EB-6: Umax - Transcript

[00:01.298]👩‍🎤 Ate-A-Pi: Hey Blake, tell me about UMAX.

[00:05.31]🤵 Blake Anderson: Yep. UMAX is a facial scanning application that allows young men to take pictures of their face and get ratings and recommendations on how to become more attractive. I think to do the story justice, this starts back a few months ago. I previously worked on an application, Riz GPT, now named Plug AI.

[00:35.741]🤵 Blake Anderson: and
I noticed that there was a major gap in the male self-improvement space. When we look at men's mental health issues, that sort of thing, men are increasingly lost in today's society. And so I started scanning social media and I was like, okay, where is there a wedge? Where's a good place to enter?

[01:08.866]🤵 Blake Anderson: Um, but the issue with looks maxing is that the information that people are getting is

[01:17.418]🤵 Blake Anderson: It's not always high quality and it's not always personalized. So that was kind of the original thesis in setting out to build UMAX. We launched a little bit of a.

[01:27.33]👩‍🎤 Ate-A-Pi: So just to interrupt you there, what is looks maxing? Because I think, okay, just to, I think, you just graduated from Tulane like last year, correct? So I think there's this whole section of social media, which I think most ML AI people are really like, you know, not exposed to at all. So what is looks maxing?

[01:53.326]🤵 Blake Anderson: Yeah, looks maxing is essentially a term for male beautification or trying to maximize your looks. This is a practice that women have engaged in for decades through methods like skin care routines, eyebrow grooming, taking care of your hair. But for the most part, men don't really practice any of those. Men, you know, obviously working out is really big. But beyond that, guys don't really talk about taking care of their face.
and their physical appearance all that much. In the past couple of years, it's risen to a lot of prominence on social media because guys are starting to realize that there are things you can do to become more attractive beyond just working out and beyond just advancing your career. And yeah, it's a super fascinating, I think, macro tailwind in terms of
what social media has done to people's perception of what it means to be attractive and the importance of it.

[02:57.354]👩‍🎤 Ate-A-Pi: So if I can, are, do people basically take pictures of themselves and kind of, you know, compare or ask for ratings or something like that? Like is that, is that something that they do?

[03:13.798]🤵 Blake Anderson: Yeah, so across various online forums, Reddit, looksmaxing.org, Discord, people upload pictures of themselves and ask people to rate them and to give them recommendations on what they can do to become more attractive. I think we all know that what we see in the mirror isn't necessarily what other people see, right? When you take an inverted picture of yourself or a selfie on your camera, when it flips, you're like, wait, that's what I look like?
So this kind of signifies that we aren't great assessors of ourselves. And so that was one of the original reasons that we thought that we would find product market fit with UMAX. It's like, how do we leverage AI and ML vision to deliver this to people instantaneously and actually...

[04:09.616]👩‍🎤 Ate-A-Pi: So if I open up the UMAX app, it's iOS and Android, or only iOS? Both, iOS and Android. So it's iOS and Android app. You open it up, and you take a photo of yourself from the front and from the profile from the side.

[04:15.746]🤵 Blake Anderson: Yes.

[04:28.25]🤵 Blake Anderson: Exactly. One photo front, one photo from the side.

[04:31.99]👩‍🎤 Ate-A-Pi: one photo from the front, one photo from the side. And after taking that photo, you get a rating. And the rating has an overall, a potential, a jawline, cheekbones, skin quality, masculinity, and then there's a tap to share with your friends.

[04:50.752]🤵 Blake Anderson: Yes, correct.

[04:52.442]👩‍🎤 Ate-A-Pi: Right, so in the backend, so you're sending these photos to a backend. What's happening in the backend? Is it a classification? Is it, you know, what's going on?

[05:04.83]🤵 Blake Anderson: Yeah, great question. So we have a combination of ML Vision as well as AI Vision models. So when I say AI Vision, I'm referring to things like GPT Vision. And what we did early on was we found a number of open source ML libraries that could essentially detect whether or not you had acne. That's an easy one to detect. But then certain things like detecting an individual's hairstyle, that's not really easy to do with ML Vision.
but AI vision models do a pretty fantastic job. So that said, we have a ton of different features that we're classifying. The presence of acne, what shape the individual's face is, how strong is their jaw contour, right? And then based on this set of a hundred values, we are generating ratings, as well as recommendations, right? So if an individual has the presence of a lot of acne,
That's something that we are definitely going to recommend the individual to take care of. So we have a prevent acne routine, which involves applying benzoyl peroxide and salicylic acid, right? Proven ingredients. And yeah, that's generally how the actual back end architecture works.

[06:26.775]👩‍🎤 Ate-A-Pi: So how long did you guys take to put it together? So how many people are in the firm right now? Just.

[06:34.218]🤵 Blake Anderson: Mm-hmm. So initially it was just me. I'm somewhat of an engineer, a bit of a designer, as well as a marketer. And so I designed, built, and launched the initial product with part-time help from two other people. That took me a little bit over a month. And then from there began to scale up. I brought on a CTO, someone who's...
real expert in full stack engineering as well as ML, brought on a full-time person to handle marketing. And so now we have essentially three full-time people, or yeah, three people full-time and three people part-time.

[07:19.014]👩‍🎤 Ate-A-Pi: and you're up to, I think, like 3 and 1 half million downloads, something like that. And.

[07:24.563]🤵 Blake Anderson: Yeah, so in the past three and a half months, we've done over three and a half million downloads and over one and a half million downloads.

[07:31.73]👩‍🎤 Ate-A-Pi: So one and a half million revenue. So something like, I saw some number was like, something like six million for the year or something like that, like estimated.

[07:39.442]🤵 Blake Anderson: Yeah, we're at six million ARR. So essentially we're doing $500,000 a month right now. These are recurring subscriptions. So I guess an important note to add here is that the way the app works, you do your scan, you get your ratings and recommendations, and then you can do a scan in a week to come back and see how you have progressed and get updated ratings and recommendations.

[08:04.134]👩‍🎤 Ate-A-Pi: And when I go to social media, I see a lot of people with kind of like their results, right? Like they've used the app for a few months, and they've taken the recommendations on, and you see some dramatic changes in results. So have you received a lot of positive feedback from people? Do people send you like?

[08:25.439]🤵 Blake Anderson: Yeah, see.

[08:30.458]👩‍🎤 Ate-A-Pi: you know, videos and you know, do they send you like, hey, you know, this has changed my life? Like, do you see things like that happening?

[08:37.778]🤵 Blake Anderson: Yeah, yeah. So what they refer to them as are glow ups, right? And we call the recommendations that we have your glow up routine. So people will use the app, they'll start a skincare routine, maybe when they previously didn't have one, they'll start taking care of their face in ways that they didn't before. Maybe they let their beard get out of shape. Maybe they had a bit of a unibrow, right? Things that people generally find unattractive, they'll start taking better

[08:42.538]👩‍🎤 Ate-A-Pi: glow ups.

[09:07.374]🤵 Blake Anderson: glow ups organically on social media, and they'll reference the UMAX app, which is, it's been pretty awesome to see. And I think that one of the critiques that we've heard is like, do you think that your app is creating insecurities in people when you're assigning a rating? And that's certainly a valid concern and something I've thought deeply about from the start. The sort of feedback that we've received, and what I've noticed is that
I think that insecurities are often just desire for change. Insecurities can be decelerating. They can make us anxious. Maybe we become paralyzed, but on the flip side, they can become accelerating forces that push us towards where we want to be. So when I was younger, I was very scrawny. And now that was at the age of like 14, 15, but.
Since then, I've been lifting almost every day for my life. And I'm incredibly grateful that I was able to recognize early on that I wanted to be in good physical shape. And so I think that this same sort of mindset applies to a lot of our users.

[10:19.699]👩‍🎤 Ate-A-Pi: Um...
So, as you said, what is on the flip side? On the flip side, do you receive, kind of like, I'm very disappointed in this, like, what is this app doing? Like, I am not, you know.

[10:37.522]🤵 Blake Anderson: Yeah, what's really interesting is that we almost exclusively receive that feedback from people who aren't part of this community that we're targeting, who haven't used the app, right? What's really interesting is that when someone gets a good rating, they're oftentimes really happy, right? And then they're going to continue to work hard. But what we always see is that when people get bad ratings, they will send us emails and be like, hey, I just want to say thank you so much. Now I know what I need to work on and I'm going to use this to...
to push myself to work harder than ever before. And I think that the looks maxing community as a whole, there have been some bad press and bad articles about it. But what's really interesting to me is, if you're someone who's an active participant, you'll realize that it's almost never about shaming other people for being bad looking. I've seen that, I can count on my hand the number of times I've seen that.
It's almost all about like what can you internally do to improve yourself as an individual?

[11:40.37]👩‍🎤 Ate-A-Pi: Yeah, so if you look at it, if you look at the LookSmaxxing community as a community of self-improvement, kind of like the continuation of the Jimbrow movement, right?

[11:52.242]🤵 Blake Anderson: Exactly. What's super, super interesting here is that on social media, what we've kind of seen is weightlifting and the gym bro movement has kind of had its heyday. It's been on the decline over the past couple of years relative to a few years ago, but looks maxing has kind of been taking that place. And so, yeah, I think it's a super interesting shift.
across tens of millions of people.

[12:27.295]👩‍🎤 Ate-A-Pi: So it's interesting for me in a couple of ways because I think in Asia, I think there has been a lot more consciousness, especially in Korea on like male beauty standards and a lot more kind of like work on, how do men look better, et cetera, et cetera. But in the US, it has slowly percolated, it's not yet.
I think it's finally arriving here. And it's interesting to see that it's arriving in a particularly American way, which is the self-improvement movement. Because that's not the way that Koreans or Asians don't focus so much on improving themselves. It's not so much of a, hey, if you work hard at it, you'll improve yourself, et cetera.
But I think it's very funny that it's arriving in that self-improvement kind of movement, and almost as a continuation of the Jimbrow movement.

[13:29.046]🤵 Blake Anderson: I think it's, there's an element of like the American mythology, right, that is so pervasive across everything we believe, right, that if you work hard, you can achieve a better life, right? Sort of the American dream. And it seeps its way into like, it's the underlying principles for so many of these movements.
exact same goes with Jimbrows, same goes now with looks maxing. And if we look at a lot of the content on social media that young men engage in, self-improvement is arguably the most pervasive niche.

[14:18.106]👩‍🎤 Ate-A-Pi: Right. Let's talk a bit about launch day. Right. On the day that you launched, did you do any product testing before that? Did you launch to your friends as a test flight app? Because at some point, it took off. The time before it took off, what was that initial launch feeling? Right.

[14:39.662]🤵 Blake Anderson: Mm-hmm. So, like I said, I built it in a month.

[14:47.791]🤵 Blake Anderson: Leading up to that, I had...

[14:52.914]🤵 Blake Anderson: I had dove or yeah, I dove super deep into the niche itself. Right. I became an active participant. I started engaging in these communities. I needed to think like one of the users in order to.

[15:04.958]👩‍🎤 Ate-A-Pi: So was it on Reddit or Discord or both or?

[15:09.014]🤵 Blake Anderson: Yeah, Discord, Reddit, TikTok, Instagram, everywhere. And so I'm a big believer that like build something for yourself, right? And at the time, like I wasn't really an active looks max, right? I never really had a true skincare routine. So I dove deep and I was like, okay, I need to become one of these people and I need to think like them in order to build the best product for them.
And so that, in addition to the development, those occurred in parallel over the course of a little bit over a month. Then we launched in the start of December. And one of the key strategies that I had during the launch was I wanna test all the different angles that we can have to portray this application. Do we double down on the ratings in the recommendations?
in the centralized repository of information that we had, right? Where you could click on hair and then click on improve your hair quality, right? Um, and then, so we really just started like split testing, like, see what would work, what didn't work. And over time we doubled down on the things that did work. Um, and we slowly trickled up for about a month and a half until we popped. And, uh, then people started sharing organically with their friends.

[16:38.083]🤵 Blake Anderson: And we were doing like $10,000 a day in revenue. And then we went up to almost $20,000 a day in revenue within the course of a day. And then we've remained there. So that was kind of like the first month and a half, two months.

[16:57.063]👩‍🎤 Ate-A-Pi: But even getting to 10k, I would say like 9.9% of apps don't even get a 10k a day. When did you get that first feeling that heading up to the 10k, even before that?

[17:10.398]🤵 Blake Anderson: Yeah, so I had worked on a previous app, Riz GPT, now named Plug AI. And that taught me a lot about what it meant to build a novel app. Now at the time, there was actually an incumbent Riz app, but they hadn't really reached that inflection point. And so a big thing that I think about is building something truly novel.
I think that one of the biggest issues with most application developers is that they make like a 1% improvement on an already popular product. Or maybe they try to combine two popular products and they squash them together and they think that will be their, that will lead them to long-term success. My

[17:57.862]👩‍🎤 Ate-A-Pi: And PlugAI was a cut and paste kind of like chat recommender. Like if you're using, if you're chatting with someone on a dating app, you kind of cut and paste the conversation into PlugAI and PlugAI would give you a suggestion, right? Is that right?

[18:10.802]🤵 Blake Anderson: Yeah. Yeah, it's a little bit more than that though. I think of it as an AI dating assistant where you can also get pickup lines, right? I don't think any other app had that at the time. We had an AI dating coach, right? Where you could talk to it and get recommendations. And then, yes, you could take a screenshot and get recommendations. But we did a lot of novel things in Plug that hadn't been done before. So the pickup lines, the dating assistant,
ability to adjust the spice of the response, that's something that I think is actually really big. Both of these bring me back to the point that I think that the reason that 99.9% don't get to $10,000 a month is because most people don't build something truly novel. They build like a 0.1% improvement.

[19:08.368]👩‍🎤 Ate-A-Pi: Right, an incremental thing which doesn't really stand out in any way.

[19:16.082]🤵 Blake Anderson: Like a great example here is like a social workout app. It's been tried like a hundred plus times. Another good example is actually the first app that I ever built. It was a college marketplace app. It's been tried a hundred times, right? And if something's been tried a lot and failed, it's usually not a great sign.

[19:39.734]👩‍🎤 Ate-A-Pi: Um, what is the wildest, um, or most unexpected use case that you've heard? Like, has someone come to you and said like, Oh, I got a, I got a, you know, spontaneous marriage proposal, you know, I got, you know, some, some crazy, some crazy story, like what's the, what's the wildest thing that you've heard?

[19:57.502]🤵 Blake Anderson: In regards to UMAX or Plug.io? Yeah, I think that some of the funniest ones are people trying to rate their dog, right? Or people will rate their friends. That's always pretty funny to me. That wasn't something that we expected from the start. But to be honest, it's...

[19:59.222]👩‍🎤 Ate-A-Pi: And yeah, you max.

[20:25.286]🤵 Blake Anderson: Yeah, there's not like that many crazy things that you could use it for per se. Personally though, I think that the craziest story is having met like a number of people in person that have used the app organically. That's always really shocking to me when I mentioned that I built it and people are like, no way, wait, like I use that app. It's awesome. That's always a really cool feeling.

[20:53.286]👩‍🎤 Ate-A-Pi: Yeah. Are there, are there UMAX like meetups? Are there, are there meetups that people, people do?

[21:00.455]🤵 Blake Anderson: No, not really though. I actually went on a trip with some of our influencers, some of the social media guys that we work with to Hawaii to make content and one of them posted on his story and he got like 50 people DMing being like, hey, like we should meet up. I'm in Hawaii right now. But no, they're not in person you Max Meadows.

[21:24.258]👩‍🎤 Ate-A-Pi: Um, is it a crowded space? Are there, are there a lot of people, are there a lot of copycat apps now, now that now that you've been, you know, you led the way.

[21:31.206]🤵 Blake Anderson: Yeah, yeah. There have got to be at least 20 apps that have copied exactly what we did. There's one app in particular, I'm not going to name, but they copied us and they have been relentlessly trying to steal our creators, trying to poach our employees.

[21:40.264]👩‍🎤 Ate-A-Pi: Mm-hmm.

[21:56.418]🤵 Blake Anderson: have been saying that we copied them when you can look at the release date. And like, it's very clear that we did not. And so that's been a pretty difficult battle in the sense that I have like strong moral principles, right, that I will not forgo, whereas they don't seem to be operating with many. So just a small example here is like, there are these little hacks that you can do to try to get someone's Instagram
You can have like mass bot reporting. They tried to do it to us, to my knowledge with them. I haven't tried to do it to them, right? And that's just not something that I would do as an individual. They tell creators that we're scammers and that we're frauds. I don't do that on the flip side, but clearly we aren't scammers. So those have just been definite challenges, but they haven't really been able to.
steal much traction from us, so we're all good.

[22:57.974]👩‍🎤 Ate-A-Pi: How do you market? What is your main, you don't have to reveal any trade secrets, but what is your idea of what it is to market such an app? What is your, what do you need to do? Yeah.

[23:09.138]🤵 Blake Anderson: Yeah, I think that it's cliche, but it's true. The best way to market an app is to build an app that people organically want to share with their friends. And so there are a couple of ways to do that. One is to figure out some little k-factor optimizations that are just going to keep picking it up and up. I think Laps is a good example of a company that did that.
On the flip side, something that we do is we just think about what is a very inherently viral feature and how do we package that and display it in a way that people are going to want to share it with other people and people are going to want to tell their friends to come download because it's really funny or fun or unique. That I think has been the biggest. That's number one. Then number two is...
you know, social media distribution and just like going wide, like we're on every single platform. If you search up UMAX, you'll find us Reddit, TikTok, Instagram, Facebook, whatever it may be.

[24:22.314]👩‍🎤 Ate-A-Pi: What do you, would you ever do a generative thing? Would you ever like, you know, display what someone would look like if they didn't have acne, for example, right? Like, is that something that you would think about doing at some point?

[24:35.578]🤵 Blake Anderson: Yeah, we actually have a feature like that. Right now it's not, so right now we're using like a, actually, I can't say the actual model that we're using because we do have competitors that copy this feature from us, albeit their version is a lot worse. But we have an AI generated like version of the individual that we call it like you as a 10 out of 10, like what you would look like if you were a 10 out of 10.

[24:38.174]👩‍🎤 Ate-A-Pi: Mm-hmm.

[24:47.376]👩‍🎤 Ate-A-Pi: Mm-hmm.

[25:02.87]🤵 Blake Anderson: Something that we thought a lot about doing too though is like with each recommendation, we can show the individual in the picture and then the individual implemented that recommendation. So with like prevent acne, if we identify a few spots here, we could clear it up in the image. I think that could be pretty powerful.

[25:24.446]👩‍🎤 Ate-A-Pi: How do you decide what kind of features to put into the app? I mean, you probably have a list of hundreds and hundreds of ideas. How do you make a decision on what stays in, what has to be developed, and what doesn't?

[25:37.906]🤵 Blake Anderson: Yeah, I think that this is another good example of you really can't know unless you are one of your users, right? If you're just trying to objectively analyze what to put in, you will have a list of 100 and you won't have any way of really assessing other than just implementing it and split testing and seeing the results. But
Being someone who engages in this community and has tried to, to some extent, embody the community, I think that, I just think about like, what I would want in the application, like what would make me wanna use it every day, and then add that in.

[26:27.082]👩‍🎤 Ate-A-Pi: How would, are there any natural extensions that you would think about, for example, does this naturally progress into dating or does it progress into sales of beauty products or recommendations of beauty products? What do you think naturally kind of plugs in or is that something that you would look at?

[26:37.986]🤵 Blake Anderson: Certainly.

[26:42.368]🤵 Blake Anderson: Yeah.

[26:57.042]🤵 Blake Anderson: Yeah, 100%. So this is what I had briefly touched on in the beginning, that we kind of identified this wedge in the male self-improvement space. And that was the thesis for launching UMAX. Now, the long-term vision is not just UMAX in its current form. And it's not just purely in this looks maxing space. The way that I kind of view self-improvement as it exists today is there's like somewhat of a triangle.
We have attractiveness. We have weightlifting, right? This is physiological, your body. This can include running. So this is kind of like exercise and how your body looks and feels. And then we have this motivation, disciplined mindset sort of corner of it. Right now we exist exclusively in this attractiveness corner. But what I'm really interested in building is a male self-improvement.
that embodies all of this. This would be a brand that offers full service, full stop solutions for every one of these. One of the biggest questions is, what is the distribution of applications look like here? Do we try to package this all into one application? Do we have a few different applications that are all by this UMAX or whatever new name it may be brand? And then...
step further beyond the mobile applications is, as you had touched on, physical products. UMAX is a good example of an app that you might not necessarily need to use every day. You scan your face, you get your recommendations, you move on, maybe come back in a week to see how you progress. But it is a UMAX skincare serum. That is something that you could use every day.
So I don't know exactly what the future holds in terms of like next, like what the exact next steps look like, but I think I understand what the big vision looks like here.

[29:07.422]👩‍🎤 Ate-A-Pi: What would you say the geographical spread of, is it like 80% US, 20% elsewhere? Like what is the geographical spread that you look at for your app?

[29:17.566]🤵 Blake Anderson: Yeah, it's somewhere in the range of like 65% US or I should say 60% US, 30% Europe, 10% rest of the world.

[29:27.05]👩‍🎤 Ate-A-Pi: And age-wise, are they, you know, is it like 18 to 24? Like what is the major, what is the largest age group?

[29:36.702]🤵 Blake Anderson: Yeah, it's primarily 16 to 25.

[29:39.71]👩‍🎤 Ate-A-Pi: 16.
Um, so if you had, you know, if you were to look forward 10 years, where would you want it to be? How would you look at this progression? Um, you know, you've talked about this kind of brand element with kind of three, this triangle of like, you know, the, uh, physical fitness, I guess, and, um, you know, attractiveness, uh, and kind of motivation. So how would you, how would you.
you know, what would you think about looking, you know, 10 years ahead?

[30:13.47]🤵 Blake Anderson: Yeah. I think that my high level thesis here can best be explained through an analogy. So when a guy has an issue with erectile dysfunction or maybe a receding hairline or a number of these physiological health issues, HIMS is a platform that has done an amazing job as branding themselves as the solution to all of these problems.
If you ask a guy who spends any amount of time on the internet nowadays, like, where do you go when you have one of these issues? Hymns will probably be the most popular answer. Now on the flip side, I do not think that exists currently, uh, in regards to self-improvement, right? When it comes to maybe you went through a breakup and you really want to improve yourself, you want to get more fit, you want to become more attractive, you want to develop a more hardened mindset.
better discipline, more focus. This doesn't really exist right now. And it's also a space that is obviously growing when we look to people like Andrew Huberman or Chris Williamson or even Joe Rogan. They are people that have been capitalizing on this upward tailwind in male self-improvement. So taking all of that into account.
I think that building a brand with this sort of reputation of hymns, but as the solution for self-improvement is a mission that I feel particularly driven to as an individual. I've always been super interested and attempted to embody self-improvement in myself. And I think providing guys with that.
You want to get more fit. You want to become more attractive. You want to, whatever all the previous examples, you go to UMAX. You go to whatever this brand is. And I think that is something that could pretty fundamentally change the trajectory of male sales.

[32:24.628]👩‍🎤 Ate-A-Pi: Um...
Cutting back to AI, for example, when did you first note that there was something special with the new kind of like the tailwind from the AI revolution? Like when did you actually feel like, oh my God, these models are now interesting enough to deploy in product?

[32:47.942]🤵 Blake Anderson: Yeah, super, super interesting here. So I first started playing with GPT-3. I first discovered it in October of 2023. I wasn't much of an engineer by any means at that point. I built a super simple, I forget what the package is called in Python, but I essentially built a super simple text box and enter, get result.

[33:19.886]🤵 Blake Anderson: like a piece of software, I guess, just making one API called the GPT-3 in October of 2020. Sorry, October of 2022. So this was before the release of ChatGPT. And I was like, wow, this is pretty amazing. I'm kind of surprised that no one's really using this. And then a little bit over a month later, ChatGPT came out. And that to me was paradigm shifting.
That's when I started experimenting with building applications on top of it. The first thing I built was an automatic slide deck generator where you could input any prompt and it would generate a beautiful slide deck for the purpose of presentations in school, right? Cause I was in college at the time. And then I continued just kind of toying around with things. They didn't dive too deep. I graduated from school and I was like, okay.
I do not want to get a full-time job. I want to pursue the path of entrepreneurship, but I also need to make money. I don't come from a background of wealth. That's what led to the initial development of PlugAI.

[34:33.887]👩‍🎤 Ate-A-Pi: And from PlugAI, why did you make the decision to transition to UMAX? As you know, what drove that decision making? When did you feel, okay, it's time for the next one?

[34:46.654]🤵 Blake Anderson: Yeah, so the idea behind Plug.ai to begin with was I was in a fraternity in college and guys were constantly using dating apps and would go to each other like, what should I say to this girl? Or how do I slide in? Whatever. And myself and I brought on two other co-founders who I had been working with before. So the three of us started it together and we were like, okay, we think that we can build this and we think we can do it pretty well.
Plug AI started blowing up. We got to a little bit over $200,000 a month in revenue. We're still doing 200,000 a month in revenue. And we kind of realized that we weren't gonna be able to scale it that much further. We, it would have been a slow journey up and we decided to put the thing on autopilot so we could pursue new applications. I get very energized by building something new.
just willing something into existence to me is really cool. And that upward momentum as well is an incredible feeling. And so in October, I left PlugAI. I sat around for a couple of weeks just thinking, what do I build? Where are the opportunities? Where is AI not currently being leveraged? Right.

[36:14.458]🤵 Blake Anderson: API release was scheduled for the coming weeks. And so I started thinking a lot about like what I could build with GPT-4 Vision. And then, you know, that crossed with the thesis I had in male self improvement just kind of felt like a perfect fit.

[36:31.978]👩‍🎤 Ate-A-Pi: amazing. Let's just say that you're probably the only two-time profitable AI founder in existence right now.
I think profitability numbers are better than like Coherent and like most of the large AI firms. So.

[36:44.001]🤵 Blake Anderson: Thank you very much.

[36:52.798]🤵 Blake Anderson: Yeah, yes, certainly. Yeah, it is really interesting to me.

[37:02.31]🤵 Blake Anderson: Though obviously there's the trade off of like it's just smaller scale smaller scope with these sort of things um But what's really interesting is that like I really felt that with RizGBT. I don't feel that with you Max though Like I know that we're profitable, but I also know that there's so much more that we can do

[37:21.262]👩‍🎤 Ate-A-Pi: Indeed. You must have lots of VCs asking you to ask you to invest. So what do you what do you how do you deal with that? Do you have you have you thought about taking a round? Is it something that you don't want to do?

[37:33.822]🤵 Blake Anderson: Yeah, I've definitely thought about it. Um, there are a couple of trade-offs here. Uh, one thing is like, we are making a lot of profit. So like we can, to some extent, bootstrap. The big question is like, how big do we want to scale the team? Um, as well as I have a, um, a somewhat negative view towards VCs as a whole. Um, I don't think that I really respect founder VCs. I think that there are a lot of VCs that were never founders and
have strong opinions about things that they don't really understand. And what's the tricky situation is that like, I think I've heard a number of stories of people thinking that they were taking money from a VC who would be totally on their side and it would be a great experience and they would just be there to provide them with connections and help them out whenever and then it just becomes this absolute terrible situation where they have to answer to them.
and the VC is blocking them from doing anything that they want to do. And so I'm certainly concerned about getting myself into that position. And being that we have the money to not be beholden to anybody, I think that we are in a pretty strong position as a company.

[38:46.522]👩‍🎤 Ate-A-Pi: Indeed. Has there been any moment when you've been using the AI that you have some weird sensation? This thing is doing something very unusual, and why is it doing it?

[38:56.214]🤵 Blake Anderson: Okay.

[39:00.028]🤵 Blake Anderson: just AI as a whole.

[39:01.594]👩‍🎤 Ate-A-Pi: Yeah, yeah, as a whole, as in like, you know, chat GPT, GPT Vision, or any of your models. Like, have you ever, you know, had like an anomalous kind of like, what is going on?

[39:13.562]🤵 Blake Anderson: Yeah, I think that when you experiment with some of the more jailbroken models, like Mixerl, or even like ChatGPT in the early days, or some of these Gemini or Sydney jailbreaks, to me it's always super interesting because we see this very put together form of large language
prompted into submission to just always, you know, stay within the guidelines. To me, seeing like the weird paths that it can go down in like...

[39:55.398]🤵 Blake Anderson: explaining how it does think that it's alive and how it does think that it has emotions. And now I don't believe that they're sentient in like current form. Though.
I do think that it's more nuanced than people make it out to be. I had a really interesting, so Walter Isaacson was my professor in my final semester of Tulane and we debated this endlessly, whether or not consciousness was binary or existed on a spectrum. I'm of the opinion that consciousness exists on a spectrum. And now, so I think that if this is a rock and this is a human, right, like a bear.
might be somewhere around here, and a dog might be somewhere around here, right? I often think about like, where do large language models fall in that? Are they more conscious than an ant? Right? I don't think that that's like a ridiculous assumption, but I don't know, it's super fascinating to me, and I don't think that I know enough about the subject to really take a strong position.

[41:02.902]👩‍🎤 Ate-A-Pi: I think no one, the exciting thing for me about AI is perhaps not even the guys who put out GPT-3, GPT-4 are fully aware of all of the capabilities, right? So I think it's anyone who uses the models can have an experience with them and figure out what, I always tell people try it for yourself.

[41:17.195]🤵 Blake Anderson: Sure.

[41:32.478]👩‍🎤 Ate-A-Pi: Don't just like, you know, take a review, right? Just try it for yourself.

[41:35.454]🤵 Blake Anderson: Mm-hmm. Yeah, so I actually posted something about this on my Twitter recently. I don't know if this has been a discussion, but it just came to me. And I was like, well, all the people who are weighing in on whether or not AI is conscious right now, that we hear in the media, they're all ML experts, and people building these foundational models. And that's great. They obviously understand them mechanistically very well. But.
I haven't really heard people take into account the opinions of psychiatrists, psychologists, philosophers, right? And so that's definitely a space that would be interesting to hear from more. I think that for the past almost 100 years, we've kind of hailed the Turing test to be like, this is it. The Turing test, that is the determinant of consciousness. And now we just kind of set it aside like it means nothing.
So that is, you know, the shifting of these gold posts is very interesting to me.

[42:40.906]👩‍🎤 Ate-A-Pi: George Hotz has a saying that AGI is anything that the computer can't do yet. So as soon as the computer can do it, it's not AGI anymore.

[42:50.162]🤵 Blake Anderson: Mmm, yeah, absolutely. Absolutely. Yeah. I agree with that.

[42:57.962]👩‍🎤 Ate-A-Pi: Um, you know, one of the things that I find is that the company's putting out APIs, um, you know, opening high and throw big, they also have moderation, moderation barriers and the moderation often is like, as soon as you try to do anything remotely fun, they're like, oops, you know, they're got a, got a cut, got a cut you off. Like that's the end. Right. Um, and as someone who like basically built apps, which were definitely more like social and.
more fun in that sense. Did you come across that at any time, either at Plug.ai or at, you know, UMAX, where you get...

[43:35.97]🤵 Blake Anderson: come across what oh it barring us from.

[43:38.978]👩‍🎤 Ate-A-Pi: Yeah, because it just says like, look, you know, this is, this is like, this is bad language. I can't, I can't allow you to say that. Like, you know.

[43:44.266]🤵 Blake Anderson: Yeah, yeah, with, with plug AI, we got it all the time, right? Because when we, you know, if someone's having a flirtatious conversation, and maybe the, the other person, the partner says something a little bit sexual, and then we feed that into the API, they're like, nope, can't touch this. So we had to have figure out some work around there implementing other models, that sort of thing.

[43:49.322]👩‍🎤 Ate-A-Pi: all the time.

[44:14.114]🤵 Blake Anderson: like these open source models that are less restrictive. But in general, I think it's a really interesting, I think that they are obviously airing on the side of the risk aversion. And so I understand it from these big private companies. I don't like it as an individual. And I would say, I don't think that they should do it, but I understand why they do it.
in respect to the rest of society and how they view things.

[44:47.27]👩‍🎤 Ate-A-Pi: Yeah, I think it misses the totality of the human experience. Right? So, and so basically what they're saying is, we'll give you the AI that you can see in the office, but it's not gonna be the AI that is gonna be with you the whole day. Right? That's typically what they wanna say. Go ahead. Yeah. No, yeah. That's just a corporate viewpoint that they have, I guess. So.

[45:05.502]🤵 Blake Anderson: It's, go ahead.

[45:14.758]🤵 Blake Anderson: Yeah, I think that there's a good corollary here to like the just free speech of individuals. And if we have as individuals have a free speech, right, it's because sometimes it's important to say things that are unpopular, right? Otherwise, we just go down this road of this group think where, you know, people can't express truly, truly novel opinions.
And I think that's also where it's a great source of inspiration and genius in our society. And when you kind of restrict the free speech of models, I think that it's a dangerous path to go down.

[46:01.234]👩‍🎤 Ate-A-Pi: Um, when you, when you had plug AI, when you, uh, did you, did you get responses? Hey, um, are you, are you using a, are you using a bot to talk to me? Did people, did you know the recipients ever like, you know, suss out like, hey, you know, are you, did that happen once? Like, and anyway, yeah, good.

[46:22.578]🤵 Blake Anderson: I don't think that it has reached that level of pervasiveness across society yet. In part because I don't think that the bots are actually better than good humans I'm flirting right now. What will be interesting is like in a year if they are or when they are. That will be pretty crazy.
if there actually exists alpha for every individual to begin using a bot to do their flirting instead of themselves. Yeah, that's a-
that will be a very interesting world where it's just bots flirting with one another. People talk a lot about the future of customer support. It's bots arguing with one another, but the future of flirting where it's just bots flirting with one another, that's pretty funny to me.

[47:20.562]👩‍🎤 Ate-A-Pi: Yeah, what if they had a good time and then they came back to you and they're like, okay, go date that person. I had a good time with them. You will too, right? So if, you know, what do you believe that no one else does? This is a Peter Thiel question, right? What do you believe that no one else does?

[47:30.071]🤵 Blake Anderson: Good night.

[47:41.899]🤵 Blake Anderson: Yeah.
Um, I haven't thought about this one in a little bit. It's a great question. Um.

[47:56.686]🤵 Blake Anderson: Thanks for watching!
I think that...

[48:07.762]🤵 Blake Anderson: And so I don't know if this is actually that unpopular, but I think that the majority of people actually just don't work remotely hard and don't really have anything interesting to say or do.

[48:26.066]🤵 Blake Anderson: and it's very pessimistic of me. I like to think of myself an extrovert. I love spending time with people. But when it comes to entrepreneurship, as well as people as employees, I hear people say all the time, like, oh my God, I've been working so hard recently. I think the vast majority of people work significantly less hard than they think they do. And that may be a popular opinion, I don't know.

[48:53.333]👩‍🎤 Ate-A-Pi: It's definitely a more SF opinion, right? Like it's a kind of SF, like, you know, that is a small subcult's opinion, but perhaps not, you know.

[49:02.594]🤵 Blake Anderson: So I think it applies to SF people as well. I think actually extra applies to people in San Francisco and in tech. And I am not an SF person. I spent a cumulative of two and a half months in San Francisco for HF0. And that is all the time that I spent there.

[49:18.642]👩‍🎤 Ate-A-Pi: How is how is H of zero?

[49:21.747]🤵 Blake Anderson: HF0 was pretty unbelievable.
I think that I have to preface this with like most of my life. In high school, you know, I had very normal friends. In college, more or less the same. I was in a fraternity, right? No one had like super high aspirations. And I definitely felt like a, you know, a little bit different from most people, given that I was, was very high achiever. After college, I moved back home to start working on these projects.
I moved to San Diego where I'm at now, the start of December, and then I get to HF0 in January. And, like, the caliber of people and the average intellect there was just mind boggling. In addition to the fact that everybody there was intelligent, but they were also brilliant. And I think those are two very different things.

[50:24.946]🤵 Blake Anderson: And just being able to like spend day in and day out with people where you can like cross-pollinate these creative ideas and thoughts on the world, it was probably the best two and a half months of my life.

[50:44.922]👩‍🎤 Ate-A-Pi: Indeed. What is AGI to you?

[50:51.544]👩‍🎤 Ate-A-Pi: I kind of gave away my definition, but what is AGI to you?

[50:58.266]🤵 Blake Anderson: I think AGI...

[51:04.43]🤵 Blake Anderson: So to me, this is...
I think that if you were to put AGI in somebody's body right now and it was able to move and function like them, and I wouldn't be able to tell that individual, that their mind was being run by an LLM, to me that is AGI. So I don't classify it as like, okay, if I'm in a chat interface, blah, blah. I like to think about the hypothetical, like what if one of my friends right now, his brain was just swapped with an LLM.
Would I be able to notice that he was in a real person? So I think that, yeah, exactly, exactly.

[51:40.682]👩‍🎤 Ate-A-Pi: Physical touring test, basically. Yeah, physical touring test. When do you think we get to AGI? Like what, like AGI and super intelligence. Like if you were to put a number on a date.

[51:55.034]🤵 Blake Anderson: Yeah, number and date is so hard because I think I'm just so unqualified for this question. I generally err on the earlier side. I think that 2025 makes sense to me that we will have a model where you could replace my friend with, given it has the ability to move. Even now is like, if any of you took GPT-4 and you gave it the correct tooling,

[52:01.454]👩‍🎤 Ate-A-Pi: No one is qualified.

[52:24.722]🤵 Blake Anderson: I probably wouldn't really be able to tell. So who knows, maybe we're already there.
I think that, yeah, to me, if somebody says, we are 100% at AGI right now, I'm like, I don't feel strongly that you're wrong. ASI to me is a significantly more interesting concept in that AGI to me kind of feels like it exists on a spectrum, right? Whereas ASI is like, when the model itself is able to just...

[53:04.366]🤵 Blake Anderson: to hit true exponential growth on its own and understand the implicit rules of systems. And so I don't know if we're gonna get to ASI with just more and more data or if it will be like a rule-based intelligence. But if I had to guess on that, I don't know, maybe like 2029.

[53:30.868]👩‍🎤 Ate-A-Pi: If you were to send a message to yourself, if you were to take a little time capsule, write something down, send a message to yourself 20 years in the future, what would you write on it?

[53:48.824]🤵 Blake Anderson: That's a good one

[53:55.658]🤵 Blake Anderson: This is cliche, but probably to continue being present and just like embracing every moment in life. Something that I've noticed recently is like the dilation of time and like the speed at which time moves throughout your life significantly increases, right? And I'm only 23 right now. And so I can only imagine that it will continue to do so in 20 years, right? Because like with each passing moment,
each individual unit of time is a greater proportion earlier on than later on. So one year right now is about 5% of my life. 20 years is about 2.5%.

[54:38.326]🤵 Blake Anderson: The solution to slowing that down is practicing presence and some level of mindfulness on a consistent basis.

[54:47.37]👩‍🎤 Ate-A-Pi: Amazing. Blake, thank you. Thank you for your time today. And this was an amazing session. And as I said, you are definitely, it's an amazing story. I mean, I don't think I've ever met someone as successful as you in their 20s. So.
You know, it's an amazing story and thank you for being with me today.

[55:18.242]🤵 Blake Anderson: Thank you very much for your time. It was a pleasure speaking to you.