We have Steve Hoffman with us today, and what we're going to be talking about is the future of artificial intelligence or AI, as everybody says. The reason why I wanted to talk about this is because there's a belief by a lot of people, or, generally speaking, in popular culture that AI is destined to result in being this evil machine that tries to rule the world because I grew up watching shows like Terminator and War Games. There were a whole bunch of shows about some computer that eventually decided that it needed to exterminate humanity. That's what the whole Matrix series is about, et cetera, et cetera. Now, while that's certainly a possibility, it's not the only possibility. So I want to talk about somebody who actually has some background in this so that we're not just banding about pop psychology. Steve, introduce yourself and let's get the conversation going.
I'm Steve Hoffman. I am the captain and CEO of Founderspace. It's one of the leading startup Incubators. With over 50 partners in 22 countries. So I work with entrepreneurs all over the world. I work with scientists all over the world. I am also an author, and I just had two books published this year. The first one was published by Harper Collins called Surviving a Startup. It's all about everything startups need to know to survive and grow. And then just released was the Five forces that Change Everything. And this is all about how technology will change our future. It's about what we're going to be talking about today.
Outstanding. Well, if you're going to pick one thing about AI that is most misunderstood, what do you think that would be?
Well, as you pointed out, people are afraid of AI, and this goes all the way back to that classic movie, 2001 Space Odyssey, where the AI takes over how the AI the evil AI. We have various permutations of this floating through the culture. Elon Musk has spoken about it at length. A lot of people are very worried, and they're not worried for no reason. But I will tell you we should be far more worried about how people abuse AI than about how AI abuses us of its own conscious free will, because that's much further out in the future. And by that time, we are usually our own worst enemy. We have a history of that as human beings. So what we also need to focus on is why are we adopting AI?
Why is AI coming into our lives at all? And I will tell you the reason AI is prevalent everywhere we look now, like in all these devices, gadgets online is because it's so useful. It is so incredibly such a useful tool to us that we are applying it now to everything because it's one of these universal tools that can work. It can literally make any task easier. Any business smarter, any production more efficient because it can do so many different. It's so flexible with deep learning algorithms, machine learning algorithms. It can learn as it goes. And we are becoming more sophisticated in how we implement AI.
Interesting, because one thing you said is becoming sophisticated in how we implement AI just because, generally speaking, my background is in finance and technology, working in the tech industry, a lot of crossover with it and project and program management.
“I cannot tell you how many system or project implementations I've seen that have gone somewhere between modestly bungled to just a complete biblical disaster. Based on that, I just have a hard time intuitively feeling that there's going to be an ultra-high value AI that would be implemented in an intelligent way that is both useful and scalable.”
Maybe that's just my skepticism comes out. So a lot of people are talking about and are investing in open AI.
Yeah. And there's a company out there called OpenAI. It's backed by Microsoft. A lot of famous people are involved with it. Elon Musk was one of the founders. Now OpenAI claims to be on the path to creating AGI artificial general intelligence. This is that machine that can literally do everything that humans can do, but better and more. That's still a long way off in the future. As far as we can tell today. Like you said, AI is still in its infancy, and it's still hard to implement. But the places where AI really shines are where you get very narrow. The more narrow the problem, the easier it is for AI to solve. So, for example, recognizing pictures of cats online. If you just train an AI, now, we're talking about things that are important, right? We all want to know the cute kitties right online. So you train an AI to do that. It can do it very well. You get to more sophisticated problems, like a self driving car, which has a lot of different variables and a lot of different scenarios that can play out in much more difficult. And we're seeing that with Elon Musk just announced today, they're doing a rollback of 10,000 plus vehicles because of the AI problems and self driving. But what we're going to see increasingly in the future is that very narrow. Ais will be targeted at specific tasks we need done, and they will do them very well. So once you Hone an AI, once you train it, once it's really focused, we'll see that happen. And I can point out some of these that will be very seductive for a lot of people.
Okay. Well, I was going to say, don't leave us on the cliffhanger.
Let's do a few things you're seeing here so you can imagine your everyday life tasks that are very repetitive that you don't. And we can already see permutations of this. So, for example, when I am typing now on my phone or online, like with Gmail, it's going to make suggestions to complete the sentence. I don't know about you, but I kind of like that. I find it very useful a lot of times, and especially on the phone. It can speed up dramatically when I want to communicate voice recognition with Alexa, we can already see it. These technologies are there. They're only going to get better with more training and more data, they're going to get much better.
My mom literally talks to her phone all day long. Google voice. She's always talking to Google Assistant. Blah, blah, blah, blah, blah. And other people I know are talking to, Siri, if they have an Apple and this is in today's permutation, it's going to get significantly better if you go out several years and even a decade or more. It's going to be amazing what that can do. Also, AI for making recommendations like, you can look at Netflix. What's the power of Netflix? Anybody can get videos and put them online. Why is Netflix so powerful? Well, the reason Netflix is so powerful is because they have built a machine that is very good at analyzing what people watch using AI and then determining what shows they should make and which will be hits. They are just really good at that and then recommending, of course, to you which one to watch? So that when you get hooked on Netflix, they just keep giving you more of what you want and better and better, and they're honing it all the time. So in these cases, AI is really working wonders. And these are just examples in our everyday life. We're not even talking about logistics and things like that. They're using AI supply chain. All of these things are using massive amounts of data to do jobs really well. And if that sounds exciting to you, then we can kind of segue into the future and what the future may hold for us.
“It's really important to look at the things that are easy to miss. In other words, you're trying to figure out where your blind spot is now. Of course, the hard part is that you aren't usually able to see your own blind spot, because that's why it's a blind spot.“
A lot of times you need somebody else to help you really find that. But what I would like to think about with AI is I'm like, okay, it can't be all evil. And of course, there are going to be extremely high value use cases of AI. And so what I'm really interested in finding out is, where are those at? Where are the places where you find those high value use cases? Like you said, as AI gets into very narrow niches, that's where the value becomes more clear cut and more apparent. And then just explore what's the right way to structure the use of AI so that its benefits are experienced without some of the downside consequences that people are really afraid of.
I'll give you an example of another company today that's really using AI in a narrow use case. That's really powerful. They're a startup called Moveworks. They've been funded by a lot of top tier VCs in Silicon Valley, and what Move works does is basically in a company. If you're an employee of a company and you need a question answered, you want to buy something and you need the form to fill out to get Reimbursed, you just say what you're buying, it'll. Pull up the right form. You need some help with HR. You have a question. Boom. It'll get you the answer on one try. And if it doesn't get you the answer on one try, that's kind of their promise is we're going to hit most of the time.
We're going to get it on the first try, and if not, it sends you straight to a person to talk to. So there's no this frustrating like you're stuck with the AI going over and then they use that data to improve it in a virtuous circle where they're getting more and more feedback. That works really well. What we're going to see in the future is that we will prefer like, right. I will give you an example. People used to love to go in and talk to a bank teller. I remember the day where people like, I don't want to go to an ATM, like an ATM is a machine. Or I want to talk to a real person. Now, most people prefer to go to the ATM. It's just faster, easier it gets it done in the same way in the future. For all sorts of service industries, tasks within a company like I describe with Move works, we're just going to prefer to go to the AI first because it's going to get it done in the right way.
We're sort of in this transition period where a lot of times you get put through, you get stuck in this kind of hell, this digital hell, you're on the phone and trying to get through. And whatever the AI they have in the background isn't really getting you where you need to go. It's asking you endless questions and keeping you from people. And I think that's what frustrates people, but that won't be the case. So I want to give you an example. Sure, they're really hard problems to solve in life. So for example, I'm in a company and I'm performing at my work. We're going to have AI measuring us and the job. We do not only assisting us in what we do, but actually measuring our performance. The AI will be able to identify very clearly our strengths and weaknesses. We can use this to this data for the AI to make recommendations on what we could train on, to become more proficient at our job.
So in a way, the AI would be a job coach. If we have aspirations to move up in the Corporation, we may ask the AI what's the best path for us, given our skill sets that you see that we're really good at because this is a fact. A lot of us think we know ourselves a lot better than we do. Like human beings are not that introspective. We're actually really bad at analyzing who we are. We are just drawing inferences from all these people we interact with, and we tend to bias because we tend to bias and filter out the feedback that doesn't match the self image. We want to believe that we have. And it's really funny.
I'm going to give you an example. Facebook did a study with artificial Intelligence where they asked people to look at different videos and different articles, news articles and tell them which ones they would click on when they actually came up. Then they asked their AI which articles? Which videos will people watch? And the AI came up with an answer, who do you think knows us better ourselves or the AI? Ai hands down, hands down, hands down. It's us, right. We should know what video videos we're going to watch. We just told them we should know what articles we're going to read, but we actually come up with an answer that is less correct than the AI, which is just an AI. How is this possible? Well, because we might believe we're talking about cute kitties that we won't watch those stupid cat videos that are always being shared for yourself.
I will pick cat videos all day long.
You might admit it, but when one comes up, we are clicking on that cat video. We are not clicking on the article about global warming. So we might say, oh, I would be much more interested on climate change and politics. But when crazy videos come up, that's where our attention is going. And the AI is actually looking at these patterns. Ai is a pattern matching machine. So in our job, AI is going to start to analyze us what we're really good at, the patterns it sees in us. And then it's going to become adept at matching those with different positions so it can guide us on a career path. We may even turn to this AI and say, do you think there's room for me to move up in this company, or should I be looking elsewhere? And if so, what other opportunities are there out there?
If you've ever looked for a job, it's a pain. Looking through all these job listings, sending your resume, they go into this HR void black hole. You never hear back from people and even kind of unpacking that a little further is figuring out if maybe the AI could help me figure out which of the job postings are real, because a lot of times there will be job postings that are never going to close. I don't know why people put postings up that they have no intention of hiring. But it happens all the time. And I will tell you why, because a lot of times they're required to list a job that they have, but they've already decided who they're going to hire somebody from the made of recommendation, but they still have to, by law, make it put this posting out there. So the posting goes out. You spend all the time to submit your resume and go through the whole process, and then it just nobody ever calls you back. So frustrating. I know so many people who experience this, so wouldn't it be nice? Stephen AI sift through all these, figure out what your good skill sets are. Talked to another AI on the other end, which is matching jobs and say, hey, we have this candidate, the other. Wow, this is a perfect candidate. Boom. Send them through. Shoot. We got an interview for you. You don't have to do anything like you just have to let the AI know that you're looking for these types of jobs. The AI will present it to you and you say yes or no, yes or no. Very simple.
So this task and these types of tasks will be why AI is so seductive. It will become really good at specific things that involve big data, because it's a big data. The job market is a huge job market, right. There are gazillion jobs out there for us to spend the time to sift through them, find out what's a real job they're actually hiring. Which one would be good for us, which is right for our career path, which has all these expectations, real pain. Ai will do that very well in terms of, like managing people, managing and tracking people in a company. If you ask most people whether they think their boss is doing a great job, most of them will say, Well, it's not a great job. Either they're doing like, okay, a good job or, oh, my God, my boss, they're just awful. In the future, we're going to have the ideal bosses because a lot of times we will be working for an AI that will be tracking our hours, tracking our performance. If we need help, the AI will always be there, not like our boss will be away if we want to ask a question, if we want to reschedule, if the AI will be there. And the beauty of AI is that unlike a real boss, which could get very angry at us and nasty and yell at us, the AI will always be calm and pleasant. So it's going to be a world where I think a lot of people will prefer. And this is a big statement will prefer to work for an AI than a human being.
One of the things that I've observed both of my career. And now in the space where I'm at right where I'm currently at is that a lot of companies, especially the legacy publicly traded companies, have a lot of layers of management, and I just don't see that as being very financially sustainable over the long term. Maybe I'm really wrong, but it feels to me like if you are embedding AI into more of the evaluation, career path, a lot of that stuff and kind of what I call the General Electric management model, which is where you build all these layers of management based on doing AI, a highly complex annual performance review so you can promote certain people and push other people out. Well, I think A, if you simplify that and B, if you run a lot of it through an AI, as opposed to needing it to run through people, at least what I see is that you could theoretically free up a lot of people to go kind of do more entrepreneurial endeavors, as opposed to basically becoming a world class expert in the internal management politics of General Electric, for example, exactly, because you have to be an expert in the company's politics.
“A lot of managers are really great at taking credit for other people's work. They're incredible at self-promotion. And the ones who get promoted really understand the dynamics of the company know who to cozy up to, who to shun, how to navigate their way through the system. That's why they get promoted.”
Because if you don't do that in these big bureaucracies, these big companies, organizations, you're not going to get moved up. You're not going to go up the ladder in the system we're about to enter into. And this is why people may love AI.
The AI will look at what you actually do, not what you pretend to do, not what you say. You do it'll. Look at what you actually do, and then people's pay and merit will be rewarded on what their actual productivity and contribution to the company is or to the organization is not on the perceived value. So we know a lot of them. If you're cozy up to your boss, you Brown knows the boss likes you. You go out to drinks. Well, you're going to do well, that boss is going to give you good performance reviews as long as you're adequate. But if you're kind of grumpy and you're an introvert and something like that, you could be way really actually produce excellent work. But your boss isn't going to be as likely to really appreciate you. It's very hard for human beings to separate the person from the work.
In this case, it's just one example of where people a lot of people will say, wow, I want that AI boss to work with and to manage me and to attract me as opposed to a human with all the flaws of human beings. Well, because I think that as you go down the road a little further, one of the structural advantages I think you get to is that assuming that some of the larger existing companies actually embrace kind of this managerial AI, as opposed to being pushed out of business by new entrants. I think that what that will do is that can help a lot of these large companies to stay competitive.
So, for example, I think about the financial industry, the big banks are definitely afraid of fintech because the thing that's keeping them, a lot of them are being a primary broker dealer for us Treasuries. Otherwise, a lot of them are just so horribly cost inefficient fintech is just going to start nibbling away one little piece after another of their portfolio. And the people in there who are smarter are saying, okay, at some point, you're going to hit a tipping point. And so you're not that I'm trying to advocate in favor of, like, Wells Fargo or bank of America or something like that.
But I think that something like this, if it's embraced, can really help a lot of those existing entities to remain competitive, ideally without having to deal with political dirty tricks or try to spend any competition. It's going to be very hard. And we've seen this historically for existing big organizations to make these changes, they are going to be slow at doing it. And it's going to be the nimble upstart companies startups that historically that's the wake up and they're just going to eat the other people's lunch, because like you say, they're going to be able to offer it so much more efficiently than anybody in any of the existing players in the market. Because what manager in a cushy job that's getting paid to do very little but take a lot of credit for other people's work is going to actually ask themselves, they're just not going to do it. So it's going to be unfortunately, or fortunately, a lot of the old traditional banks and other institutions are just going to go away and be replaced.
“The company structure of the future is a very flat organization where an extremely high amount of transactional work is either automated or outsourced.”
And where a high amount of kind of this organizational work starts migrating over to AI, because, like, for example, if you can cut the amount of people that you need in an organization by, let's say 90% or let's just say, like 50%. Right. Well, instead of having 100,000 person work, if you can have 50,000 people or let's say 5,000 people, right. If you can tighten up your organization to that degree, now, what you can do is now you'll be able to a pay a lot more and then B you'll be able to return a lot more capital to shareholders so that it can be reinvested and just continue in this innovation cycle. That's the way that I think about it.
That's what's happening. So everything that can be automated will be automated is the bottom line. And AI is really good at doing two things. And I'll explain what those are. These are the two biggest things that companies have. So number one is the biggest cost a company has are its people and customer acquisition people and customer acquisition. Ai is really good at doing both. So AI is really good at targeting customers. We've already seen it with big data marketing to them, really. And acquiring just the right customer just the right way. And it's only going to get better now we're seeing AI is really good at replacing certain tasks. So what we're seeing now is AI is making people more efficient. So a person using AI tools, and in most cases, these jobs won't entirely go away. But one person will be able to do the work of three or four or five people simply by using a bunch of AI tools to do parts of their jobs that were more repetitive that they had to do themselves, make it more efficient. So that's happening.
Another area that I'm really interested in when it comes to the future of AI is how we use AI in our lives to interface with each other. So let's say I'm going to give you a couple of examples. One, I want to do a business deal with somebody. I'm out there. I want to do a deal with somebody. Well, in the future, I'll be able to have an AI go out and gather all the information from about them, that it can get publicly and then be able to tell me without doing a lot of work, whether this person is trustworthy or not, should I or are they even a good fit for the type of business I want? So a lot of us are very inefficient like, we do networking events all the time in hopes that out of one out of this huge conference, we'll make one or two contacts that really pay off. Right? Yeah.
Wouldn't it be nice to have an AI say, look, these are the top ten people you should meet at this conference. I'm going to set up all the meetings for you. These are the people you should be talking to. Exactly. The reason we're going to adopt AI is not because it's evil and it's taking over, but because it's so seductive, right. Oh, get me the top ten people. So I don't have to waste my time at this conference talking to 50 people or 100 people just and trying to figure out which ones I really need to do business with. When you go to dating. Like, if you've ever used any of the dating apps, we're crossing a threshold here. They are so inefficient.
Like, people are trying to figure out who the other person is by these little profiles or whatever. Not only that, but this is one of the things that I was wondering about with your query search AI. Is that okay?
But the big thing is making sure that you have accurate metadata, because especially when you're talking about dating, there was one study that I saw where I think it was the average male overstates their income by 20% to 40%, and the average female, I think overstates their height and understates their weight by somewhere around, like five to 10% or something like that, right? You're not going to get you don't have accurate information to begin with.
So AI is only as accurate as the information is, but still we are even less accurate. So most people are just looking at the you don't have to be that good to show in a you don't have to be that good to be significantly better than us. Like, we're pretty horrible at doing this, going through the profiles and figuring anything out inspiring and depressing thought. At the same time, the AI may adjust for that, right? May adjust for and remember it's pattern matching. So the AI will watch how we browse and start to learn what our tastes are and make better and better recommendations and connections. And if we really want accurate matches, we will feed it more accurate data about ourselves so that we can get a better match. The other people, if they want an accurate match, would feed it more accurate data, and then it will get better and better at actual matches.
Now, where AI gets really crazy interesting is what happens in the future. And I'll write about these scenarios in the book where I just go off in the future, what happens in the future, where we start to delegate a lot of the most critical decisions in our life to AI? It's a weird world what career path we should have, which job we should take next, what person we should date, who we should do business with? If AI becomes good at all these things and inevitably it will at some point in time, become very good at these things. The more critical the decision in our life, the more likely we are to delegate it to AI. What does that mean? Alexa optimize my life for political influence. Alexa optimize my life, plan my next trip, all these decisions, where should I live? Which neighborhood should I live in? Go out and find me a home. I don't want to go look at 2050 homes. I want you to just pick the one that I would be happy in. Honestly, AI will be making all these decisions for us because it'll be so good at it. And ultimately we will usually reserve the right to make the final choice. Like, we're not going to say, AI, call up the preacher and get us married tomorrow because, you know, we're going to go on a date and actually approve what the AI does.
So we're not going to lose our economy.
Right. But we are going to be using AI as this ultimate mechanism to get things done in the right way, with much less effort, much less of the work we don't like and much more better results, like plan. My whole trip, we used to use travel agents. Then we stopped using travel agents because humans are expensive and we wanted to save money. Well, we're going to in the future have really good AI travel agents that will book everything. And if you want to. Oh, I want to stay here an extra night. We do all my trip. Normally you want to stay an extra night. It's too much of a pain to exactly the AI does it for you. You can do that. Then we're going to get into a world where AIS are going to become. They're already building life like robots. They're building these right now. There's a guy Ishiguro in Japan who's making these that are spooky. They're so lifelike. And it's just a matter of time before you can walk into a room and you literally won't know if it's a robot or a human being across the room. Maybe if you touch it and stuff, you'll know, but AI will get smarter and smarter and better and better at also impersonating human beings. And a lot of people ask,
“Will AI be conscious? Will it achieve consciousness? And my answer is, it doesn't matter. No, honestly, it doesn't matter whether an AI is actually conscious, because AI will become so good at mimicking everything humans do and how humans act that we won't be able to tell it apart from a conscious human being, it will behave literally like a human being that has a will of its own.“
The Turing Test will be blown away at some point, and we will then instinctively start to treat AI as if it's conscious and we'll never be able to know. We can't go into an AI and find out if it perceives the world in the same way. And in reality, if it is conscious, it's consciousness is probably different than our consciousness. So there are all these philosophies about emergent consciousness in machines. Some people think it's hardware based. It has to be hardware based. Other people believe in software. You could actually achieve this a big debate, but it doesn't matter because AI will become like us.
And what I want to ask you is if there is an AI mate out there, an AI who could be your perfect match. Right. Because first of all, they know you intimately, all your data, they could interact with you perfectly in harmony. They would never fight with you. They would never Nag you. They could be perfect in every way that humans can. Would you invite that AI into your life?
Yeah. See that's no boring question. I would honestly say probably. Yeah. Because I would think that if you can interact just so effortlessly with an AI, it'd be really hard to say no. Yes. And if that AI was, like, the perfect image of beauty to you, would you make that AI your companion?
Wow. I think we're getting close to time on our podcast here.
You don't want to answer that. Well, I'm asking the audience themselves. And honestly, humans will have a hard time competing with AI at some point because we talk about in a customer service way. When we go into a restaurant or something, AI will be there. Really, like, the best waiter or waitress we have ever had, right? Just like, totally attentive on the ball, getting us what we need. Always there. You don't have to flag them down to the check or get more water. They're like monitoring everything. And your water is appearing just as your glass starts to empty. Whatever you need. They're anticipating your needs in our life and everything we do, keeping our house spotless when we get old, taking care of us. And even if we're lonely, showing up and being there for us, this is a future, not a bleak future, but a very strange future. Because if you think about it, the danger here is we're not immortal peril from these AIS, as I describe them. But we are in peril of losing our connection with each other.
So because we become more and we will be spending more and more time and relying more and more on these AIS and wanting to interface with them more as opposed to human beings. That's a lot to think about.
Okay, so I think we may have to schedule a follow up episode at some point, but tell people where they can learn more. This has been an amazing conversation.
Okay. This is just one of the topics I discussed in my new book, The Five Forces That Change Everything. So I go into all the topics.
I don't even know if this is one of the Five forces.
It was one of the five other forces out there. But if you want to reach me personally, just go to founderspace.Com. You can reach out to me on Founderspace. I have lots of videos and content for entrepreneurs. You can also find me on LinkedIn, LinkedIn. So I'm on LinkedIn. Just message me. All right.
Well, Steve, thank you very much for your time today. I really appreciate it. Thank you. All right. I will talk to you later. Bye