Rich Palmer CTO of Gravyty

(Joel Beasley at 00:00:00) Today, we are talking to Rich Palmer, the CTO and co-founder of Gravyty, and we discuss using artificial intelligence for philanthropic purposes, how the quest for AI dominance is the new space race, and the humanization of technology versus the mechanization of people. All of this right here, right now on the Modern CTO Podcast. Here we go. This is the Modern CTO Podcast. So what are you up to today at Gravyty?

(Rich Palmer at 00:00:40) Today, let's see. What were we working on this morning? We're working on reinforcement learning, if you've heard of that before. So it's where, you know, Google beat the AlphaGo champion last year, I think it was.

(Rich Palmer at 00:01:02) And then, you know, IBM beat Gary Kasparov, chess champion in the nineties. They all did that through reinforcement learning techniques and similar things. So we're doing something kind of similar. It's a huge hard problem. So we've spent a lot of time just sort of figuring that out this morning, as we do most days lately. It's awesome.

(Joel Beasley at 00:01:26) It's just the morning. We're just solving the world's problems.

(Rich Palmer at 00:01:29) Yeah, well, and that's crazy, you know. And nobody's ever—you know, we also use convolutional neural networks, which is pretty cool. Most of the time, you know, the best way to think about that is if you're tagging, I don't know, like you're on Facebook and you're tagging a friend's face. You're labeling data, and then next time it sees an image, it can say, oh, this is Joel, this is Rich, this is Genie. Right? People usually do it with images.

(Rich Palmer at 00:01:56) Not a lot of people do it with text, and that's what we're doing right now. So again, we're not solving a problem there. We're sort of hindered by our technology at the moment. So it's just been so much fun.

(Joel Beasley at 00:02:09) So that's sort of like the assisted AI stuff, like assisted learning or reinforced learning where you have a human there that's watching what its guess is and then you're telling it if that is the correct guess or not.

(Rich Palmer at 00:02:19) Yeah, in certain parts. So some of our stuff is fully automated, which is the part I'm excited about. But we also do, it's called human-in-the-loop AI, which is like—to pick on Facebook again, but when they're trying to tag, what were they trying to do? They're trying to tag articles for abusive language and stuff. That's a human. They hired a thousand people to do that alone because AI is not able to do it just yet. Google has some similar stuff, a bunch like Twitter to figure out who's being a poor character on Twitter. It's all human-in-the-loop AI. And then, yeah, exactly what we do is, whether it's right or wrong, the outcome for us is not objective necessarily. You know?

(Rich Palmer at 00:03:06) If someone gives a gift to a nonprofit, that's a great outcome. But whether it's right or wrong, exactly what we're saying, requires some human intuition to make a determination. So we have people, every hundred batches, thousand batches, just check it out and see if it's still happy.

(Joel Beasley at 00:03:23) And so you're doing this AI stuff, but it's centered around philanthropy?

(Rich Palmer at 00:03:28) Yeah. So the three-word description of our company is AI for philanthropy, which is basically—you never think about AI or big data when you think about nonprofits. But they have more data than most big companies do. Right? Every time you swiped your card and bought a hat at the school store or made a donation or went to a game or did anything, they track all the data.

(Rich Palmer at 00:04:02) They just have no idea what to do with it. So that's a lot of what we do is sort of liberate that, go to big public data, get behavioral stuff from social media and everything like that, and combine it together, which is cool.

(Joel Beasley at 00:04:15) Yeah. So how did you get into this? Is this something that's personal for you? Or—

(Rich Palmer at 00:04:19) I spent the early part of my career on Wall Street. So I was doing, really early on, portfolio analytics and quant systems. And then eventually, you know, Wall Street's sometimes a tough place to keep your soul intact in certain situations. So I quit that.

(Joel Beasley at 00:04:38) Most situations.

(Rich Palmer at 00:04:39) Most situations. Turns pretty dark sometimes. I quit that to go to Berkeley, California to start a, to bootstrap a startup. It was basically like Etsy for food is what we did. So I went from, you know, buttoned-up high-paying Wall Street to really poor hippie food startup.

(Rich Palmer at 00:05:01) And that didn't work out for a variety of reasons. Actually came back to do sort of more of a fintech startup at Wall Street, but we raised about $120 million there. So not really a startup, just a, yeah, behemoth. But again, you know, sort of—it wasn't exactly for me towards the end. And then I actually, along the way, had suffered a brain aneurysm rupture.

(Rich Palmer at 00:05:35) So working hard, always, you know, doing pretty well, and then that just takes me out of the game for a little bit. Really hard recovery, couldn't walk, couldn't talk, had to relearn all that stuff and decided I needed to basically do a second chance at life kind of thing. So I focused on starting a company that really matched my interests. Ended up going to Babson College. So some people know about them, but their claim to fame is they're the number one entrepreneurship education school.

(Rich Palmer at 00:06:06) So for, like, twenty years or something. And then so that was my thing. I was like, all right, let's try a different version of life right now and sort of go meet some new people, try some new opportunities, right? And there I met my co-founder who was a fundraiser, like these schools have frontline fundraisers that basically help, you know, talk to donors, build relationships, close gifts. And he was getting his MBA at the time too. I had no idea what fundraising was, so I asked him, like, what's your job about? And he's a smart guy, but the limitations of his understanding of the data and tech and the school in general, you know, he was doing a lot of the work in just Excel spreadsheets or, quite frankly, with sticky notes and a bit of luck.

(Rich Palmer at 00:06:58) And so as we were talking, we're like, hey, what happens if we combine, you know, my skill set with your skill set, which is basically taking all that data, modeling it, and predicting outcomes. Turns out we did really, really well with that. We shopped the idea around for a little while to hospitals, peers and causes organizations, all kinds of—and now there's so many different types of nonprofits, and everybody seemed to have the same problem. So then we were like, okay, this matches both of our interests, and it's a socially conscious business focused on AI. Let's go for it. And then we got funding. So that's sort of the abridged version of the story.

(Joel Beasley at 00:07:40) So where are you at now? How big is the team?

(Rich Palmer at 00:07:42) So we are based in Newton, Massachusetts, right outside of Boston. Boston real estate is very expensive, so we're doing stuff here. We went through the MassChallenge program, which is a really big accelerator out here. We're in a building that's sort of a joint venture between MassChallenge and the city of Newton. So we're really happy here.

(Rich Palmer at 00:08:07) We're at, depending on the day, eight to ten people now. We think by the end of the year, we'll be up to fifteen to twenty folks depending on how the winds go. And we're pretty heavy on tech right now. So probably about 70% of our people are tech-focused, which is necessary to build algorithms and move fast enough and do some stuff that has never really been done before. And, you know, when we raised our pre-seed funding, we said the purpose of this is to have a bunch of questions and answer a couple of them.

(Rich Palmer at 00:08:45) And then once we've answered them, if we still, you know, deserve to be in business, let's go and raise a seed round to answer more questions. So we're constantly answering all these questions and doing really well. So that's—I think we have a big year ahead of us on that.

(Joel Beasley at 00:09:00) And then how did you pick Boston? Were you living in Boston? Or how did you become familiar with that?

(Rich Palmer at 00:09:07) Yeah, so me personally—so my wife is an environmental attorney and we were both down in the Virginia DC area when my aneurysm happened. I was actually working from home during that year down in that area. Her, it was a clerkship, so it had a finite sort of time. And I had applied to Babson and several other schools, and Babson just was the amazingly correct choice for me.

(Rich Palmer at 00:09:38) So we both decided to move up to Boston, and we've been here for four years now or something. Between—yeah, I think about four years now. And it's, besides the winters, it's, you know, really smart. A lot of really smart people, very culturally diverse. From where I'm sitting, there's a hundred schools in a ten-square-mile radius. So it's a fantastic place to have a nonprofit-based—you know, have your customers be nonprofits because I can just basically drive around all day and have plenty of folks to talk with. So and, you know, from an AI perspective, there's all sorts of weird articles around, but one said there's about 20,000 people who are capable of building these AI systems, and about a quarter of them live in the Boston area. So it's the perfect place to be for attracting talent and that kind of stuff.

(Joel Beasley at 00:10:31) Oh, wow. That's like AI hub, Boston.

(Rich Palmer at 00:10:35) Oh, it's crazy. You know, I talk about it a lot. I'm a mentor at the Techstars autonomous technology accelerator where it's all AI and ML and moving drones around the world and all kinds of stuff like that. Every single school here has some sort of accelerator or think tank around AI, ethics of AI, training educators, training business people, all kinds of—everybody's thinking about it, which is awesome.

(Joel Beasley at 00:11:04) Yeah. We were talking about Brad Feld this morning. He's one of the co-founders of Techstars.

(Rich Palmer at 00:11:09) Techstars. Yep. Yep. Yeah. I used to live in Denver for a little while.

(Rich Palmer at 00:11:15) In between my failed startup and, you know, the REL side, which was the one after that. And I always wanted to go over to Boulder and meet him.

(Joel Beasley at 00:11:25) Oh, yeah.

(Rich Palmer at 00:11:25) Never got around to it.

(Joel Beasley at 00:11:27) Oh, man. I've read—

(Rich Palmer at 00:11:28) The book. I can't remember what the book is, but they have a pretty popular book that I read. I can't remember what it is though.

(Joel Beasley at 00:11:35) Oh, yeah. He's a writer. Yeah. Yeah. I spent some time in Waltham—

(Rich Palmer at 00:11:40) Nice.

(Joel Beasley at 00:11:40) Which is right around that area. I remember just—we were at this cabin thing, cabin-style house, or it kind of felt like a cabin anyways. But it was right on this big lake. And so, to me, Waltham is like that house and that big lake. That's my memories of it.

(Rich Palmer at 00:11:55) Yeah. It's awesome. And I mean, in Newton—so we're in a, oh, it was somebody's, you know, Victorian mansion a long time ago, then they turned it into a library for about fifty or sixty or seventy years or something. And then the parks and recreation department moved in, and now we're here. So we're actually on the corner of a really big park, and they're refurbishing a pond that's here.

(Rich Palmer at 00:12:23) I mean, you know, it's cool to be out in the suburbs and not just stepping all over everybody like you do in the city.

(Joel Beasley at 00:12:30) Okay. So you're big into the AI. Did you—there's this article that came out. I saw it last night. I'm not sure when it came out, but it felt like it was within the past week. They're talking about how China has now declared, essentially, the second space race but with AI. Have you read that one yet?

(Rich Palmer at 00:12:46) Yeah. Not maybe not that particular article, but, you know, one of the interesting—yeah. The story is really interesting because one of the, you know, one of the really interesting things is that, you know, I was talking about AlphaGo and like IBM and all these, like, really interesting, you know, success metrics of AI against humanity. Those algorithms—there's another article that's been floating around. Those algorithms have been around for decades.

(Rich Palmer at 00:13:16) Right? So the concepts aren't new, but what has changed, and people talk about this sort of exhaustively, is that we finally have, you know, the processing horsepower to go through it fast enough. That maybe takes a back seat to the fact that we now have enough data to come up with meaningful results. So that's the combination of right algorithms, right business case and data, and fast enough processing power.

(Rich Palmer at 00:13:46) So one of the things that, you know, it's a little bit controversial to think about it out loud, but, you know, if you—all else equal, so algorithms are equal, processing power, let's say that's equal between, you know, the United States and China. The movement of data though is wildly different between our countries. So within the US, there's really no stopping data moving around the borders or gathering data from citizens and stuff versus China has a lot more controls over that. So a lot of that space race concept, popular opinion, I think I share it, is that they're gathering way more data than we are, which means they're, you know, poised to take a bunch of steps ahead of us if they can use it properly.

(Joel Beasley at 00:14:33) Dude, are you listening to the show all the time?

(Rich Palmer at 00:14:37) It's crazy. Right?

(Joel Beasley at 00:14:38) No. Because this is what I'm talking about. I've been talking about this at—I didn't, before I even saw this China thing. Right? We were talking about AI and who's going to win this race.

(Joel Beasley at 00:14:50) Before I saw this—I just saw this for the first time last night, but I've been thinking about it since about two years ago when I noticed an increase in AI funding because I was doing some, you know, private equity work due diligence. And I was like, dude, what's with the increase in all the AI projects? Like everyone wants AI. All of a sudden it became this investor hot topic. Right?

(Rich Palmer at 00:15:08) Right.

(Joel Beasley at 00:15:09) So I'm thinking about it. And then I'm brainstorming one night with one of my buddies, right, who's a CTO. And we're kind of talking, and he said, oh, we could do this with AI, we could do that. And we both have fifteen years plus coding experience, so we are very aware of what we could capably do. Right?

(Rich Palmer at 00:15:27) Gotcha.

(Joel Beasley at 00:15:27) And I said, you know what, man? I said, I don't think the code and the processing is where I would put my bet. I said, you know where I'm going to put my bet? And he goes, where? I said, I think if we do a startup, what we should do is we should build a tool that allows—like create a standard in the data so that, like, let's say we want to do—the example I gave was, I want to—I believe that in the future, personality data for these AIs is going to become more and more interesting because after everyone solves all these basic functional problems, the next thing is humans are going to want to talk to it and interact with it more like another human.

(Joel Beasley at 00:16:04) Right? And what are humans but collections of stories. Right? We all have our stories.

(Joel Beasley at 00:16:10)
Stories is how we communicate information. Storytelling is so important. So I said, what if we developed a tool by which we could design stories in ways that AIs could consume them and have amazing detail and be able to have this base story that then the AI could build upon? And what we do is we create all the stories, all the ideas, all the things like that, and then we wait for the technology to catch up. And then they'll be able to consume our data sources and write algorithms or whatever they need to write to take our large, vast data stores of different personality stories and all this stuff, and then consume that. Because the arbitrage, the thing that everyone's going to need the most of, is going to be clean, formatted, organized data that they know that, alright, I can have my machine learner go in here and consume all this data and it will understand the concept of empathy.

(Joel Beasley at 00:17:04)
But right now you can't do that because you don't know if the data is dirty. Forget structure of the actual organization, the actual data. You don't know that it's pure, that you're actually teaching that empathy. You know what I'm saying?

(Rich Palmer at 00:17:16)
Right. Well, and it's interesting. There's one thing that I read a paper recently. I can't remember who it's by, high-profile person. It's called "The Unreasonable Effectiveness of Data," which sort of talks about improving any of these things in AI, machine learning. Everything hinges on your data. The better your data is, the better your results are, which is awesome. And then from an innovation startup standpoint, it's fascinating to think about how a new AI company can come to be, right? Or who's gonna win, right? And a lot of what I've been seeing or formulating an opinion on is that there's sort of three types of companies in this space. Like, one is a vertical company, right? So you solve a very specific problem for a very specific end user, and you know all the data. You can evaluate whether you can solve this problem because you have all the data at hand. It's not a wide problem. It's just a very deep problem, right? That's what we do. That's what a lot of new AI companies are doing, right?

(Rich Palmer at 00:18:19)
The other side is sort of a horizontal one. So if you imagine an Alexa or—there's an article recently, a very long one. I think Steven Levy wrote it about, you know, Amazon's use of AI and sort of coming into it as a horizontal problem because you can pull data from all of these different sources because it's formatted, because it's all integrated with one system. I certainly, and many of my AI friends, have no business becoming a horizontal company because we don't have the data.

(Rich Palmer at 00:18:56)
And then I have some investor friends that, you know, they ping us every once in a while asking our opinion about AI and, exactly as you said, there's a lot of interest out there, money wants to get thrown around. Every time somebody says, "Oh, I'm a standalone AI company. I generate my own data and my own insights," I raise the BS flag very quickly, and I say, "I guarantee you might not know what you're saying, or you're bound to fail at this." Just because without the data, you're not gonna come up with the right insights that a customer is gonna want at the end of the day. So it's fascinating.

(Joel Beasley at 00:19:30)
It's pointless in a sense. I mean, it's less about BS, right? It's kind of more about being pointless. If you're manufacturing your own data and then getting your own result, what you need is you need a large array, like, a large variety of data, right?

(Rich Palmer at 00:19:47)
Right. It's like all the big Vs: volume, veracity, variety—all of the concepts of, like, what is big data. And if you don't check one of those boxes, you're poised to run into some struggles.

(Joel Beasley at 00:20:04)
Yeah. But we have two writers on staff and they're generating two articles a day and we're consuming that data.

(Rich Palmer at 00:20:10)
Yeah, exactly. Oh, and can I just use a corpus? I learned that corpus is corpora, which is a fun word. Really smart. So couldn't we use corpora? The problem is you can, and so can everyone else, you know? So there's no real advantage that—you know, for training, it's great, and maybe you tweak some parameters to make a smarter outcome. But until you get sort of those proprietary datasets, you're playing with the same hand that everybody else has, and that's not always a great place to be, especially as a startup.

(Joel Beasley at 00:20:48)
So you know about more about this China thing than I do. So I'm curious if you can answer this question. The gist of it, because I totally skimmed the whole thing, but the gist of it was that they're announcing, you know, oh, this is gonna be the next space race. AI is gonna be the next thing. But my question that I couldn't clearly find in the article that I read was, if you're gonna have a race, it's very clear when we went to put something on the moon, right? Because it's like, alright, you have a rocket and there's the moon, you put the rocket onto the moon. You could clearly see the goal, and then you work backwards to solve it, right? But I couldn't clearly see what they were referring to as the moon. What is the thing that says I won the race, other than just all-out war domination through digital hacking? Like, I don't know. Where is the goal? What are they running towards?

(Rich Palmer at 00:21:40)
Yeah. And I'm sort of faced with this weird duality between this past Wall Street shark life and my current socially minded, you know, survivor mindset, right? And I would hope that most countries are going towards, you know, altruistic uses of this data and trying to benefit people and humanity in general. But my answer is gonna be really weird. So in that sense, I don't know what the goal is. Maybe it's using AI to bring about a stronger economy, right? And on the flip side, it's actually very clear what the goal is—yeah, like you said, domination and war. But if that's our focus, then someone winning is also losing at the exact same time in my mind. So I'm not sure what the outcome is there.

(Rich Palmer at 00:22:30)
And then I've been reading some stuff about it where the goal of AI and information warfare is to almost have it be like nothing happened at all. You know, it's like with the indictments for Russia there, there's all these subversive, sort of inception-level changes that can be happening around you because AI made it possible. So you don't even know if you're losing. You don't even know you're at war until it's too late, kind of thing. So it's very—it's either gonna be amazing or really, really scary, and I hope it's somewhere in the middle.

(Joel Beasley at 00:23:09)
Yeah. I think it'll probably end up being somewhere in the middle. I do, however, think that we should visualize some sort of goal or set some sort of goal as either an industry or a world or a country, or—we should set some goal so that we can hit it.

(Rich Palmer at 00:23:31)
Well, and this is one thing that came up in the past couple of weeks: data is inherently biased, right? So even though our algorithms are built by people who could be biased, our data is too, right? So data is either created by, you know, a human, which—we're really flawed creatures, right? So that's biased. Or the mechanism for creation of automatic data—so, like, think about Internet of Things or something, right? That was created by humans who are biased. So we're starting from a bias, right?

(Rich Palmer at 00:24:05)
And then there's all these really intense ethical quandaries that are coming up about facial recognition, right? So a lot of the corpora are, you know, like 70% white males. And so when you try to run the system across all sorts—I just wrote an article on this—like, trying to run systems across genders, ethnicities, races, all that kind of stuff, you're already at an imbalance and a bias. So one of the coolest things—I can't remember who brought it up—but, you know how doctors have to take the Hippocratic oath, right? Do no harm. They're trying to start one for AI software engineering. And I think it's—I don't know. It's sort of like self-policing, I guess. But I think it's awesome. You know, instead of "do no evil" like the old Google model was, it's like do good. You know, do good first.

(Joel Beasley at 00:25:09)
I'm kind of just having stuff flow through my head right now, so I'll just let it out. Okay? So I'm kind of seeing in my head that there's a possibility. Like, if we let—if we leave it up to the governments to develop the AI, right, if we say, oh, we're not gonna work, we're just gonna let the governments lead. We're gonna let China lead. We're gonna let every—we're gonna let the governments lead. I think if we let governments lead versus if in our mind we let the private sector lead—and then I'll back up for a second here. I think that one of the ways for us to win, right, one of the ways for us to win is for us to develop an overwhelming amount of good AI, right? And just be consistently developing AI for good so that it's a disproportionate amount of altruistic good AI in relation to warfare-style AI.

(Rich Palmer at 00:26:03)
Sure.

(Joel Beasley at 00:26:03)
Right? So we would have to actually intentionally do that because from a macro level, I just feel we're in a better position than not addressing it at all. Or I'm sure any normal sane person would agree with me, right? So let's develop a disproportionate amount of good AI, and that smarter, better AI can help us when we deal with anything that—mischievous AI that might come up. But I don't think that the government's money is necessarily to develop good, genuinely good AI. I think it's more to develop process optimization and things like that, right? Which is fine, right? But that means it presents it, or maybe to develop a meaner, stronger AI so they can defeat the other government's meaner, stronger AI.

(Joel Beasley at 00:26:43)
So I think maybe the private sector needs to figure—put some, put its cash into some—developing these better AIs that will then fight the other. But I really believe that there will be some sort of weird AI digital cyber, like, world war situation, and not in a "go be crazy" way. But here's the thing: humans are curious. We're curious analysts. You get that individual that makes it, and the hypothesis is that it could go hack the whole other government's network and hijack everything. And then, you know, it's just whether or not they push the button, right? And then the curiosity in me says they push the button.

(Rich Palmer at 00:27:29)
Right. Which is crazy. And I mean, one of the things you can think about is back on the China versus US—I mean, I'm sure there's a broader competition, but we're the two behemoths at the moment. Yeah. Regardless of your political leanings, China's putting more government funding into AI projects. I don't recall the top—yeah, it's orders of magnitude greater than the, you know, the current administration or the former administration, so on both sides of the aisle, which is really interesting. And then, you know, so then you're like, I agree. You don't—you look towards the private sector for some innovation, right?

(Rich Palmer at 00:28:04)
The accelerator that I'm part of, too, they're a bunch of vertical AI startups that need to rely on government data, but the government has 50 forms you have to fill out before you can even show up at the door and all kinds of stuff. So they're trying to break it down. So you're exactly right. I feel like they might focus on process. And that's—process and warfare are great, but not altruistic, and it's not AI for good.

(Joel Beasley at 00:28:31)
Yeah. So where China wins is this. So my sister lived there four years, and she just moved back. And so I was getting to talk to her a lot about how it works over there. So in China, they all have one app. There's one app. It's your social network. It's your payment gateway. It's everything. We have different things. We have our PayPals. We have our bank accounts. We have everything separate for us so that when the government wants to tie it together, they would literally have to go to every single one of these independent companies everywhere. It's this big deal. All the data schemas are, you know, proprietary, so they're very different and need to be mapped if you're gonna consume the data, all this stuff, right? Yep. Cost, time, money, whatever.

(Joel Beasley at 00:29:11)
But in China, it's all in one system, man. And it's owned by the governments. They have everything. They know everything. And then they don't have to negotiate with the companies. They don't have to—they just walk in with their army and say, "Give it." Like, they aren't like us at all. They just do whatever they want from a government standpoint.

(Rich Palmer at 00:29:28)
It's interesting because our government—there's people thinking about this stuff, which is cool. So the senator, I believe, from Washington State—her name is escaping me right now, too—I think she's gonna have a hard time with one part of it, but she's trying to create AI-based legislation in a positive way so that the government can think forward on it. The first page tries to define AI, which I think doesn't deserve a definition in my view because it's nonsense, like, in general, right? It means whatever it means for the use case that you're solving. It's not like a whole encompassing, you know, theory of the universe for what AI means, right? But I think some people are hopefully taking some smart moves, but who knows how long that'll take.

(Rich Palmer at 00:30:18)
And then the other thing that's really interesting from—okay, abandon the government, look at the private sector. When we started Gravyty, you know, we weren't an AI company at the beginning. It happened—it happened slow, not slowly, but it happened, maybe six months after we were formed, just because, you know, people didn't want process and visualizations anymore. They wanted something to do work for them, and AI was obviously the best way to do that. But people still to this day try to evaluate us based off of SaaS metrics. So software-as-a-service metrics.

(Joel Beasley at 00:30:59)
Yeah.

(Rich Palmer at 00:30:59)
Which—AI, it's still too new. Like, what are the real metrics to measure us? And does the old way of moving private, you know, funds into companies still make sense? I would argue no. But I'm typically, you know, the metaphor of standing on a soapbox in London Square or something saying, "Oh, change the way you do your investment thesis." It's not working so well for—it's fine for us, but many of our friends, it's not working well for. So it's almost like a fundamental shift across policy, investments, ethics, and data collection that's gonna make us win. And any one of those things we mess up on, it might set us back pretty far.

(Joel Beasley at 00:31:49)
I have a goal. I have the goal for the AI space race.

(Rich Palmer at 00:31:53)
What's that?

(Joel Beasley at 00:31:53)
Ready for it?

(Rich Palmer at 00:31:54)
Yes.

(Joel Beasley at 00:31:54)
Where the machine, the artificial intelligence, is fluid consciousness. So when you interact with it, it's no different than a person. And not only that, so that's part one, to be able to learn and then understand and whatever. But the part two would be for it to be able to do it maybe 50 times, maybe 100x faster than humans. Like, I have to wait for my daughter to reach the age of 15 before she starts. I mean, she's six months right now. But she's, you know, walking at a year maybe, right? She's gonna talk—one, two. I mean, that's two years, 24/7 being, or well, you know, ten hours a day being on—

(Rich Palmer at 00:32:37)
Yeah.

(Joel Beasley at 00:32:37) Before she gets some basic functional stuff, right? So I mean, I kind of consider the computer right now to be how she was when she was physically delivered, when she was born. Her processes, vision functionality.

(Joel Beasley at 00:32:53) She's got some movement functionality. She's got some basic functionality, right?

(Rich Palmer at 00:32:58) Yeah.

(Joel Beasley at 00:32:58) But she hasn't really learned to walk yet, you know? She hasn't really learned to talk yet, right? But once she does, you know, so I think you're going to have to be able to boot up a machine from nothing, right? And then have it, not even do it as—I mean, it takes humans to mentally mature, like, 25 years, right? You notice differences in your mental maturity every decade, right? So we have to condense that 18 to 25 years down into, I don't know, a couple weeks, a couple months, right, of just straight machine learning. And when we hit that, well, once you can create a person or once you can create a machine that can think and interact and consume as fast as a human mind, once you get there, then you can do anything in the world because you can just make people on demand.

(Rich Palmer at 00:33:57) A couple months back, there were two AI books on his desk. One was a book called The Master Algorithm, which talks a lot about that. I'm sure some of your listeners have read it or heard about it or whatever, but where we're at, artificial intelligence right now is, yeah, it's a newborn baby, right? There's an article I read yesterday that AI, in its best form at the moment, is not able to defeat a moth in trying to discover new smells, like to learn smells. A moth's brain is faster than the best AI we have right now. So it's like we are definitely at the beginning. And then so it's just AI for specific tasks. And then this book, you know, there's some incendiary thoughts in there, but in general, you next go to artificial general intelligence, which is us, humans, right? We can learn anything. When I see something new, I pretty much, you know, an apple—when I see an apple as a child, it only takes maybe one or two more times to know that a green apple looks like a red apple, looks like a rotten apple, looks like a picture of an apple. It's like I learned it very, very quickly. And the thinking, which is the scary dystopian thinking, is that once you hit artificial general intelligence, the AI is not going to be satisfied with that. And it'll quickly move to artificial super intelligence, which is the, you know, the overall, you know, exploration of that book. Which is how do we do things that, you know, our human minds can't fathom.

(Rich Palmer at 00:35:33) So if AI right now is the distance—you know, you imagine using your phone—AI now is the distance between, we're in the middle of the spectrum and, let's say, a primate is on one side. So the distance between AI and us using a phone is that. You hand it to a primate, they'll see it, they can hear the sounds, they can sort of understand what's generally happening. The distance between us and AI superintelligence, though, is the distance between us and an ant. So, like, using a phone.

(Joel Beasley at 00:36:10) Dude, that's what popped into my head. Yeah. Before you said it, I was like, probably an ant.

(Rich Palmer at 00:36:15) Yeah.

(Joel Beasley at 00:36:16) And then you said ant, and I'm like, why did both of our minds go there?

(Rich Palmer at 00:36:19) Well, it's like, you think of the strength of ants and the fascinating, all the stuff, but in what context does an ant understand what's going on with the phone? Like, what's going on with the phone? So, like, that's, if we get it right, it's beautiful. If we get it wrong, I don't know. I'm, maybe I like my humanity a little too much, but I'm worried I'll lose it at that point. So it's kind of crazy.

(Joel Beasley at 00:36:43) Right? Isn't that insane? Oh, man. Just to think about the possibilities of even when you start getting out of the Earth, like, an AI's omniscient thought can be transmitted instantly across the universe, but yet we are limited to us.

(Rich Palmer at 00:37:00) Yeah. Quite limited in our own—I started, I've been watching in what little free time I have, I've been watching, what's the show? Altered Carbon on Netflix, which basically—

(Joel Beasley at 00:37:14) They had that crazy NCS, yeah.

(Rich Palmer at 00:37:17) It's AI, and it's this theory of, like, oh, they figure out how to store consciousness in a data storage so you never really die. You know, you just move to new bodies and stuff. And it's like, holy cat. They explore it in a fun, sensationalized way. But in reality, if we do that, my God. It's awesome. I don't think we as a species are prepared for all the little, you know, implications of what AI could be if we let it grow.

(Joel Beasley at 00:37:47) Dude, I wish I had this show, like, years ago because you could talk to my family and my wife and stuff. Yeah. I've been talking, I'm 30—I've been talking about this since, like, well, I actually gave that one specific example of storing the consciousness in computers when I was, like, 16. I remember when I was, like, driving around talking about it with my brother, like, while I had just gotten my license. And he would just sit there and be like, dude, you're crazy.

(Rich Palmer at 00:38:12) It's like, now—

(Joel Beasley at 00:38:13) Now I'm like—

(Rich Palmer at 00:38:14) Premonition. You know? It's like—

(Joel Beasley at 00:38:17) No. I'm like, no, no, no. Because here's what I'm doing. Here's the context. Okay? We are—if you step back and look at macro patterns, we bring technology closer and closer and closer and closer to us, and the end result is us being one and inside of technology. So it's really not that crazy to throw it out there. It's crazy when you walk up to a person today and say, hey, can you believe that we're going to be living in a computer in 50 years? And they're going to be like, you need to go to a mental institution, right?

(Rich Palmer at 00:38:44) Right.

(Joel Beasley at 00:38:44) Well, if 50 years ago, you went to someone and said, hey, you would have all the knowledge of the entire universe in your hand in a screen and in a device a billion times more powerful than anything we have right now, more data than anything you have, and if you do, go to the mental facility, that would be impossible.

(Rich Palmer at 00:38:57) That's amazing.

(Joel Beasley at 00:38:58) Right? So it's like, if you just look at the pattern of technology, we constantly put our brain technology closer to us, increase technology. We had a conversation the other day about how when we imagine an invasion, we imagine other creatures coming from outside of the Earth to us and them being humanoid shaped, right, like an alien coming. But we were talking the other day on the show, I said, dude, just imagine, like, it doesn't have to be humanoid shaped. What if it's just silica? Then if you actually look at it with that perspective, it's taking over the world.

(Rich Palmer at 00:39:29) Well, they talk about, so if you think about our ability to manipulate atoms right now. So we're doing that in all these really big, sophisticated pieces of machinery, but we're using our human brains to figure out how to manipulate that stuff. Imagine if an AI can go on its own. So you hit artificial general or superintelligence, and all of a sudden, AI could manipulate atoms in its own way. It can manipulate the physical world in its own way. And then you're exactly right. Like, what is the outcome of that? It's just, I find it fascinating. So on the same note, I was leading sort of a normal life until at some point, I started reading the books Ray Kurzweil puts out on the singularity and when humans are no longer, you know, machines are no longer distinguished from people. And we should have started a podcast then because that stuff is wild to think about. And, you know, one thing that I struggle with still is, in the context—let's say even in the context of my little corner of the AI world—is this the humanization of technology that's happening, or is it the mechanization of people? And I think it's different, but I haven't quite figured it out yet.

(Joel Beasley at 00:41:02) Well, it's just the same thing, two perspectives.

(Rich Palmer at 00:41:05) Right.

(Joel Beasley at 00:41:06) It's the same event that's occurring. It's just what perspective you take.

(Rich Palmer at 00:41:09) Right.

(Joel Beasley at 00:41:09) It's happening. It's just your brain is trying to figure out how do you process it.

(Rich Palmer at 00:41:13) Yeah. The humanization of people could lead to, you know, job loss, and it could lead to all these scary things.

(Joel Beasley at 00:41:20) It could. I mean, look, it already has, dude. Yeah. Like, I will bring it up on a positive note. We are coming out to Boston. Okay? We're doing a Modern CTO World Tour.

(Rich Palmer at 00:41:31) Yep.

(Joel Beasley at 00:41:31) Have you read any of our, like, propaganda about this yet?

(Rich Palmer at 00:41:35) Not that—not the tour, but I'll be here, and I would be happy to be involved in whatever way.

(Joel Beasley at 00:41:42) Yeah. We want you to come hang out behind the scenes. So what it is, it's this high-energy geek night out, right?

(Rich Palmer at 00:41:46) Yep.

(Joel Beasley at 00:41:47) And we're doing it to kind of give back to the college computer engineering world, right? So it's this one night, and we're going to transform the college auditorium into this technology paradise. So it's going to open with this drone show inside, right? Lights, drone show, media screen, live DJ. The drones and the music and everything will be synced, and so this, like, you know, five to 15-minute opener, insane high energy, right? Then after the drone show, I'll introduce, come in and you'll say, hey, what to expect as the host and everything. This monster aerial drone will be flying around, right, equipped with a 4K lens, taking shots of the crowd, right? So it'll be live-streamed not only at that university, but it'll be live-streamed at every computer engineering and engineering university in the entire world, like there's 7,200 of them, right? And then I'll interview the special guest, which will be a really, really big high-profile guest. And so there'll be one at each—so it's a world tour. So there's 16 locations and one of them is, you know, in Boston. And then we're going to have Elon Musk-style flamethrower, right? And one of the computer science kids is going to get to shoot off the flamethrower, right? So they're not going to let us do it inside most likely. Hopefully they would, but they probably won't. So what we'll do is we'll fly that 4K drone outside and let the individual shoot it off outside and it'll be streamed, the video stream back into the media screen on the inside, right? Then we're going to bring up a local technologist, right? We're going to bring up local technologist like a cool CTO and some cool robotics engineers that are from whatever city that we're currently in, and then we're bringing up, we do a little live Q&A from the crowd, from some of the engineers and the students in the crowd, and it's all free for the college students. It's at the college, it's all free for the college students. Then we're going to bring up on stage, we're going to do a show and tell. So we're going to bring up like a Boston Dynamics robot or like a Cassie, one of the bipedal walking robots, right? We're going to get to play with it and see it, and then we're going to announce winners of the hardware and software hack where each group will win $10,000 given away on the whole tour. Giving away over a million dollars in hardware and software hack prizes and stuff like that. Then we exit with a high-energy drone show finale, right? It's going to be insane.

(Rich Palmer at 00:43:56) When is this? Tomorrow? Can this be tomorrow? This sounds awesome.

(Joel Beasley at 00:44:01) Yeah, right? We're kicking it off, I believe, spring, in the spring, so like three months we're kicking off the tour. We're kicking it off in Florida, and then we're hitting 16 cities around the world, and it's going to be the most—have you ever heard of anything like this before?

(Rich Palmer at 00:44:15) No. I mean, I'm imagining what I saw in the Olympic opening ceremony right now with the coordinated drones and all that stuff. So I am very excited. That sounds awesome.

(Joel Beasley at 00:44:25) Yep. We are so pumped. It's going to be the biggest event in engineering, like the most fun thing that anyone has ever experienced, and it took a whole team of us to just imagine it and create it and come up with it. And now we're executing it, so we're really pumped.

(Rich Palmer at 00:44:43) That's good.

(Joel Beasley at 00:44:43) Maybe we'll put you on our list of advisors for AI advisors for the tour. So we will run AI stuff by you, say, hey, we should do some AI thing on the media screen to raise awareness of this and that, and you can kind of give us some direction, just a 15-minute call. Would that be cool?

(Rich Palmer at 00:44:59) I would love that. And, you know, this is a big research city too. So the closer you get to research is very interesting from a technical standpoint. But then there's also this really big groundswell about the ethical use of AI. So we touched on some of it as we were just talking. I'm just planting some seeds now. Those would be very interesting talks or showcases or whatever for AI. So I'd be happy to share whatever I can be helpful with. That'd be awesome.

(Joel Beasley at 00:45:33) Awesome. Thank you so much, Rich, coming and hanging out with us today. If people want to find out more about you, how would they do that?

(Rich Palmer at 00:45:39) Sure. They can go to just www.gravyty.com if they want to find out about the company, or you can find me on Twitter. It's just Rich M, you know, Palmer is my name.

(Joel Beasley at 00:45:52) How do you spell Gravyty?

(Rich Palmer at 00:45:54) So the, you know, proper word domains are difficult to come by in 2018. So Gravyty is spelled with a Y in the middle. So it's G-R-A-V-Y-T-Y.

(Joel Beasley at 00:46:11) Excellent. So type in Gravyty or Rich, and they're going to find you.

(Rich Palmer at 00:46:13) Yeah. They'll find me.

(Joel Beasley at 00:46:15) Awesome. Thank you so much, Rich.

(Rich Palmer at 00:46:17) Awesome. Thanks, Joel. Appreciate it.

(Joel Beasley at 00:46:24) Thank you so much for listening to the Modern CTO Podcast. Share this. Get the word out. Thank you guys so much. I couldn't do it without you. I appreciate it. You guys are the absolute best.