AI’s future is more than technology. It involves enabling innovation, revolutionizing education, and promoting accessibility. Join us in this episode as Kevin Dewalt, co-founder of Prolego, talks about the cutting-edge developments, challenges, and exciting possibilities that AI brings to our world. AI is changing our lives in amazing ways. It has made programming easier to access and self-driving car technology is advancing rapidly. And today, Kevin talks about these transformations. He sheds light on how AI is poised to revolutionize everything— even learning, making education more accessible and personalized. Get ready for a paradigm shift that’s not just on the horizon but happening now. Tune in now!

—

Watch the episode here

 

Listen to the podcast here

 

The Future Of AI – A Conversation With Kevin Dewalt

We have Kevin Dewalt who is the CEO and Founder of Prolego, which is a company that has taken the biggest businesses and helping them integrate AI into their workflow. I’m glad to have you, Kevin. AI is such a big buzzword. It’s something that we’re all talking about. You have the inside scoop on how this is going to affect our interaction with a lot of these companies that we see on a daily basis. Tell us a little bit about yourself, how you got here, and what you think AI is going bring to the businesses that you’re offering.

First, thank you so much for having me on the show. It’s an honor to be associated with the quality of the guests you have on the show. I might provide a little bit about my background and my perspective because that will help the audience understand where I come from. I’ve been doing this work for many years. I made my first neural network in the early 1990s as a grad student at Stanford when I was trying to figure out what classes should I take. I saw this neural network and said, “This sounds amazing. AI thinks like a computer. I can build something that’s going to predict the stock market.” I quickly realized that the five-neuron network that I built in MATLAB was slow and not very useful.

One way or another, for the past years, I’ve been involved in software innovation as a founder and venture capitalist investor in big companies, small companies, and the government. I’ve done the gambit. Years ago, I started Prolego because I felt like the time had finally arrived when the technology was good enough to build a business and start solving real problems. That’s what we do. The perspective I bring is somebody who’s out there working with the biggest companies, helping to identify their problems, think about their strategies, and what they can do as a business to leverage this technology to make their businesses operate more efficiently. That’s the perspective I bring in Prolego and what I do.

Everybody knows what artificial intelligence is from a very superficial standpoint. They know the word. They know that it’s some brain in a computer but what I’m most interested in is this. What is the future going to look like when we have artificial intelligence built into a company that you’re working with? We will get a little bit further into some of the topics that are cool and interesting but I want to know from a company that you’ve worked with. How have you brought artificial intelligence into their business? How has it changed them for the better?

Everyone has been hearing about AI for at least the last few years. When we talk to business leaders in public, sometimes we get the reactions like, “I’ve been hearing about this stuff but what’s different? What’s happening now?” The challenge of building and solving real problems with AI came down to money. To use what I’m going to call narrow AI, which is everything before ChatGPT came out and some of these generative AI technologies, it was expensive and hard. To try and solve a business problem with the technology, you had to get data curated by infrastructure and get expertise, and there was a lot you had to do.

 

 

As a result, there were not a lot of profitable business cases at most companies. You had to be at Google, Apple, Netflix, or a tech company with the data and the scale to leverage this technology. Everything changed with the arrival of ChatGPT and these generative models. The big shift for the industry and companies is that AI has now become a general-purpose technology. If the previous generations of AI were building a laptop from scratch, the current AI technology is going to Best Buy, buying a laptop, and being able to use it for whatever you want.

That means there are many ways that we can deploy this technology across different types of products and services for a much lower cost. The nature of the technology is going to usher in a change in how we build software. For the past few years, if you want to build software to solve a business problem, you have to explicitly write rules and tell the computer exactly what you want to do. If any of your audience has ever done any programming, they will be familiar with typical programming languages.

 

 

We’re entering a new area of software engineering where instead of telling the computer exactly what to do, programmers will be able to take AI technology, give it direction, and have it figure out problems. The implication of this is that we are going to have much more software because what used to take a team of 20 people and 20 months will be done by a team of 2 people in 2 weeks. We’re going to introduce software at scale everywhere very quickly. When we look at the future, that’s the lens I look through it. I can’t try to predict exactly what’s going to happen. Will it be good or bad? I don’t have a crystal ball better than anybody else does.

For example, when I think of artificial intelligence, I think about the Google Assistant software. That’s something that’s going to change all of our lives for the better. It would be so great if I could have my assistant get me a reservation at a restaurant without me having to go through Resy or Yelp, book a reservation, and look at the exact time. I have to manage my calendar. I have to manage my wife’s calendar. I have to make sure that my kids don’t have anything. The idea of having somebody do that for you is the future that I want. For the layperson, in some of the businesses you work with, can you give an example of how that artificial intelligence work will translate to a better experience for them?

Let’s use your example because that’s a good one. The reason that you don’t have an automated solution to solve that painful and simple problem like travel planning, which is the example I used in my past book of how tedious it is, scheduling and planning this stuff is still tedious and hard. Everybody knows it’s a pain, and companies have been trying to solve it for decades but no one has figured it out. When you try to tell a computer how to do this, it’s hard because computers don’t generalize well when you tell them exactly what to do, and there are too many edge cases.

You want to find the time when you’re available and someone else is available. You want to go to this restaurant, not that one, at this time of day. There are all these exceptions and boundary conditions. Traditional programming is not very good at that. These new AI technologies can handle general direction and figure out problems by going out and gathering up the data they need to do it.

From a software engineer’s perspective, I don’t have to tell the computer how to figure out everything to coordinate schedules. A model says, “Look at my schedule. Look at my wife’s schedule. Do a Google search, find out restaurants in this area that are available, find me a good time, and come back with three suggestions.” It can do that now. That’s what’s already here.

Give me an example. Does Google offer that?

We will find out because they are getting ready to roll out a release they call Gemini. Google is effectively scared to death of OpenAI and Microsoft. The company pulled the fire alarm, went on all alert, and combined the best technologies they had. They swept away the bureaucracy. They brought Larry Page back in to drive it. He’s personally driving the release of their next models. They’re training it on five times the computing power that OpenAI used.

Is it going to become a commodity similar to other industries where you have large companies that are the source for AI? Is it going to be more democratized where a smaller company or a handful of programmers could offer the same services but not at the same scale? What is the future of the business of AI going to look like? Is it going to be top–heavy? Is it going to be more distributed?

It’s bottom-heavy. It means that the two-person engineering company or startup can do things at Google’s scale. It’s hard for me to see how a lot of legacy companies are going to survive. I’ll give an example. Let’s pick a software. I’m sure your audience knows the company Salesforce. It’s a customer relationships management software that companies can buy, connect their data to, and start using to help their sales and marketing more effectively. It’s a large rules-based application that has been built over decades. To get it up and running in your company is hard. It’s easy to give them your credit card but to give them your data, customize it, get it working, and get your salespeople using it is difficult and expensive.

We’re going to see a generation of smaller companies that can offer most of the same solutions at a fraction of the cost where you connect the AI to your data, and it will start helping you with your sales. That’s what’s coming. It’s going to be democratized. The great analogy is the iPhone. Elon Musk can’t buy a better iPhone than I can. That’s where we’re headed. Intelligence will become a commodity. We can talk about some optimistic things for society whenever you’re ready.

AI is commodity intelligence. It will become a commodity.

I would love to because that’s what I’m all about. I love looking at the future and an optimistic life because honestly, I can’t wait until I have the ability for AI to do a lot of the scut work that I’m not interested in doing. There are so many admin tasks that are required to be an adult in 2023. That is not what human beings are created for like responding to my jury duty letter or doing stuff like that. Let the computer handle it. What are you seeing down the pipeline that you’re most optimistic about?

Let’s take that example. You are right that there is a lot of tedious day-to-day stuff, and it’s easy to look around you and say, “When intelligence is democratized, I don’t have to do that,” but that’s not the biggest social impact. What I see coming is the ability to provide that intelligence to everyone. I have members in my family as everybody does who as middle-aged adults struggle to do life, make decisions, think critically, and make good choices.

Social scientists have been talking about this for years. If you can invest in people’s childhood education and increase their competency by single-digit percentages, there are massive social implications. You make somebody 5% smarter in preschool, and suddenly, it returns tens of millions of dollars to the economy. It’s very clear. Imagine, we can instantly provide these people with a tool that is going to be free on their phones that helps them make better choices with basic life stuff. We are looking at an instant lift in productivity.

If you can invest in people’s childhood education and increase their competency by single-digit percentages, there are massive social implications.

I have family members who say, “How do I get to the airport?” Especially when you look at people who English is not their first language navigating the world when they have a lapel pen that’s listening, helping translate and tell them what to say, and telling them, “That’s what I need to do,” the social impact is massive. Suddenly, the gap between what is a frustration for you and me and what is a survival life skill for everybody else, that’s going to go up, and that’s an exciting future.

All of those things that you talked about are getting me so excited to see come down the path. Universal translation alone is going to break down so many barriers. The idea of talking to anybody without any of this stuff in the way is going to be significant for us as a species. The idea of human beings being optimized is something that I also appreciate, especially with all the health data.

I’m in the medical field. Wearing an Apple watch and seeing minute-to-minute what my values are for X, Y, and Z has not been explored to the level that it should be. Everybody could have an AI health coach making sure that they optimize their health in the best way possible. The insights that we’re going to get from all the data are going to be pretty interesting because there’s too much data for human beings to realistically look at. We have to leverage some of these other ways to look at the data so that we can get the same insights.

There are seemingly straightforward and simple breakthroughs coming out. The general media does not fully appreciate the impact but I know you will.

Can you tell me about those?

I’ll tell you one. I’m sure you know that one of the big challenges in the healthcare system is getting the data out of paper, notes, and everything and into a structure where you can process it. Everybody hates these. I go to my doctor’s office and they have to spend half the time looking at the computer and typing into an EHR instead of talking to patients. Doctors hate it. Patients hate it. The only people who like it are the people who are trying to do billing. That whole thing is a disaster.

OpenAI has released a computer vision component to ChatGPT where you can take a picture of something and convert it to data. I was contacted by a big bank that said, “We want to do analytics in our loan applications.” I know about this problem. One of the big problems is that everybody sends in loan applications, and they have their chicken scratch filled out. How do you convert that into data that can be used by a computer?

I took a loan application and filled out my chicken scratch handwriting. I took a picture with my phone and said, “Turn this into JSON format,” which is a type of data format where you can take a text document and return the structure and information in the form, and the computer does it. I sat there and watched my phone take everything I’ve written out and convert it into a data format that we can use.

That’s not the future. That’s now on the phone, and OpenAI is going to make that available for their API in early November 2023. Every company that has data sitting around on paper is now going to have a solution to convert that into an electronic format without all of these expensive systems immediately. That’s a small incremental change but imagine the impacts of that in healthcare.

That’s going to be huge. The human beings were not meant to write all these forms that we have to fill out. That’s not an enjoyable process for anybody. I’ve not met one person in my life who enjoys filling out forms. Honestly, there’s so much pushback to AI in general. Even learned people who I meet on a regular basis have this hesitance toward artificial intelligence because they fear the loss of control and that it’s an existential threat. What are some of the arguments that you would push back on with those people so that they get the same optimism that you have?

They’re correct for being vigilant and concerned. We have been a human technology society for thousands of years. Every time an innovation comes along, there are winners and losers but generally speaking, tech society moves in the right direction. When the printing press came out, it caused wars but there were not many people sitting around and going, “We should have never come out with a printing press.” It’s how just it happens.

We are a human technology society. We have been for thousands of years. Every time an innovation comes along, there are winners and losers.

We all make the same mistakes whenever there’s new technology. We look at the world and where we are, think about introducing technology to that world, and take a linear approach to how the world will change with that technology but that never happens. The world completely changes when the technology arrives. There’s a lot of understandable fear from smart people who are making this leap. I’ll give you one that your audience will be probably familiar with. It’s this idea of disinformation and misinformation. We’re all concerned about what’s happening with deepfakes, social media, and this fire hose of craziness that we see from whatever your political bend is. We see it happening on all sides.

When you imagine taking AI, taking the bad actors, and giving them this weapon, it can supercharge the internet to the point where I’m going to log in and see a fire hose of Dogecoin scans and QAnon everywhere but what they’re not considering is the point that you raise, which is the ability to process this data with AI. A major challenge with this information is that the bad actors are flooding the zone. To use Steve Bannon’s quote, “I’m going to give you so much that you can’t manage it, and you don’t know what’s true or not.”

Somebody goes on Joe Rogan. I’ll pick Robert Kennedy Jr as a person who’s incredibly persuasive and has some ideas that I agree with and some ideas that I find highly skeptical. If the person goes on Joe Rogan, and they’re persuasive and starts spinning out information, there’s no way for Joe Rogan or anybody to fact-check that in real-time.

By the time anybody does, it’s out there, and everybody is onto the next thing. We know that the truth does not survive in a free marketplace. That hypothesis fails but you can imagine that AI can do this, sit there in real-time, sift through these things, and say, “Maybe you should ask that person what evidence they have for this claim.” Suddenly, it allows us to filter so that we’re not sifting through a firehose of InfoWars from our relatives over Thanksgiving.

That’s the underlying basis for all of this. The world has become so complicated that it’s tough to process all the data, and we need help. Human beings are imperfect creatures. We need to leverage these technologies to expand our capabilities. That’s why I have an optimistic view because anytime we have tried to expand our capabilities, there have been bad things that have happened but Martin Luther King used to say, “The arc of history is long but it bends toward justice.” It bends toward progress. We live in the best age of human beings ever. To be a king in the twelfth century versus now as a layperson, I would choose being a layperson now over being a king, and you would be a fool not to. The idea of AI is another adjunct to how we’re going to progress as a species.

 

 

You will appreciate this one. Whenever I hear people long for the good old days, I have three words for them that end that conversation, “Surgery without anesthesia.” That ends no matter who you are. You think about that.

I tell people now when I’m talking to people, “This has a 99% success rate, and there are certain surgeries that have an 87% success rate. This is not a surgery you want to get into,” but then back in the day, it was 50% or less for a lot of these things and heroic actions, and now we’re so focused on these few percentage points. It’s a different society. I was wondering if somebody was young and they wanted to get into artificial intelligence. They say a buzzword. You started at Stanford. What would you say to a young person who is interested in getting into this field so that they could be at the ground level for this explosion of growth?

It’s the same advice I give to CTOs who want to get their next job as the head of AI, engineers, or products. Start doing it. Start learning how to build things. Start using ChatGPT. Start using the tools. Shockingly, there are so many people who are paralyzed by indecision. What I’ve learned in the past few years of technology is the expert is the person who’s not afraid to try. Take action. Join a team. Don’t talk, think, worry, or obsess over the articles.

 

 

There are a lot of valuable topics like AI fairness that are very worth considering but if you want to be a part of this act, do it. Try writing an application that leverages GPT-4. You will find out that it’s not that hard. It just takes some determination. You don’t have to learn all the programming stuff you had to learn years ago. The technology allows you to give it general instructions. That’s my advice for anybody. Start doing, building, acting, and trying to solve problems. Upload a spreadsheet to ChatGPT and see what it does. Take a picture of a form on your desk and say, “Tell me what this means.” Start writing about it, talking about it, and sharing an offer to help other people. Start doing.

Is programming more accessible now because of this? As somebody who’s not in the programming world, it’s very daunting and intimidating. People are talking about artificial intelligence but is it accessible to somebody who doesn’t have a programming background?

It’s continuously gotten more accessible throughout my entire career. The tools get better. I started a program when I was fourteen years old on the Commodore 64 when I was a kid. It was very painful to do anything. It’s getting easier. What is a program or not a program? That line is getting blurry. There’s an impression. They’re like, “Since AI can do some programs, it’s getting easier. We won’t need programmers anymore.” What I would tell people is that the opposite is true. We’re going to need more programmers. There will be people that we need for lower-level complex stuff you need a PhD for and things where you need creativity, good problem-solving, and good communication skills. Leveraging these tools to do things is going to be a universal skillset.

Let’s talk about a few of the end-product technologies that I’m most excited about. How far away are we from full self-driving? Is that something that you think is going to be realistic? More importantly, how do you think the people who are making this technology are going to handle all of the negative concerns that people have with ethical considerations and everything like that?

This is not an area of my expertise. I’m not working with an automotive company. I will share with you what I’ve learned from my education. On self-driving cars, in 2017, I bet my wife we would have self-driving cars driving around my hometown of Savannah by 2021. The loser had to keep scooping the kitty litter for the next decade.

Are you in Savannah, Georgia? I was married there.

Where?

At the Telfair Museum of Art.

We will have to follow up afterward because we’re donors and very active board members. We will have to follow this up. We’re great supporters at Telfair. Public service announcement, if you come to Savannah, please visit the Telfair Museums.

I love Savannah.

We’re going to have a chance to meet face-to-face the next time I’m in Boston or when you’re in Savannah. I lost the bet about self-driving cars but now, they’re going to happen quickly. The breakthrough that’s come out is companies are realizing they can use AI to generate the training data to train cars. One of the challenges of self-driving cars is there are so many edge cases. There’s the kid with a balloon versus the ball and the kid that runs and stops. It’s also interacting with bad drivers. It’s the other people. Companies have now realized they can identify these edge cases and generate tons of data automatically using generative AI that they can use and train the models.

Tesla and others are investing massively in this. This is going to allow us to bring the technology on much quicker because before that, there was only so much data of driving around Pittsburgh streets in the snow, down a hill, at a stop sign, and when it’s icy. It’s too much information. That is what’s coming. I suspect this technology is going to come quickly. Every incentive is driving this thing forward. Consumers want it. It’s going to make us a lot safer, and the industry wants it. There’s no faction aligned against it.

It was shocking to me even when they said Tesla’s self-driving car made a mistake due to some operator, and somebody died. We kept going. It didn’t shut everything down as you would have expected. Most problems in the world can be solved when you get infinite resources and incentives driving behind them. There are not too many problems. The cure for cancer might be one of them but when you align to every incentive possible, we’re going to get there. That’s why I’m optimistic.

 

 

Another one that I’m excited about is AI tutoring, which you highlighted a little bit. You invest a little bit in somebody’s education at the kindergarten or pre-kindergarten level, and there are lots of benefits down the road. Human beings are very similar to plants. You give them an algorithm of how to succeed, whether it’s a little bit of fertilizer or a little bit of water, and they’re going to flourish. The same thing happens with very young children. I have a two–year–old. I see her taking in information at a level that is still remarkable to me every single day. How far away are we from that? Is there anything that you’ve come across that gives you a lot of optimism about this? I do think it’s going to be important.

We’re there already. I find myself turning on voice-activated ChatGPT, “How long do I need to grill the steak on the grill for?” It will talk to me back and forth. We’re there already. The biggest companies recognized that this is one of the biggest markets in the world. It’s getting people smarter with AI. You have AWS putting $4 billion in philanthropy. I saw in the New York Times that OpenAI is now going to raise its next round at a valuation of $80 billion. OpenAI now is valued about the same as Boeing. That’s how fast this is happening. Google is throwing everything in the kitchen sink at this.

These companies are pouring everything in, and this is one of the most obvious ones. It’s going to be general purpose. Help kids learn smarter in any language regardless of background. They’re all in. That will be free general purpose ubiquitous everywhere. I’m beyond optimistic. I would bet everything that we’re going to have that. We’re talking about months, not years. That’s coming.

I was speaking with one of the policymakers for science education on another show. She was highlighting how the idea of talking and interacting is such a basic human instinct. That’s how we’re supposed to learn. It’s active learning, and active learning is so much more powerful than passive learning and watching a screen. Much of our last decade has been passive learning for young children. The transition to active learning she thinks is going to be exponential growth for humanity.

Honestly, I remember conversations more than something that I’ve watched on television. It’s the same for kids. Before we even had writing, that’s how we were learning. That’s something that will be powerful. The idea of talking to your phone and talking to Alexa and all of these different devices has been a great breakthrough but now, with the addition of artificial intelligence being able to give them any information that they want, that’s going to be huge. I’m looking forward to that.

I’m always writing scripts for YouTube, blogs, and proposals. I’m writing all the time. I now use AI to write. Typically, what I will do is I’ll open up my phone, start dictating to ChatGPT, and say, “I’m about to write a blog post on this.” I’ll dictate what I say and give it instructions, “After I give you a chunk, you’re going to go back and summarize and edit it.”

It’s a way for me to get a lot of ideas out there quickly without physically typing. I’m not getting any dumber. I’m just using a tool. Instead of banging away on the keyboard, I’m telling it to organize my thinking, and then I have to go back, edit it, clean it up, and everything because it’s still not that smart but that’s how I work every day. Once you get into the workflow where I can talk instead of sitting down and staring at a blank piece of paper, it is easier.

That’s something that I need to incorporate into my life. Academic writing, which is the main focus of my writing, is so difficult. I dread every single time somebody asks me to write a research paper or something like that. I hope that’s something that can help that field. We talked about self-driving. We talked about tutoring, which is also going to be a big deal for us. What are some technologies that I might be missing that you’re excited about?

The biggest one in terms of an impact that may not be as evident to a lay audience but will be clear when they think about it is the ability of these Large Language Models to access data and databases. If I open up an app in the airport, it’s almost like a United app. There are only two things I want. What gate am I going to? Is it on time? I have to log in, click on this, and go to these nine screens to get to these too-easy questions. At some point, I was like, “I’ll look at the board. I don’t want to deal with this thing.”

It’s not that the people making these apps are stupid or want to annoy us. Getting data out of databases is a hard problem. You have to use a programming language called Structured Query Language or SQL to figure out, “How do I get data from different places?” To be able to do this at scale, the data has to be organized in databases. That’s not easy. It’s one of the biggest challenges of any programming. How do I get the data out in the right fashion?

Large Language Models can now generate SQL from Natural Language Queries. I can say to my phone, “Which gate do I need to go to?” It can take what I’m saying, translate that into a query for a database, get the data, come back, reason across it, and give me an answer. That is the most generic and exciting technology because it means it’s going to change the interface. It’s not just about having a chat interface but it’s how we can access technology without going through the painful dashboards and applications like we do now. That’s going to impact everything.

Much of the information is stored in databases. People don’t realize that they just see the end product. All of the data that I have to sift through to write a scientific article or something like that is all databases.

How do you get it out? It’s not easy for people to do that. That’s a major challenge. Why is it so slow to find out my flight? It’s because United is processing millions of customers and they have a massive database. To find your flight at your time, that’s a difficult problem but it’s one that language models could do before you ask the question.

Honestly, I could talk to you for another hour but we’re getting toward the end of our time. I always end with three questions that I love to ask my guests. The first one is something personal to me. I always want to know where you gain your inspiration from. For me, it’s science fiction. When I look at the ideas of the future that are promoted by utopian science fiction like Star Trek, it’s something that I look forward to. I can’t wait until we have robot butlers so that I don’t have to do my laundry.

I feel like I’m making myself out to be super lazy but the human experience should be something that’s devoted to something other than tedious work. A lot of the visions of the future that you gave through artificial intelligence hopefully will be able to remove that from our experience. Where do you gain inspiration from? Is there any content out there that drives your day–to–day?

At the risk of flattering my host here, it’s podcasts honestly. It’s so practical for me to have this thing and put my AirPods in my ear when I’m walking the dog, doing dishes, or folding laundry. I always have a podcast on. I enjoy audiobooks and learning but generally speaking, it’s selected podcasts out there that I listen to and I find inspiration from. It’s usually ones where the host is very deliberate about how they think critically. That’s usually where I gain most of my inspiration.

What are the podcasts that you listen to if you don’t mind me asking?

Jordan Harbinger is a great one. Sam Harris is another one. The FiveThirtyEight Politics podcast, although since Nate Silver has left, is probably not as much as my go-to but what I enjoy about them is how critically they think about data and how carefully they analyze. That’s a great one. Peter Attia’s health podcast is one of my go-to’s along with his amazing book Outlive, and the Huberman Lab podcast. Those are my go-to ones if I have to pick from my phone. There are some Philadelphia sports that I will go to as well when I want to turn off my brain.

What did you think about Outlive? I finished it.

This will have to be a second episode. My other interest is longevity and curing aging. I’ve been a board member of the SENS Research Foundation for a couple of decades. His book was the best book I’ve ever read that marries interventions that you can do as a person and modern medicine. It does it in a way that’s evidence-based. One of the things I’ve always struggled with is finding the balance between the non-pharmaceutical interventions you could do as a person. Don’t smoke. Exercise. Wear your seatbelt. There’s a couple out there for nutrition and hot and cold therapy. There’s all this stuff out there for which the evidence is not completely clear.

Trying to figure out, “How do I prioritize? Where do statins, blood pressure, and colonoscopies come in?” He captures all this and helps you figure them out, “These are the things that are going to kill you. Focus on that.” That was very valuable for me because it’s easy. Even as somebody who studies this stuff all the time and works for the researchers, I get lost in the morass of thinking about sauna therapy and not putting it on that same level as getting a colonoscopy. If you stay back and look at the evidence, these things are very clear where you need to be spending your time. That’s what I loved about it.

He’s very practical. I had Aubrey de Gray on. I’m sure that you know he is one of the originators of the whole longevity movement. He was very much 15 to 20 years down the road talking about some of the societal implications and some of the technologies that will not produce anything but may be the answer like Yamanaka factors and things like that. He’s giving a prescription for somebody on how they can exactly live longer. I appreciate that. He’s also a physician, which I appreciate.

Second question, where would you hope that artificial intelligence would go so that it has the most benefit for humanity? That’s a very general question. I want people to know that this is a huge benefit. We have talked a little bit about the granular, “I won’t have to look at my flight tracker anymore.” Where do you think it’s going to have the biggest benefit?

When we have models that have ChatGPT’s performance that fit on your phone, that’s going to unleash a wave that we don’t even know how to think about yet. People are actively working on that now. Week by week, they get better and better. When you can have a model that has seven billion parameters and ChatGPT-4’s capabilities, the accessibility speed and what you could do with it are going to unleash creativity. If you do any practical thing, you have to send data to a server somewhere where these models are running and get a return. It introduces a lot of practical challenges for innovation. When we get powerful models on our phones, look out.

That’s interesting. I never thought about the difference between having something handheld versus relying on a server because that’s how self-driving is. Is it located in a car?

Most of it is on board the vehicle. A car is an expensive device where you could have customized hardware trained especially for these things and you could get downloads. I’m assuming that’s what happens when my Tesla gets a software download. I don’t believe they’re sending any data back but it’s possible I’m wrong about that.

Here’s the last thing that I wanted to ask you because we didn’t touch Prolego at all, which is cool because I love hearing experts and picking their brains. You’re a thought leader in this space. I appreciate talking about some other stuff but I do want to ask. Where do you see Prolego in 5 to 10 years? Where do you hope the business applications of AI go in 5 to 10 years or so?

We are going to be driving the change in how software is built in big enterprises. The overwhelming majority of applications built at banks, insurance companies, and any company have a lot of software they have custom-built for their operations. That’s all going to be rewritten using these generative models. We’re already starting to do some of that.

As a company thinks, “I have this legacy system that runs my operations I’m going to have to replace with something else,” they’re going to want to use generative AI to do that. That’s where we have our expertise because we do a combination of strategy, product planning, and actual engineering. That’s where we make our biggest impact. It’s helping companies make this transition.

I want more companies to start using artificial intelligence because it makes more value for the employees that are doing a lot of this mundane work. All the best to you. Good luck with everything. It was interesting to talk to you. We could talk for another hour about longevity, Savannah, Georgia, and all sorts of different things.

Let’s save it for a future episode or a meal together in one of our respective cities.

We have a live show in Boston where we have a recording studio. If you’re ever in Boston, I would love to have you on again.

Let’s do it.

To our audience, thank you so much for joining us. If you want to check Kevin out, he’s on different sorts of social media. More importantly, we will see you in the future. Have a great day, everybody.

Thanks a lot.

 

Important Links

• Prolego
• Outlive
• Aubrey de Gray – Previous episode

 

About Kevin Dewalt

Kevin Dewalt’s professional passion is helping the world’s largest companies integrate the transformational technology of AI. In 2017 he co-founded Prolego to help Fortune 500 companies take advantage of the biggest technology shift since the internet.

He has helped drive AI adoption at diverse companies such as Bosch, Raytheon, Cox Enterprises, T. Rowe Price, Lockheed Martin, Cloudera, Transamerica, Protective Life, and FINRA. His work touches every aspect of generating business value with AI, from state-of-the-art deep learning techniques to board-level strategies for $100M multiple-year investments.

In addition to his technology expertise, Kevin is also a highly regarded AI advisor to some of the world’s most influential business leaders. He is author of the widely read strategy book Become an AI Company in 90 Days. He also wrote the world’s first AI comic book, Adventures in AI. Globally, more than 30,000 executives have taken his MIT Bootcamp, AI in the Enterprise. Kevin has also been a guest lecturer at Berkeley, Stanford, and Carnegie Mellon.

Kevin founded Prolego after a distinguished 30-year career as a technology innovator, entrepreneur, and investor. At In-Q-Tel, the investing arm of the US intelligence community, and at the National Science Foundation, he helped make over 50 venture capital investments of more than $50M.

He is the founder of four startup technology companies and an early investor in more than 10. As an early strategic investor in Palantir, he helped founder Stephen Cohen to deploy the company’s first AI product in the US intelligence community.

As a board member of SENS Research Foundation, Kevin works with philanthropists like Peter Thiel to develop cures for diseases of aging.

Kevin was the top graduate in his US Coast Guard Academy class, with a degree in electrical engineering. His expertise in AI began during his master’s degree at Stanford University, where he completed original research under legend Dr. Bernard Widrow.

Although his lifelong dream of playing professional golf was prematurely cut short – at birth by an utter lack of talent – he and his wife enjoy playing golf near their home in Savannah, Georgia.