. That’s not feasible.” So really nothing changed. But what we’re seeing with adaptive software, I’ve seen a couple of schools that have done this really, really well. We’re actually seeing students with the right level adaptive software outperform students who have one-on-one full-time mentorship because the softwares can be even better at adapting to that student’s needs and isn’t trying to keep a mental log of what a student knows and can put back repetition of what they haven’t learned yet at levels that are better than a human.

26:48 – The norm is suboptimal

We’ve grown accustomed to the status-quo, but the truth is that adaptive education is sorely needed.

The norm today is just really suboptimal. It’s not because it’s revolutionary. It’s just a standard classroom environment is really bad. And it’s not because the teachers aren’t doing their best, it’s just the structure is bad. My wife was a 3rd-grade teacher and it’s impossible to teach a random selection of eight-year-olds division, teaching the same class of 30 people the same material. And they’re going to be on wildly different levels. And take that for every subject and that’s going to be true. Whereas if it was more adaptive to who they are and what they need, every single student would do far better. And so I’m really, really excited for a world where that’s true.

33:36 – The value of scaled help

Austen thinks the best future learning model is one where you get small doses of one-on-one human help, plus a lot of automation.

Having scaled help is 10X better than the old way of you either had to go it alone or buy a book or pay somebody incredibly high rates to do one-on-one training. Rates didn’t have to be incredibly high. It’s expensive to do anything one-on-one with somebody sitting there watching you. So yeah, I’m really excited to see what that level of software looks like. And it probably starts out, it’s adaptive enough to bucket you and then you have a human in the loop as necessary. And then over time, sure, it’s automated. The marginal cost is zero. We’re not at that level yet. But what we can start to do is have everybody the experience the outcome of a full-time personal trainer for 1/10 of the price. Or even in my opinion, slightly better outcomes than a full-time personal trainer for 1/10 of the price. And that’s really, really cool. That’s the power of technology.

41:57 – The biggest challenge for Lambda

Getting students to fully commit to the education and job search journey is difficult. But Austen and the team are brainstorming ways to solve this.

The difficult part of Lambda School isn’t teaching people to code necessarily. It’s a lifestyle change. So we actually have to bet on people being able to make a complete lifestyle change. Some people are closer. But think about if a personal trainer only made money if you lost weight, they can have the best program ever. They can nag you all the time about making sure you’re working out. But then, if you’re eating bad on the other side, you’re not going to move the needle. So we’re discovering over time that there are ways we can filter for that. But in the long run, I think really it comes down to, if you can get a student to financially commit to do that, then we can offer way more. We can actually reduce the tuition when it does work. We can offer stronger guarantees. So incentives have to be a two-way street. And I think we started out by saying, “We’re going to guarantee everything and a student has to guarantee nothing.” We learned that if we can get just a little bit of skin in the game from a student, we can place pretty much everybody in that world. So we’ll have to see how that plays out in the long run.

45:12 – Extrinsic vs. intrinsic motivation

Studies have shown that extrinsic motivation can be an incredibly powerful tool.

They tried an experiment once that was literally, “If you agree to put your phone in this box when you come in the door and not touch it until you leave, at the end of the week we’ll buy you a Chick-fil-A chicken sandwich.” Participation of not… Phone usage went from everybody using a phone to nobody using a phone overnight. Is that a blunt instrument, that $3 chicken sandwich from Chick-fil-A in order to cause students to do that? Yes, but then you can watch outcomes shift in a way that I think educators would be nervous to understand how much motivation a $3 chicken sandwich can provide a 12-year-old. Is that a blunt instrument? Yes, but at the end of the day, I’m pretty outcomes-oriented. And if giving a student a chicken sandwich will make them way better at math, I’m not morally opposed to giving them a chicken sandwich. That’s something that I had underestimated, the importance and the power of extrinsic motivators versus entirely intrinsic.

Austen Allred (00:06):

A standard classroom environment is really bad. And it’s not because the teachers aren’t doing their best, it’s just the structure is bad. My wife was a 3rd grade teacher and it’s impossible to teach a random selection of eight year olds division, teaching the same class of 30 people the same material. And they’re going to be on wildly different levels. Take that for every subject and that’s going to be true. Whereas if it was more adaptive to who they are and what they need, every single student would do far better. And so I’m really, really excited for a world where that’s true.

Ben Grynol (00:45):

I’m Ben Grynol, part of the early startup team here at Levels. We’re building tech that helps people to understand their metabolic health, and this is your front row seat to everything we do. This is A Whole New Level.

Ben Grynol (00:58):

When it comes to behavior change, thinking about extrinsic motivation versus intrinsic motivation, it can be very different things in the approach. But For Austen Allred, co-founder of Lambda School, a company that focuses on building and educating people to become developers, people who don’t have previous experience in software engineering. Well, it’s analogous to Levels in the fact that some people who start to use our platform don’t have prior experience in metabolic health. And so, we have to start to think about if we’re creating a product that helps people to understand their metabolic health and all their biometric data, what intrinsic motivation do we need? What extrinsic motivators do we need in the platform as far as giving people the appropriate feedback and the timely feedback to be able to make meaningful behavior change?

Ben Grynol (02:02):

And so, Austen and I sat down and we talked about everything pertaining to education, what he’s doing with Lambda, how it applies to Levels and what we’re building from a product and community standpoint. It was a very meaningful conversation. Here’s where we kick things off.

Ben Grynol (02:16):

Are you staying in Utah now? Is that your new home base?

Austen Allred (02:26):

Yeah, well I mean, we have a house here, but then I do like a week in San Francisco and I do a couple of days in Miami. I don’t know. We’re figuring that out. We’re going to be pretty nomadic, so we’ll see. But it’s the right-

Ben Grynol (02:39):

Why not?

Austen Allred (02:40):

It’s the right cheap home base. It’s really nice, right next to an airport, to buy a giant house with a pool for like half a million bucks.

Ben Grynol (02:49):

It’s crazy.

Austen Allred (02:50):

That’s great.

Ben Grynol (02:51):

I mean, that’s the thing about living in Ohio or living in Utah or Wichita, Kansas, you’re just going to get a lot of value for money, and now-

Austen Allred (03:00):

Oh, it’s crazy. Yeah.

Ben Grynol (03:03):

Especially with what you’re building, it’s like you’re building knowledge workers. And knowledge workers have the ability to be anywhere. And now with people stripping away the heuristic that being in a cluster tech market means that you’re smart, now people are like, “Oh, you could be from Fargo” and it’s totally cool and no one really gives a shit like, that’s pretty interesting.

Austen Allred (03:23):

When Lambda School had a San Francisco office, we had a Utah office, I was like, “Yeah. Remote. All of our students are remote.” But it’s so good to be in the office with everybody. When we made the hard transition to fully remote, I was like, “Why am I in San Francisco?” I think there’s some folks like, “That transition sucked.” But it wasn’t my fault, because there was COVID, right?

Ben Grynol (03:45):

Mm-hmm (affirmative).

Austen Allred (03:45):

It would’ve been way more difficult to transition from office to no office without it. And there’s still some people that are just like, “Look, I need an office that I need to be hanging out.” That’s tough because they’re not going to be happy in the long run, but I think 95% of the company was like, “My life is so much better now. If you do it the right way, it works really well.” So I’m not quite like I would never go back to being in an office. I think it depends on the mission of the company and what you’re building, but for Lambda School, I can’t imagine trying to get everybody into an office. Even like hiring, like how did we hire people in San Francisco when everybody else was trying to do the same thing? That’s bonkers.

Ben Grynol (04:27):

Mm-hmm (affirmative). Yeah. You compete for talent on such a different level. And then as soon as you open it up and you start to look at talent where there aren’t geographic constraints, you realize that especially when it comes to dev work where it’s just about execution of task and you’re like, “There are some really smart people in these hidden pockets,” and it opens up a new world. It opens up… It is a strategic advantage especially if you do have like a fully remote team.

Ben Grynol (04:56):

What’s really hard is when you get into the hybrid models where it’s like, “Well, some of us are in an office and some of us aren’t,” and then you get these two distinct cultures and people are like, “Oh, you missed the office conversation.” And you’re like, “Well, we made a decision.” And you’re like, “Well, we have to figure out a better way of having unity in the way that we communicate.”

Austen Allred (05:15):

There’s a time when 1/3 of Lambda was in San Francisco, 1/3 was in an office and then 1/3 was remote. We had like a couple remote execs. We had an exec in Utah. A lot of the execs happened to be in San Francisco. And pretty often, we would hear from people like, “I’m not getting the information I need. I’m out of the loop because I’m remote.” And then I would talk to the people who are in the office and be like, “Hey…” And they’re saying the same thing, right?

Austen Allred (05:43):

And so I think it was easy to have that remote scapegoat and people were like, “Oh, we’re not very good at passing information because it’s only getting passed around in the San Francisco office.” But it wasn’t even getting passed around in San Francisco, right? So it created a forcing function for us to be like, “Okay, we need to document this stuff. We need to share it more broadly.” We’re getting to a point where every once in a while I hear, “Okay, I get it. You don’t need to send me all this information from what’s happening in this other org that has no relevance to me whatsoever.” And we were the polar opposite of that three years ago. I don’t know how it’ll shape out in the long run, but I think it’s something we were forced to get much, much better at.

Austen Allred (06:27):

When I look at some fully remote companies that are building everything in a handbook or with a lot of Looms and have a really strong document culture, being remote forces you to do that in a way that would’ve been healthy in an office, but you use the office as a crutch to not do as much of that. So I think similarly to how Amazon has to be better at giving you certain types of information than it would if it was just a department store. For example, navigation. You just rely on the physical layout for stuff. But sometimes when you go into a physical store, it’s still really difficult to navigate, whereas Amazon has to have that more nailed down.

Austen Allred (07:06):

So I think it’s one of those you eliminate a lot of the crutches that you used to have and not even realize that you have. And that’s probably doubly true in an education sense. Like once you removed that, “Can I look around the classroom and see what everybody’s doing right now?”, you can actually replace it in a lot of instances with something that’s better where it’s not knowing where a student is isn’t based on looking around the room and talking to them individually and seeing what they’re doing. You just have data in real time about everybody. And once you have that data, not only can you get the data but you can actually guide and shift that experience in real time in a way that you could have done in a physical classroom but you just weren’t going to. So I think when you eliminate the crutches and you use technology to replace them, oftentimes you can actually replace them with something better, which I think is really interesting.

Ben Grynol (07:58):

What you’ve done from a business model perspective is interesting in how that applies to people’s lives, right? So the parallel between Levels and Lambda which is just very interesting is we’re focused on giving people different opportunities with their health, right? We show them how food affects their metabolic health. It’s data or feedback that they’ve never seen before and they say, “Wow, this is life changing.” The parallel is that it doesn’t take that much scrolling through Twitter, let’s just use that as a single platform, to find some tweet that says, “I have changed my life because of Lambda School,” right? There are all these unbelievable community stories about, “I was working in some minimum wage job and I had zero experience. No technical literacy at all, zero experience with computers, and I really didn’t know anything and I felt overwhelmed. But I signed up, went through Lambda, and now I’m making a very healthy income and I bought a house and my life has changed.”

Ben Grynol (08:58):

And so it’s like, when you hear these stories, they replace Lambda with Levels, and granted one has to do with economics and furthering your career path and the other has to do with health, so they’re unrelated in that respect, but it’s giving people a different opportunity to really feel agency over their life in sort of these different aspects of finances or career trajectory, and then like health and wellness. So yeah, it’s just so cool what you’ve done. But when you started to think about building out community, what are ways that you’ve engaged people to do it? Is it organic or is it something that you guys sort of build within the business model?

Austen Allred (09:35):

Yeah, I think it’s really interesting because the way we started Lambda School was actually just by teaching free classes. It was basically, in the early days, we threw everybody in Slack and then we would have a Zoom that was scheduled and we would @channel everybody, “Hey, the Zoom is going live right now.” We would have thousands of people. At the time, Zoom didn’t have a big enough overflow so we’d also live stream it to YouTube. So you had Slack, there was a chat there. You had a Zoom chat and then you had YouTube chat.

Austen Allred (10:09):

It’s just like we used… Slack was our early crutch where it’s not built to be a place for a school to happen necessarily, but it was something that we could pull off the shelf and your people are divided up into cohorts and they have channels and they can send messages and they can drop code snippets and they can message with each other, they can message with an instructor. We can create help channels. So it was super chaotic in the early days, but it was just like, “Let’s just get people in the same digital environment and let them interact with each other.”

Austen Allred (10:43):

And there are a few places that falls apart and is suboptimal. The first is that Slack is not meant to be a permanent medium. It’s free flowing. And if you miss it, you can scroll back. But if you don’t, if you miss it, you miss it. It doesn’t do a good job of… I mean, it has archive and it has search, but it’s not built for that the same way… So now we’ve built tooling where if I need help on an assignment, I go to what we call the Hub. And I can I start asking my question in the Hub and I often see that “800 other of people have asked this question before and here is the best answer for that.” So that is kind of bifurcated from the “Let’s be together as a community.”

Austen Allred (11:27):

Before COVID, we started doing something that was really cool. We had in-person meetups, which in the early days part of the thesis of Lambda School was, A, we wanted to reach people who aren’t in a position to move to San Francisco necessarily. And so, specifically me, I moved out to Silicon Valley and lived in a car and I moved there from small town Utah with 4,000 people. There literally was not a code bootcamp within a several hour drive of that location. And when I would go back, everybody would say, “Hey, I want to do what you did. I want to figure out…” And I didn’t attend a bootcamp. I taught myself everything, but, “How can I follow that same path?” And there wasn’t a great thing for me to recommend. I couldn’t point to something and say “Do that.” So the original thesis was just, “Let’s reach all the people in hard to reach places and let’s do it using technology.”

Austen Allred (12:24):

And then if you look at code bootcamps financials, a non-trivial amount of their spend goes to real estate and physical location. And then they base everything off of how many students can I get in that room at that time, which then means I have to turn that room over four times a year if I want to break even which then means I have to have my school be three months long because I need to get people out of that classroom. So it’s when you break those barriers down, the cost structure is different, the accessibility is different. So really, actually the earliest focus was just building something online, which feels obvious now in a way that was not obvious in 2016, 2017. And from there you realize, “Oh, here are the advantages we have. Here are the disadvantages. We can build this to overcome disadvantages. Now that we have this advantage, let’s lever up on that.”

Austen Allred (13:20):

It’s something that we get slammed for in the press. But when we have a giant live lecture, we don’t focus on the teacher to student ratio for that lecture itself, or we call it a guided project. We can have 500 people be in that if we want to, and then they are working in subgroups with other people where they can get assistance in other ways. And so traditional educators are like, “Oh, you have a 1:500 teacher to student ratio. That’s bad.” And the reason that’s bad is because they have a mental model of, “Oh, if I want to get help, I have to wait in line for 500 people.”

Ben Grynol (13:54):

Yeah.

Austen Allred (13:55):

But really, no, your help is coming from a giant team of other people who are there instantly. And that’s something that’s very difficult to do in a physical environment, but it’s obvious in an online environment. And so we can break up and subdivide into classrooms instantly of people who are stuck on Section 3 of Unit 2 in the fourth problem set. Now they’re instantly together with someone who is focused just on that and working with them. And it’s way better than a single teacher running around to all the students trying to handle everything individually.

Austen Allred (14:34):

And we’re just starting to get to the next level of that, which is moving away from the notion that… We had something that we called Flex, which is if a student can’t complete a sprint or can’t wrap their mind around a unit, instead of just pushing forward and hoping they’ll pick it up along the way, they can roll back at no cost and repeat it. And that makes a ton of sense in our world in a way that wouldn’t necessarily make sense in a physical environment. So we had that, but we’re slowly moving away from that to, “What if a student wants to speed up? What if they know this thing already? How can they fast forward that?” Sometimes you can cut a month or you can cut two months off of the learning because they’ve already done that. Or “What if I move at a faster pace or what if I move at a slower pace? How can the school and the software automatically flex around that and just use the scheduling of instruction time and help in real time to adjust to all those people?”

Austen Allred (15:35):

So what’s kind of the next step is, I’m sure in the future we’ll have people that takes three months to get through the same content that it takes somebody else a year to get through. And then you can accommodate schedules that in the early days it was difficult to be on the East Coast because we ran Lambda School 8:00 AM to 5:00 PM Pacific, which meant your full time on the East Coast starts at 11:00 or noon and it ends at 5:00 or 6:00. Or sorry, at 7:00 or 8:00. And that’s really late. And then part-time was 5:00 to 8:00 Pacific or 6:00 to 9:00 Pacific, which meant some East Coasters for part-time were getting done at midnight. That was the way we structured things. Now we can break it apart. So whatever three-hour stint you can attend, it can be 6:00 to 9:00 on the East Coast and 6:00 to 9:00 on the West Coast and the shared time will be kind of at the crossover of those and we’ll have help on both sides of that.

Austen Allred (16:28):

So the flexibility, the adaptability, the ability to adjust to what a student needs at any given time, it’s just orders of magnitude better than it was in a physical classroom. And that started because we were trying to solve for all the “problems” that remote learning has. If you could build a school the way that Lambda School is building it, it’s actually detrimental to go to a physical class classroom if your goal is to learn things well quickly, which is super interesting.

Ben Grynol (16:58):

Yeah. I mean, you changed the approach, right? You focused on the incentive. The incentive is to educate people and put out great resources. That is the incentive. And so by doing that, you go, “What steps do we have to do?” instead of indexing on like, “We need to cover our real estate costs this month so we just need to fill this with people and get them out as quickly as possible so that we can get to max utilization again.”

Ben Grynol (17:25):

One of the things that we’ve been working towards is it’s similar to what it sounds like you’ve got with somewhat of a guided journey, which is like micro support where it’s like what you mentioned with like four people are working on Problem 4A, Section 1, like something so granular. And we find the same thing where we’re getting these micro communities that people come in with, A, a different understanding of metabolic health. Some people know a ton because they’ve done a lot of self education. Other people have no lens on it. And so, you need to sort of find the pockets in the groups that can support each other.

Ben Grynol (18:02):

But again, being geographically distributed, knowing everybody has different interests, we’re aware and we’re trying to build against that, saying, “What does a guided journey look like from a community perspective and then from a product perspective?” so that there’s like some overlap in the Venn diagram but still trying to get to that level of support that feels personalized as opposed to saying, like, “Here’s a CGM. Here’s Levels. There you go,” right? Because that’s just too generic given that we’re all… Everything about metabolic health is so nuanced and everything about learning is so nuanced.

Austen Allred (18:35):

I think that’s exactly right. It’s fascinating watching products that we kind of moved from a one size fits all model. If you onboard… We were talking about Slack earlier. The Slack onboarding is the same for every single person they’ll onboard. “Here’s channels. Here are DMs. Here’s how you adjust your profile. And good luck.” And you’re thrown into an environment that’s “Hopefully this works for you.” But I’m looking at a new kind of generation of products. So in the health space, I think about a Noom for example. Noom, it’s onboarding is more about getting to know who you are and adjusting the product to what your specific needs are.

Austen Allred (19:17):

So Noom, for those that don’t know, is it’s a weight loss app/community that tries to key off of a lot of psychology to help you lose weight. But there’s a big difference between “I’m 500 pounds and I’m morbidly obese and I eat 5,000 calories a day” and like, “Yeah, I’m 10 pounds overweight, but I can’t shed the last 10 pounds.” And there’s difference physiologically male versus female. There’s a difference in eating habits. There’s difference in exercise, you know?

Austen Allred (19:45):

And so building a product that can accommodate all of those people at once is difficult and expensive from a product standpoint. But then when it’s there, it’s so much more powerful than like “Here is how you count calories. We’re going to take a wild guess at how many calories you should be eating a day” or, “Fill out this form and then we’ll spit out ‘Okay, your calorie intake should be X and your macro should be Y’.” Versus one of the things that makes me really excited about Levels is Levels will have a level of data that other folks don’t, right? Levels could, in the long run, be a completely different product based on if your blood glucose is in low levels and if you eat rice, it doesn’t matter, versus if I eat a bowl of granola, my glucose spikes. Whatever it is that you’re measuring and keying off of, it can react totally differently.

Austen Allred (20:43):

I think that’s one of the reasons I invested in Levels. And I think it’s really, really exciting, is not only can you get all of this new data, which that’s what we learned in the early days of Lambda School, like, “Oh my gosh, I can know how a student is doing every moment of every day.” Not only I can go well beyond like, “Are they raising their hand because they’re stuck or not?” I can know this person is a 92 out of a 100 at React and there are a 10 out of 100 at Python. Those two may not always happen at the same time.

Ben Grynol (21:12):

Yeah. Yeah.

Austen Allred (21:14):

If you’re not good at React, you’re probably better at… But anyway. I think at scale customization is really, really interesting. And I think some products have been doing that already where if I log into am Amazon, it’s a completely different experience than if you log into Amazon because it knows our history. I think that level, instead of shopping or consuming content, combining that with “What should I be doing?” I think about “What is my personal dashboard? What habits should I be forming? What behaviors should I be shifting?”, that’s incredibly powerful. I think health, to me, is one of the most exciting places to apply that because it is so customized. It has to be customized. And education is another one where my health is completely different than my wife’s health even if we’re eating the same thing every day. The way I experience an educational product is completely different than the way my wife experiences an educational product even if the content is exactly the same. So I think that’s kind of the frontier. And I’m really, really excited to see what happens with that.

Ben Grynol (22:21):

In product, have you been building loops that give people personalization so that it’s… Is there like you get as granular as you can and then indexing that with community to find that support? How have you been doing that as you’ve bringing more and more people through?

Austen Allred (22:37):

One of the things we’re working on now, if I said that we’re doing a great job of it today, I’d be lying. There’s actually a study in the 1980s by a gentleman named Bloom that basically showed… And so, they did a very intensive study on different learning approaches and different learning environments. So they had a traditional classroom environment and they would take students who had the same level experience, the same IQ range. They did all of the important work in selection. And then they taught those students a set of things in a traditional classroom. And then they had another set where they had a traditional classroom, but it was mastery-based. So you don’t move on from one subject to the next until you’ve mastered… until you’ve at least completed or demonstrated understanding in one thing. And then you kind of move on to the next classroom.

Austen Allred (23:28):

And then the third was one on one mentorship, right? So you have a full-time mentor who’s sitting there watching over your shoulder. The crazy thing about that study is they found not only was there a difference in the ability to learn, but the median student in the personalized mastery-based kind of the one-on-one environment, performed at the 98th percentile of the traditional classroom. So literally, the median student was… They call it Bloom’s 2 sigma problem. The median student in an adaptive learning environment was doing better than 98% of the other students in a traditional classroom. That kind of shift and empirically demonstrated is so crazy, right?

Ben Grynol (24:19):

Mm-hmm (affirmative).

Austen Allred (24:19):

Think about who the smartest 1% or 2% of people are in a traditional classroom, and you can get half the class to that level. That’s a massive, massive shift. So what is the response in the education community to that? The response is, “Well, that’s great, but we can’t afford to hire a personal tutor for every single student. That’s not feasible.” So really nothing changed. But what we’re seeing with adaptive software, I’ve seen a couple of schools that have done this really, really well, is not only… We’re actually seeing students with the right level adaptive software outperform students who have one-on-one full-time mentorship because the softwares can be even better at adapting to that student’s needs and isn’t trying to keep a mental log of what a student knows and can put back repetition of what they haven’t learned yet at levels that are better than a human.

Austen Allred (25:19):

We’re in the very early stages of seeing this in education, but I’ve literally visited elementary schools that are based on this adaptive level of product like learning, and they can easily move a student from the bottom 10 percentile of a public school to the top 10 percentile in a year. And we’re we’re now seeing students who… And this is K-12. We’re working on getting it into post-secondary. But we’re talking students who have a perfect SAT score and 5s on a bunch of AP tests and are done with the high school curriculum at 12 or 13. And so they have this problem like-

Ben Grynol (26:02):

[crosstalk 00:26:02].

Austen Allred (26:02):

… what do we do with these kids for the next five years, right? But not only is their performance superior to a private school, they’re absolutely crushing… A student with the same background, same demographic, same IQ level is absolutely crushing a kid at a $50,000 a year private school. And that’s one of those things that is not widely distributed yet. I think they’re maybe a handful of schools in the US that are doing that. But once that starts to proliferate, the delta in performance is so strong that I can’t imagine… I don’t know what it will look like, but the people that queue into that and start treating their education differently are just going to have wildly, wildly superior outcomes to the norm today.

Austen Allred (26:47):

A lot of that is because the norm today is just really suboptimal. It’s not because it’s revolutionary. It’s just a standard classroom environment is really bad. And it’s not because the teachers aren’t doing their best, it’s just the structure is bad. My wife was a 3rd grade teacher and it’s impossible to teach a random selection of eight-year olds division, teaching the same class of 30 people the same material. And they’re going to be on wildly different levels. And take that for every subject and that’s going to be true. Whereas if it was more adaptive to who they are and what they need, every single student would do far better. And so I’m really, really excited for a world where that’s true.

Ben Grynol (27:28):

The wild thing about that is the intent can be there. So let’s say from… Let’s use the analog of a teacher and a physician where it’s like the intent is there to do what is in the best interest of the student or the patient, but there’s always a certain amount of discretion or subjectivity. So let’s use a student where it’s like, “I think you should do your math again. I think we should spend some time on that.” Whereas tooling and an algo is just “Objectively, you are not at the place you need to be. Do more math.” And then you get to the point, right? You get this feedback loop with health, where it’s giving you that personalized feedback to say, like, “Keep going, keep going, keep going” until you get to the level where somebody has enough of a routine with what they’re doing whether it’s from a education perspective or a health perspective where that platform and that tooling helps to give you that feedback loop to say like, “You are on the right track.”

Ben Grynol (28:23):

I think using that insight of the students who became better than 98% of the rest of the class, it’s like you can do the same thing with your health too, where as soon as you have that information in your own hands and you have ownership over it, you can understand it. It’s like there’s so much more that you can do because tooling becomes scalable and becomes as objective as possible. Sure, we can digress into algorithms can be amazing and they can cause some major challenges too, but that’s a different conversation. The idea is that on average, let’s say algorithms can help you to perform better or to get to better places. That’s just infinitely scalable to start to bring everything up. Whereas, the bottleneck in education is teachers, like one to one relationships. The bottleneck in medicine is one to one to one relationships with physicians. So it’s like, “Okay, well, how can you solve the problem?” It’s like, “Well, tooling.” It becomes such an easy answer.

Austen Allred (29:19):

Totally.

Ben Grynol (29:19):

And then that’s where you have to index. But the hybrid is exactly what you’re building too and that’s what we’re trying to get too, is tooling plus community, because you still need social support and reencouragement and reinforcement that like, “Hey Austen, you’re doing a great job” because the reality is, we’re all imperfect beings as humans and we don’t operate exactly like robots. And so that’s where you need sort of like the overlap in that Venn diagram to say, “We have both components and you can index a little more on one versus the other.” Some people might be heavy on the tooling. Like let’s say with Lambda, all they need is a tooling. And then other people say, “Nope, I need that community support.” So it’s like you find what works for you based on personal interests.

Austen Allred (30:05):

Yeah. I think education went through this phase kind of 10 years ago where our understanding at the time… And maybe it’s too flippant to say our understanding. But if you ask people what they thought of as education, they would say like, “Yeah, lectures with really smart people. Good curriculum.” Somebody breaking down the curriculum. And so early movers in education realized, “Oh my gosh, we can take all of this and put it online and it’s pretty close to free,” right? So the Udacitys and the Courseras of the world said they created what was then the MOOC movement, the massive open online course where we can get a professor from Harvard to record his lecture and then it’s just like everybody’s going to Harvard. And it turns out it wasn’t quite that simple, right? Because the most important pieces of education weren’t actually the lecture. It turns out we’ve had libraries for a really long time, access to of the material itself… I mean, it’s way better than not having access to the material for sure, but that in and of itself isn’t necessarily what caused learning for the vast majority of students.

Austen Allred (31:13):

I think, similarly, the problem set of how do you get the right type of data and then how do you… Eventually, in the long run everything’s totally automated and totally responsive. And then in the meantime, you’re building up buckets of community or one on one assistance that is scaled. So I think of a future fit where, when I worked out with a personal trainer in San Francisco, he must have been the most bored person in the world because he’s watching me do bench press. And he’s watched 100 people do bench press. He’s sitting there full time to see if there’s anything that I do wrong so that he can correct my form, or if I say, “Oh my shoulder’s tired,” he can adjust.

Austen Allred (31:56):

But now after everything kind of went remote, I actually hired a trainer from Tonal. So Tonal is kind of wall mounted weight lifting system. And then I basically send him data from my eating and my Tonal and my Peloton and if I do other workouts. I give him all of the data and he can adjust in a one on one way to that. He was shocked that we actually have way better data on this than we did when we were in the gym, even when he was sitting there watching me full time. He doesn’t have to sit there and watch me full time anymore. But certainly, that versus me sending him a log of “Here’s the workout I did in a physical gym,” it’s just night and day difference.

Austen Allred (32:45):

And so I’m watching that kind of develop. And now there’s a community of a few dozen people that are doing that along with him, and like you get a community aspect. At the end of the day, I feel like I have… And this cost me a fraction of what a personal trainer used to cost. It’s now flexible. If one of us gets sick, I don’t have to like… Mike was my personal trainer in San Francisco. I’d have to text him. Or if he was on vacation, I just wouldn’t work out that week because I wasn’t at the point where I knew what to do yet. I think combining those things… Eventually, maybe that gets fully automated to where my Tonal will just knows exactly what I’ve eaten and it knows all of my workout then it’s going to say, “Hey, we’re going to swap out this Bulgarian split curl to whatever that is.”

Austen Allred (33:33):

But in the meantime, getting that data, getting people into communities, having scaled help is 10X better than the old way of you either had to go it alone or buy a book or pay somebody incredibly high rates to do one-on-one training. Rates didn’t have to be incredibly high. It’s expensive to do anything one-on-one with somebody sitting there watching you. So yeah, I’m really excited to see what that level of software looks like where… And it probably starts out. It’s adaptive enough to bucket you and then you have a human in the loop as necessary. And then over time, sure, it’s automated and it’s fully… The marginal cost is zero. We’re not at that level yet. But what we can start to do is have everybody the experience the outcome of a full-time personal trainer for 1/10 of the price. Or even in my opinion, slightly better outcomes than a full-time personal trainer for 1/10 of the price.

Austen Allred (34:34):

And that’s really, really cool. That’s the power of technology. That’s the power of something like a Lambda School. That’s the power of something like Levels, that when you can do that, it opens up entire worlds of opportunity that just were not there before.

Ben Grynol (34:48):

When you were designing the business model, did you receive pushback when people were like, “Man, that’s not going to work”? Because you were [crosstalk 00:34:57] stagnant industry.

Austen Allred (34:58):

Oh, yeah.

Ben Grynol (34:58):

Let’s just say the institution of universities’ didn’t change for 100 years. And then you go through the batch in the YC batch in ’17 and everyone’s like, “Man, what are you trying to do? This doesn’t make sense.” I ask because one thing that it comes up in conversation where we never want to be incentivized to sell CGMs, right? We never want to be incentivized to sell something to people. Our mission is strictly to educate the world about metabolic health. The more awareness we can create, the more behavior change we can make in the world by empowering people to really have ownership over not just their data, but information and spread that information. The byproduct is like, “We happen to have CGMs. If you want to buy them, great.” But the goal with it is we’re trying to align our incentives, like the goal being educate the world about metabolic health, which means strip away the economic incentive.

Ben Grynol (35:54):

So we receive not necessarily pushback, but people just think like it’s a wild model to take where we’re saying, “Hey, there’s a membership model.” And you did something from an economic perspective where it’s like, “We’ll take the risk. We’re going to do everything for you and then you can just pay us back once you get a job if you get a job.” It’s very, very different. So were there things that you moved the knobs or massaged the levers along the way to say, “Hey, here’s how we figured it out to make the economics work”?

Austen Allred (36:19):

Yeah, I mean to answer your original question, yeah everybody thought we were freaking crazy, right? Originally, I went to other schools and I was like, “Hey, you should do everything online.” And they’re like, “No, online doesn’t work.” Oftentimes, their incentives were, they’re looking at their classroom and like, “Hey, we should try this online thing.” And then they would automatically go all the way to like, “Well, for trying this online thing, let’s just upload everything that we do. Let’s film three months worth of classes and put it on a platform and let students go through it. We can charge half of what we charge and that’ll be it.”

Austen Allred (36:54):

And then they would do that and realize nobody would do it. You could get somebody to pay for it, but they weren’t going to go through all the classes because it’s just a soulless difficult journey to spend three months on your own behind a computer with no help and no community whatsoever trying to get through a very difficult… Basically code bootcamps tried to build themselves into a MOOC and then found that they had all the same problems that MOOCs had. MOOCSs being massive online open course as kind of what I referred to earlier.

Ben Grynol (37:25):

Side note. That acronym has always irked me. It is the byproduct of the late 2000, early 2010s. There’s just something about MOOC that I just couldn’t get behind.

Austen Allred (37:39):

Yeah. It’s funny that it was a hot space for a while.

Ben Grynol (37:42):

Right.

Austen Allred (37:43):

Yeah. At the end of the day it was like, “Let’s record all of our lectures and put them on the internet.” That’s crazy.

Ben Grynol (37:49):

You can do that MOOC… Everyone would refer to it and you’re like, “Why don’t we just call it online learning? That seems a little bit easier.”

Austen Allred (37:56):

Yeah, totally. And then now, I think we’re going through something similar. I mean, there’s a stroke of branding genius from some folks saying there’s a cohort-based course, which is really just saying it’s an online but live thing. It’s not a MOCOC. That’s what they’re really trying to get away from. Anyway. Yeah. So we’d go to the schools and they would say, “Hey, online doesn’t work.” I’d say, “Okay. I can’t prove you wrong yet, but I think you uploading all of your lectures does not constitute a good faith effort in making online learning work,” right? And they’d be like, “Yeah.”

Austen Allred (38:28):

That was always like the 10th thing on their priority list and so it never got the attention it needed. Nobody was focused on it. And it would just kind of go away. And then we said, “Also we want to do something that’s completely free upfront and you only pay if you get a job.” I mean, not only from a financial standpoint did people… Obviously, that’s there, right? Like, “Well, if only our graduates who get hired pay us, then that’s a very different economic model than every student who walks in the door pays us.”

Austen Allred (38:58):

But one of the things that schools were most concerned about was the way that I know a student is actually committed is when they write that $15,000 check. Until they’ve made a payment, I have no way of knowing whether they’re signed up for actually doing the work. If a student’s not signed up for doing the work, they’re not going to make it. And so, we spent a lot of time. That’s honestly been the most difficult piece of Lambda School, is figuring out who is there for the right reasons and who actually has the right level of dedication, because it’s so easy to sign up for something that’s free. MOOCs did experience that to a lesser extent where like, I signed up for a dozen MOOCs that I took 30 minutes of, and I was never actually serious, right? I wasn’t ever going to finish them. I was just dabbling.

Austen Allred (39:51):

But now for Lambda School, we’re going to put tens of thousands of dollars of investment into those as people, like, “What if they’re not serious?” That’s not just a Lambda School problem. Colleges have that problem. Y Combinator has that problem, where not every company that can get through the filter for Y Combinator is actually serious about starting a company. Some of them it’s like, “Hey, Y Combinator will pay me a year salary to do this little side project, and then I’m going to bounce.” I think for Y Combinator, it doesn’t really matter because if you have Stripe, then at the end of the day, you can fund everything else and it’s fine. But for us, that’s not our model.

Austen Allred (40:30):

So we found a few things over time, required pre-course work, does a really good job of filtering people out. We start students now in the first… And for us, the big thing is it’s not even will they do the work, the coursework in the school, it’s, “Are they serious about getting a job afterwards?” And we have had a non-trivial, if everybody who attended Lambda School was serious about getting a job afterwards, we’d be set. We’d be financially very, very sustainable. We’d be profitable. That hasn’t been the case to date. And I don’t want to share numbers, but the number of people who just fundamentally never looked for a job and were never intent on looking for a job was way higher than I would’ve guessed. I figured if you’re going to spend a thousand hours in a course, you’re going to go look for a job afterwards, but no.

Ben Grynol (41:23):

So what do people do? I don’t understand it. Like, it’s-

Austen Allred (41:27):

They just go back to what they were doing before. Or they’re like, “Yeah, this is cool. I’m really interested in it. I got a free education.” Or sometimes it’s less direct than that. Sometimes it’s, “Yeah, I’m going to take a few weeks off. I just finished this really grueling course, I’m going to take a few weeks off and then I’ll look for a job.” And then you reach out to him a few weeks later and they’re like, “Yeah, I’ll start it a few more weeks.” And before you know it, six months has gone by, they haven’t written a line of code in six months and they’re back at their old life, right?

Austen Allred (41:57):

The difficult part of Lambda School isn’t teaching people to code necessarily. It’s a lifestyle change. So we actually have to bet on people being able to make a complete lifestyle change. Some people are closer. But think about if a personal trainer only made money if you lost weight, they can have the best program ever. They can nag you all the time about making sure you’re working out. But then, if you’re eating bad on the other side, you’re not going to move the needle. So we’re discovering over time that there are ways we can filter for that. But in the long run, I think really it comes down to, if you can get a student to financially commit to do that, then we can offer way more. We can actually reduce the tuition when it does work. We can offer stronger guarantees.

Austen Allred (42:47):

So a light incentives have to be a two way street. And I think we started out by saying, “We’re going to guarantee everything and a student has to guarantee nothing.” We learned that if we can get just a little bit of skin in the game from a student, we can place pretty much everybody in that world. So we’ll have to see how that plays out in the long run. But depending on when this podcast comes out, there may be innovations that have launched by then where we’re saying, “Look, actually, we can make your monthly payments really low and we can make your cost of tuition even lower than what we charge today if you’re willing to sign up to look for a job” and base what they pay on whether or not they look for a job. We’ll have to see how that turns out. But in many ways, that’s the final problem for Lambda School. And if we can solve that, then everything else is… We will still keep improving on everything else, but that’s the last problem to solve.

Ben Grynol (43:45):

Interesting. That’s almost like taking the behavioral economics lens where you’re like get them to commit to something, which is just like, literally sign your name that says “I will search for a job one and say complete the coursework.” And just by giving that nudge up front, that might change behavior in the long run.

Austen Allred (44:04):

The classic Peter Thiel, like “What’s something that you believe that is not widely held by other folks?”, one thing that I have noticed that this is super… You’re not supposed to say that in an education space. But extrinsic motivation is a blunt instrument, but it is very, very powerful. Oftentimes, 10 times more powerful than intrinsic motivation, which is, I think intrinsic motivation is incredible when it’s there, but it’s very difficult to stoke. And we’ve seen extrinsic motivation absolutely works. I think educators are one example that’s a little bit silly. But a school that I visited had… It was a private school that was pretty expensive. They spent years trying to figure out how to get students to put away their phones and be alert and attentive. And they viewed it for a long time like “This is our challenge. If we’re not more engaging than whatever’s on TikTok, then we need to make the school more engaging.” They’ve spent long time working on that.

Austen Allred (45:11):

And then they tried an experiment once that was literally, “If you agree to put your phone in this box when you come in the door and not touch it until you leave, at the end of the week we’ll buy you a Chick-fil-A chicken sandwich.” Participation of not… Phone usage went from everybody using a phone to nobody using a phone overnight. Is that a blunt instrument, that $3 chicken sandwich from Chick-fil-A in order to cause students to do that? Yes, but then you can watch outcomes shift in a way that I think educators would be nervous to understand how much motivation a $3 chicken sandwich can provide a 12-year-old. Is that a blunt instrument? Yes, but at the end of the day, I’m pretty outcomes-oriented. And if giving a student a chicken sandwich will make them way better at math, I’m not morally opposed to giving them a chicken sandwich. That’s something that I had underestimated, the importance and the power of extrinsic motivators versus entirely intrinsic.

Ben Grynol (46:10):

You got to get Dan Pink or Dan Ariely to do a little bit of researcher around that one.

Austen Allred (46:15):

Yeah.

Ben Grynol (46:16):

The funny thing is though, with extrinsic motivation, if somebody’s earning $30,000 a year let’s just say, and they see all these community stories of people, the reality is engineering jobs command at all levels. They obviously command different amounts, but at entry level, they’re still commanding such a big delta between $30,000 and whatever that and is depending on the company that you go to, whether it’s a startup or we’ll call it like a blue chip tech company or something in between. The reality is, the delta is just so big that even if you are intrinsically motivated to do the work, it’s like if you could go from making $30,000 a year to $40,000 by going through Lambda School, there wouldn’t be enough extrinsic motivation for people to be like, “Man, I’ll do that” because they’re like, “My life isn’t going to change that drastically.” It’s that they can see that they can go from…

Ben Grynol (47:08):

And that’s what I love that you were doing back to the, if you can get students to write a check, is that you are finding ways to reach underserved communities. And it’s something that we think about a lot where right now the price of CGMs and a lot of health related products are just, they’re inaccessible to many people, especially the people who need the most. Same thing with, if you’re talking about underserved communities that might not have opportunities to be exposed to, let’s say a tech scene, and you’re saying, “Hey, here’s Lambda School. You can get in it.” I don’t know if you’re still doing this where you’d say, “Just put up your hand. Let us know if you need the computer and we’ll help you out.” But that was kind of the beautiful thing where you’re saying like, “Let’s strip away all of the things, all of the barriers, and just give you opportunity if you’re intrinsically motivated enough to do the work, to find the job.”

Ben Grynol (48:00):

But that was such a beautiful thing. And it’s something that we think about a lot of. As we scale, how are we going to continue to make sure that health is not a privilege and health is equal opportunity? I think the same thing goes with education. Education should not be a privilege. Education should be equal opportunity, right? If you want to do it, you do it.

Austen Allred (48:20):

I think for you guys, it’s easy to underestimate the extent to which that’s true today. I know people who spend $6,000 a month going to a doctor and getting all of their labs run and they are perfectly in line with everything and they can adjust whatever they need to. One way of thinking about Levels is it gives that level of data to somebody for a tiny fraction of the price, right? And it’s not free yet, I’m sure. It may not ever be for free or it’ll have to be someone will have to pay for it somehow, but that democratization is happening, right?

Austen Allred (48:56):

When I’m looking at Levels, I get data that is better than someone who wants to pay a doctor. I’ve looked at some of these, “Hey, I’m going to do a test every week and monitor all of your vitals.” It can be $5,000 a month. It could be $10,000 a month. That’s not crazy at all. And there are some doctors who only take a handful of patients and they monitor you like a… I mean, it goes beyond what a personal trainer would do, right? But there’re a lot of people that do that. Levels is giving that level of data away for a fraction of that. A tiny, tiny fraction of that price. I think that’s some of the opportunity.

Austen Allred (49:33):

On the incentive side, I think it’s also important to… And this is something we’ve learned. In the early days, if I would talk to a student and say “Look, our median income is X. If you do this and you do the work, you’ll be able to get a job for X.” There were people who literally did not believe me. One of the earliest hires ever was working in a factory before Lambda School. He was an African American gentleman, a single father. And he came to me and said, “I really want to learn how to code, but I know I’m not going to be able to get a job for more than 50K a year. So I just want you to understand up front that you don’t have to continue the charade with me. I understand that the way the world works is…”

Austen Allred (50:22):

I think he either had a GED or had not completed high school or something like that, but he was one of the top performers on our pre-course code challenge stuff. He’d been really interested in kind of learning on his own for a while and he was just kind of stuck. He ended up getting a job for $85,000 a year. Now I’m sure he’s… I haven’t checked recently, but I’m positive he’s well into the six figures now. A completely life-changing shift, but part of it for him was that he had not been exposed enough to literally understand that that was possible. And I think with wealth in the United States, that’s often something that happens where until you’ve met somebody who has done X, you don’t realize that that is a thing.

Austen Allred (51:11):

I mean, I still think about I grew up in a lower middle class town where I thought a six figure job was just FU money. That was such an unfathomably large amount of money. And you kind of get exposure over time to people with different levels of wealth and you look at their career paths. Now I’ve met billionaires. And far beyond thinking that that’s a possibility. I didn’t realize when I was a kid that there were billionaires who were just tech founders who had a really good idea and it hit and they worked really hard. Growing up, I thought you had to be evil or some crazy, super genius… I didn’t know that that was possible or normal. I’m not saying that billionaire is something that’s easy for anybody to aspire to. Million things have to go right for that to be true.

Austen Allred (52:02):

But I was talking to one of our investors about how do I convince these students that these jobs are real, and he was like, “Well, what if I told you, I came to you when you were 12 and I said, ‘Hey, I will give you $15 million if you can place in the top 10 at Wimbledon’. I’ll give you all the resources you need. I’ll give you all of the training. I’d run a test on you and say, ‘Look, you have the athletic ability to place top 10 at Wimbledon’. Would that be easy for you to do?”

Austen Allred (52:31):

And obviously for me, no. Yeah, I played a little bit of tennis growing up, but top 10 in Wimbledon is like you have to be an absolute world-class out of this world performer. And he was like, “Well, that’s how it feels to some of the students you’re talking to saying ‘You can end up making six figures’. And so the promise of, ‘Hey, spend a year of your life working on this. And then if you don’t get it, I’ll give you a refund,’ that feels good, but it’s far enough away. And if you’ve never seen it, it’s so difficult to believe that you need to do a better job of breaking it down into smaller chunks. Whereas as an example, if I told you, ‘Hey, I’ll give you $1,000 if you play tennis for 40 hours this week.’ As a 12 year old, would you have done that?” Absolutely.

Austen Allred (53:22):

So some of it’s… I don’t know. It’s interesting to break down… At Lambda School, we know what it takes to get a job. I can tell you exactly what you need to do in order to be hired as a software engineer. Even though we have not charged anything until you get there, it’s still difficult for some people to believe. And so, one of the things I’m also interested in is, can I put more skin in the game in some way and break it down into smaller chunks? Or put even more on the back end of, “Look, if you do all of this stuff that I’m telling you is required to get a job and you don’t get hired, I will literally write you a check.” I don’t know that I’ll be able to pay you minimum wage starting out. The math may not add up, but I can take the risk out. I can say, “I’ll pay you a couple thousand dollars if it doesn’t work out.”

Austen Allred (54:07):

There’s some regulatory reasons that make that difficult, but how can I make it so that you don’t have to believe me? How can I make it so that you are just going to do the… How can I take the risk out not just from a “Do I have to pay anything?” perspective, but, how can I continue to de-risk where it is now just so obvious that I need to do these simple things and “Even if it doesn’t lead to the outcome that Austen and Lambda School are telling me it leads to, I’m going to do them because either way I end up in a better spot.”

Austen Allred (54:39):

So that’s kind of going beyond what we’re doing now to… At the end of the day, Lambda School will have more than a 100% guarantee. It might start with 110%, it might get to 150% of tuition, where if you do all this stuff that we tell you to do and you don’t get hired, maybe we’ll write you a $10,000 check one day. And then maybe that eliminates all of the risk and you just do what you need to do. That’s a level that we’re trying to get to now. I think that is incredibly interesting to me. How do you get to that next level of de-risking? And at the end of the day, it comes down to, if you know that action X causes outcomes Y, how can you incentivize people to do action X? That’s a fundamental human problem that I think is really interesting.

Ben Grynol (55:28):

Yeah, because you’re filling a gap that is massively needed in the world, which is just like, there’s a… We don’t have enough software developers. We need more people to be a suit in all aspects of engineering, right? Data science, like name it. And it’s like we need to be able to empower people to do this. And so, the more that you can surround people with other people who’ve seen it firsthand and they go, “I can believe this,” the caveat to it is it’s really hard to believe. Sometimes you’ll tweet out things and you’re like, “I’m finding it hard to believe this right now.” Even though it is true, it’s just… I’ll probably misstep on the exact nuance of the tweet, but I’m sure you’ve shared tweets that have said “The delta is $192,000 in their previous salary versus their starting salary.” Something absurd where somebody… Like you’re sitting there going, “I know this happened, but I’m trying to figure out how can I believe this is happening” because it just is happening.

Ben Grynol (56:27):

And that’s where you’re starting to create this movement where it’s like, the more people that can see like, “Hey, this is real.” Bring it into health and wellness, same thing. When people are seeing not that CGMs are a weight loss or that we encourage people to use them for weight loss, it’s just the byproduct of being able to manage your own health and wellness and understand that data, there are members who have lost 100 pounds, 100 pounds. And people who’ve spent 10 or more years, 20 years in some cases trying to lose weight and all of a sudden they go, “Wait. It was just a matter of me having enough data to truly understand that feedback?” Same thing with what you’re saying with your trainer where it’s like, it doesn’t matter how much your trainer is watching you do bench presses. It’s like there’s the actual data to be more objective about “What was breaking?” and like “Let’s fix that.”

Austen Allred (57:15):

Yep.

Ben Grynol (57:16):

It’s such an incredible thing-

Austen Allred (57:17):

We’ve been flying blind in so many… I mean, especially in nutrition and health, I mean we’re flying so incredibly blind that what was it? 20 years ago we had the food pyramid, which is just flat out wrong. I mean, it’s not surprising to me at all that there are people that once they’re able to analyze their data in real time, realize… Nutrition is even more so because it’s so unique to each person. My metabolic response to one food, the exact same amount of food will be completely different than my wife’s even if we work out exactly the same… Anyway, you get what I’m saying.

Austen Allred (57:53):

I think part of the power of what we’re doing in getting more data is eliminating the uncertainty to where I know exactly when one of the reasons losing weight is so hard is because you’re not actually sure at any time, right? And not just losing weight, like any aspect of health, where you’re not sure if anything is working and the change can be so incremental over time. It can take you a month to figure out if what you’re doing is actually making a difference.

Austen Allred (58:26):

So getting that into more real time, more interactive data, it can help people understand that “If I eat X versus Y, what is the delta between those two things? What is my fork in the road actually like? How will my life change based on whether I eat a bowl of ice cream or a hamburger,” right? For some people, eating a hamburger, maybe totally fine. And for some people, eating a bowl of ice cream, may be totally fine. And we just don’t know. So I think the empowerment that is unlocked by giving people knowledge and information about what the world looks like is… As investor in Levels, I think it’s a very noble mission, and I’m glad to be a tiny part of it.