Good morning. Now, in my defense, I’ve only known that I was going to make this presentation for about nine months now. So obviously didn’t quite have enough time to pull the whole thing together. And I kind of set myself up because of the topic, I kept telling the organizers that there’s no way I could turn in my presentation because things were changing so rapidly that I’m not exactly sure what I was going to talk about. And then considering that it was the early morning session, and the second day and an unknown speaker, I was really kind of hoping that you know, the 10 of us could gather around the table chat.
So I now see that this slide is probably not going to cut it. Give me a second to see the touch up if you have the others and you just kind of talk amongst yourself and just move some things around. I think you’re really gonna like this one.
Right, oh.
(Computer Voice)
Good Morning, esteemed professors and fellow entrepreneurs. I am thrilled to be here today to discuss a topic that is shaping the landscape of entrepreneurship,
the future of AI in startups. As an early stage seed investor, I’ve had the privilege of witnessing the transformative power of AI firsthand.
All right, well, that’s, that’s a start. So if I were to just use AI to create this presentation, that’s kind of how I would have started, it then would have come up with an outline of topics that you’ve already heard of, and outcomes that are already foregone, and it wouldn’t have been terribly interesting. But I’m sure you all would have been polite and pay some attention to me. But just in case you weren’t, and you find yourself buried in your laptop or your phone screen, that’s okay, too. Because I would just assume, as I will, during this talk, that you’re just wanting to share your knowledge that you’ve gained through the conversation with your colleagues on threads, or whatever other social media platform you’d like. And this is where you can find me there.
The AI then also suggested that I introduce myself, but I figured you’re here. I’m here. You either know who I am, or you don’t. And you can always look it up later. So skip that part for now. And and then it also did suggest that I thank you for being here, which that’s seems pretty appropriate. I do appreciate that. Although we can see we don’t have any of our a students that showed up. Maybe one, but that’s alright, as well.
So to kick things off, I just want to kind of get us on the same page as far as where we are today. We live in this world of big tech. We have these large companies that have basically been built on this almost unwritten rule that if you collect it, and you store it, then you own it and by it I basically mean our data. And so each of these companies have come up with new, some may consider it nefarious ways to continue to collect even more and more of that data. They then feed that data back to us with small little windows of either products or services. They call them feeds and user generated content and social graphs. But each one of those is really just there to collect more of our data that they then sell to advertisers. And that then sell back to us. Now, just in the saying that, you can tell, I believe that that is our data and not theirs. But we don’t have control over that. And we also don’t really understand how much that’s worth, some probably think is worth a lot more than it is, and others may not think it’s worth anything at all.
But recently, we had a bit more insight into this. Just last month, Reddit did their initial public offering. And in their filings, they noted that they generate $800 million a year in revenue. Now to you and I $800 million, seems like quite a bit of money, nothing to sneeze at. However, in Reddit’s situation, this is the number three most trafficked website in the United States. Number one, and number two are Google and YouTube, both Alphabet companies, and they generate billions and billions of dollars a year, whereas Reddit does not. Now also in their filing, they were very proud to announce that they had a new line item under their revenue, $60 million was coming from, Yes, you guessed it, alphabet, to use their data to then train their new AI models. So we now have a price on at least what that data is worth to one of the big tech companies. And we also have the situation where Reddit doesn’t create any of that. It is all their users that generate that information that has created this value of $60 million to Google or Alphabet. And of course, maybe worth more to to others. So you have to start asking, Is this a beginning of a breaking point for, you know, our awareness of the value of our data?
Now, in this talk, I’m going to discuss a lot of the promise and hope and possibilities of AI. And yes, I’m going to focus on the positive aspects of that. That is by no means to say that everything is positive, or that we’re ready to do all of it or that things are perfect. So I wanted to just throw a few things out up front here to to note that I am aware of the challenges, I’m not ignoring them, but want to keep things positive on that.
One of the big ones is the training data that’s being used to create the current generation of LLM. It’s been said that we may have already created the most human centric version of a large language model, because AI is now being used to create much so much more of the content that’s available online. Now, my problem with that is the fact that we use content at all to create these models. What is content? This is the information that we are posting publicly out onto the internet. And why do you post content on the internet? Well, it’s almost exclusively used to garner attention. And when you have that as being your primary goal, then it’s much less about who we are, and much more about who we would like others to see us as. So whether that’s a company website that’s doing sales and marketing, or your individual posts to your friends to show them how great your life is, even though we’re not talking about the other pieces. That’s what the models have been trained off of. Another way to look at it is basically every large language model is that kid in the classroom, who is willing to do anything to get the attention of the teacher, and distracts everyone else from the learning experience by raising their hand and squealing and being obnoxious. That’s what our large language models are today.
Another challenge within it is the models are all trained off of the output. This is the end result. This is where we’ve gotten. There’s very little information inside of the models about the process that was used to get there. What questions did we ask that we answered and decided not to do? What twists and turns were involved to get us to that endpoint? And just averaging all of that out isn’t good enough to get the results that we need to interact with computers? In a more human way, this picture is a great result of this. This picture was the result of a prompt that asked for a meal in a bowl that includes all of the foods. Now, I don’t know about you, but I’m not really eager to try the mango mint chili that I have displayed on the screen here. Though the end result may be factually correct. I’m assuming there’s some cheese in there somewhere to get my dairy
My vision as part of the future that we’re trying to create primarily focused on software. I know there’s a lot of other aspects of startups and industry. But this is this is one that I’ve been the most focused on. First key concept to that, I already mentioned it a little bit, is we have to have the ability to own our own data. This is really the thing that is keeping us attached to big tech, SaaS models, industry, and at a core right now. Now, there’s a lot of reasons why we haven’t, both from not understanding the value, and that the infrastructure is not there, how will we integrate that. But both of those challenges are at a point where they can be resolved. So just imagine, in the future, instead of you having all of your data stored on Facebook and Amazon and Google and the like, you have a single storage unit that it’s held in. And then you have your own personal AI, that will help to authenticate and allocate that data to the other services that you want to use. So this isn’t about how can I pick up your data and move it and use it and create a copy of it, that kind of defeats the purpose. It’s the ability to make direct quote, requests based on your data, only getting the information needed to facilitate the solution that you’re required. And some of this will be even zero knowledge proof where the provider doesn’t need to have access to your data, they just need to be able to authenticate that your data is correct.
And so the example that’s used a lot in cryptocurrency and blockchain is an example where you’re going to the bar and the bouncer looks at your driver’s license, the bouncer now knows your name, your address, your birthdate, that’s a lot of personal information that you’re handing them. What the bouncer really needs to know is, are you 21, or under. That’s it. That’s what we’re talking about here is your personal agent would be able to authenticate that for the bouncer say, Yes, this person is 21. And over this person is not, no need to give your name and address and phone number and firstborn child, you can come into the bar, think about how many different scenarios that that can affect and what that would do for the privacy of our data, and then also the control.
So much so that, beyond the control and use of it, there’s even opportunity for compensation, let me know how much that data is worth to you to use. You want to train a new model, that’s great, I can allow you access to all of my presentations that I’ve created. And the conversations that I’ve had, where I’m authorized, have other users authorize the use of their part of the conversation, then we start getting to real data that’s human generated for these next models.
Next up, building on this is the end of SaaS. Now, we’ve had this model for quite a while now. However, prior to the mid 90s, this didn’t even seem like a possibility. It was unheard of that I would pay you to use a piece of software that was stored on your server, and that I had no control over. And it’s become apparent over the years why that was such an important challenge for us to get beyond. But here we are. And so now your data is stored in probably millions of different facilities all around the world. And most SaaS products are very similar in what their capabilities are and what they’re doing. It’s really just a window into your data interface that allows you to connect with that. And the building blocks of those are very similar. There’s a user interface, there’s that ability to enter text data into a database, they do some sort of reporting, and and then they just specialize on the type of information. For what I do, I have three different services just to track the startups that I work with, I have a CRM, I have a specific service that’s used to track the investments that I make. And then I have file storage for all of the data. Now the CRM and the the investment tracking, both allow me to upload files into them and store them there. I don’t use that service. What they don’t do is allow me to link into my file storage, so that I have a complete set of information in one place, I have to go to each one of those. But they’re all providing very similar services. It’s just divided up based on the data that I choose to give them. We won’t need to do that in the future, we’ll be able to take our current solutions that we now call SaaS, break those down into the individual components that they are built on and then assimilate those in a more dynamic way.
This will then lead us to the third point, which is our appless future. Now, I’m not sure if Appless is actually a word, I did register a domain name on it, I thought it was kind of cool. But I also like the fact that it references one of the biggest contributors to the challenge that we’re having in this, Apple themselves. Imagine if you were in an industry, where by making one change, you could add 30% revenue back into your finances. That’s the situation we have for every industry that has an app out on the app store right now, that’s a lot of incentive for us to figure out a new way of doing things. So with these components, we will have the ability to create a dynamic exchange that will measure the value and the quality of each of those outputs, and then bring them together to create what we now call an app. Again, your personal AI or a linking of AIs, will be able to do this, basically on the fly. And that’s been the linchpin, we’ve heard this one before, where we have object based programming, and you can pull components in and out. But with each one of those historically, you’ve still had to have a human in the loop that could pull that together, make the connections, upload the program, upload the app, deploy it to the web, today, that’s no longer the case, we can have an agent that will build that for us. So the front end of this has been called ambient computing, and we’ve been working with it for years. And we kind of begrudgingly do it. And we don’t ask much of it, because it’s pretty incapable. So if any of you have used a Google Home, or an Alexa, you know, the limitations of that. It’s great for timers and playing some music, sometimes as long as you don’t want the exact song. And heaven forbid, you tried to turn on the light in the room with you and not get your whole house to light up.
But we have a new version of that. Who here has heard of the Humane AI? Can? Anyone? A couple, okay. Just to get you caught up, this pin was just released this past week, the first reviews are out. The titles of those reviews are things like worst product I’ve ever reviewed. And would it be cool if it actually worked. So we’re not there yet. But the thinking behind it is in the right direction. This is an AI pin that you wear. It has a laser projector built into it so that when you hold your hand out, it will project an image on your hand. Now most of the interface is built around voice. And it does not connect to your phone. So there is no app integration. Now what it’s lacking is the back end infrastructure that would allow it to actually be functional. A big miss, but you know, we got to start somewhere. I have literally ordered one of these just to keep it in a box so that we can sell it to the museum in the future and say, This is what we thought an AI hardware would look like. Now where this is going, though, is to create that component exchange where the pin would be able to take your request or a query. And then either have this built ahead of time based off of, you know, general requests that I’ve seen in the past, or potentially even do that dynamically. So how this changes things is instead of me having to go to my phone, thinking ahead of time, download Uber download Lyft create an account or relationship with each of these organizations give both of them my credit card information, give both of them access to my phone, which by the way they monitor. One way, if you’re ever looking for a cheaper rate on that, it’s bring up both apps because they both look for the other app look running in the background on the phone. And they’ll be able to know and detect that. And sometimes you can get a cheaper rate. I digress… So instead of having to establish those relationships, each one of those organizations will be able to compete in the component exchange for the best way to get a driver from location A to location B. And there will be other components for credit card payments, creating a user interface, GPS location. So the way that will look in the future is that your driver will be able to register on the network to say I’m available, here are my credentials. I’m willing to share my location with someone if they’re, if they can authenticate that they can pay for the ride. And here’s the data that I’m willing to share on their interface. They log in, then you as the consumer, get to create your specific query based off of whether you’re going to the airport or you’re going home or going to work. And then you will be able to build that application in real time to go out and get that reply. So each one of those components will be able to be a separate provider that is competing to provide the best service for that solution.
Now in the title, and for the audience, education is a big part of this. I’m not here to tell you how to do your job, I’ve looked at the agenda for the conference, there are people who are far more qualified to do that than I am. But what I would like to do is throw out a couple of concepts that I think are important slash interesting to consider, and maybe just some, some things to do some some workshops and thought experiments around. First one is, right now, if nothing else, we can look at this AI technology as the ability to empower humans to be creative, by eliminating the need for technical skills. Now what I mean by this, imagine, if 10/20 years ago, I would have come to you and said, I know an elder in a village, where there’s only 2000 people that speak the language, and there is no written language. But today, they can tell a story, and have a novel written based on that story, and generated in multiple languages within a matter of minutes. That’s pretty powerful. We could do that right now. And that’s never been able to be done before. You take one model, get one translation of that language, and we can spread it everywhere. Other applications of this, artists, painters, sculptors, digital artists, they all go through a process of learning the skill and the trade of that artistry. Now, painters have already been through this before, right? We have photography, that’s going to end the world of nature painting. And obviously it didn’t, it created a new form of what we now consider art, the same thing will happen with an AI only it opens it up to almost everyone. Now, the most important aspect of this is we need to stay focused on the AI’s ability to help us be creative, and not asking AI to be creative for us. And that’s kind of where we’re seeing some of the challenges right now is you just put in the prompt, that’s the only interface we really have right now. And we’re hoping for these fantastic unexpected results from that. That’s not where things will be in the future.
The next concept, this is a big one, and as big as the next one, but the big one, what does it mean to know something? Now, in my uneducated group of tech professionals, but not educators, one of the ways that we look at this is knowing something is our ability to recall. We don’t have direct access to our memories, I can’t just show you what it was that happened. And so any knowledge that I share is really based on my ability to recall that information, use my imagination to fill in the blanks of the stuff I don’t remember, and then express that to you. So we’re doing a lot of it. And, you know, as you as you kind of look at that none of those things that I’m talking about doing really create an environment of reliable information. But it’s it’s the one that we have. So what happens now, when we have the ability to augment that recall, with a digital version of our memories, a recording of this talk, a recording of the conversation you had earlier this morning, video were or pictures of all the things that you’ve done, that are now streamed to you through a device that perhaps is attached to your body? Or maybe it’s only a layer of fabric between you and your body, which we’re at today. How does that change things? Does that mean you know it? Let’s assume it’s fast. Let’s assume it’s just piped into your brain. It says it was your thought. Now, let’s take that a step further. What if you have a stream of consciousness of everyone’s memories that’s in this room? Or on this campus? Or in this world? Does that mean you know something? What is it that really defines knowing? the speed is going to be there the information is going to be available. Wasn’t our original thought, but none of them are when they start. This is one that’s much bigger than me and probably us in this room, but it is one that we’re going to have to tackle and it’s one that you’re probably dealing with in education right now. now trying to define, is it okay to use chap GPT on an assignment? What can we do to lock it down? And then what does this mean for our for our future?
The final one I have here is critical thinking. The reason why I bring this one up is along with creativity, this is one of those things that humans need to keep hold of. It’s something that given the right situation we can be really good at. If we give up effort on it, we can be really bad at it. And it can cause a tremendous amount of problems. Another one of the ways we look at this is in conducting critical thinking, one of the first things that we do is to measure quality, I get some information, I look at that, and I consider whether or not it’s it’s valid, it’s good. And that spectrum is broad for different situations, sometimes lower quality is fine, because I’m going to get a lot of information, sometimes high quality is enough to pass and continue on. But once I define that quality, then I also have to be able to understand what to question. If it is low quality. And that’s not good enough, then I need to trigger another process in that. I had a situation over the holidays, where I was talking to a few, let’s call them young adults, that had come up with a statement that I felt was probably not true. And I did not agree with to say the least. So we were having a discussion about that. And in this discussion, they continue to justify their statement through deeper and deeper knowledge inside of the same article that they started reading where they got the headline from the knowledge. And I tried to explain to them that I understand that that article was written to support the statement that it made. But what else did they do to verify this information. And they very adamantly let me know that they didn’t need to, because this was the source and the source had all of that information. So I of course, went out and use my Gen X knowledge to Google some of the items on that. And very quickly learned that I do not believe the information was correct. But just that situation, impressed upon me enough to add this into this presentation. And to bring it up to a group of educators. This is us. This is what it means to be human. This is why you can look at a picture and think something’s off. But I don’t know quite what it is. And just that notion and being aware of that, and going out and and reviewing is a part of what we’re going to continue to need to do. And then less will know how to question well, you have to know what to question. But then how do you question that? Where do you look what is authoritative? What’s valid? That’s getting harder and harder to know, because we have more and more information out there. This is this is one of those finer points that we have to stay focused on to to keep the the human side of this of this AI.
Now the other part of the topic here is startups. This is my world. I’ve already told you the things that I think are the main challenges or changes that are coming, at least in the software side. But I will discuss a few things that are more directly related to startups because of that. The first one, and this one is becoming more and more obvious, we’re going to have smaller teams, you’re already seeing big tech do downsizing, this is going to roll into other startups. But it may not just be for the for the reasons that are being talked about today. Now, of course, with smaller startups, we’re going to need less space. We of course already have remote work and future of work and that side of things. But we’re also going to have pieces of the business that we no longer have to do. It’s not just that we’ll be able to augment the things that we are doing with AI, there’ll be things that are that are eliminated. And we’ve seen this happen before in industries, it’s not anything new. It’s just kind of the pace that we’re at. And of course, the fact that we’re living through it today. So it’s more meaningful to us than to others. But there are things that just won’t exist anymore. If we have this exchange for these components for these solutions, then there may not be as much of a need for marketing. You won’t have to win based on your ability to garner attention and to get in front of customers. Because the infrastructure will be able to weigh and measure the quality that you’re able to provide. You have more of a level playing field that will then open things up for a startup versus a large incumbent that are there because they’re storing our data and it’s really hard for us to move somewhere else. The other thing that a smaller team will do for companies is it allows you to have a stronger culture. You don’t have to expand that out into more and more employees over time and layers and layers of bureaucracy. That small team can be a group that shares your vision and and, you know, beliefs for how things should be done. And you’ll also be able to focus on on those together.
This, of course, all will create an opportunity to do faster iteration. We thought we could create things really fast when we moved to cloud computing. And we had this network across the whole world. We haven’t seen anything yet. And I’m a big believer in Lean Startup methodologies and customer development and work with your customers early. But with the development changes that we’re seeing now, it’s becoming much less resource intends to just build and test than it is to continue to work on what might be the best solution for what you’re trying to solve, doesn’t mean it’s going to go away. But the resources are shifting and shifting very rapidly, where you will be able to put this up in front of maybe a test environment and through your production actually generate the inputs on what to change, more so than just the feedback from your users. Again, we can’t lose it all. But it will be a part of that. This will also allow teams to specialize into a specific part of that complete solution set. So again, any SaaS product that you’ve gone into, and it’s like, well, you don’t have as good or reporting as these guys or your user interface sucks, or, you know, this isn’t as reliable. All of those individual components will be available to many of the teams. And then those teams will all get what’s going to happen. This will allow for the teams to compete on their quality of their service. And then if we all own our own data, that will significantly reduce the switching costs or concerns of this company has my data, I don’t want to take the time to export and import, which never really works, I already hold it, I’m only giving you access to the pieces that I think you need to solve the problem that I’ve requested, I no longer have to worry about where’s it at, and where’s it going to go.
With this, then it will allow teams to be compensated on the value that they’re adding to this network. So again, it’s no longer necessarily just about a land grab and getting more people onto your network and critical mass and all of the things that we’ve come out of on the mobile, social, local, cloud based computing side, this is adding true value and true value per transaction. And this will all then culminate in a service that each individual component will be compensated for what we’ll be measuring that all along the way each and every time. This will give teams a lot of versatility options, you can be the best at a certain aspect of the process, and also analyze the network for weaker components that maybe your team is suited to go after. And you will be able to then compete within that realm. It also allows us to spin things up and shut them down much more rapidly. If we need fewer resources on the front, then we can also do that with less funding, and we don’t need as big of an outcome.
At the end. This will create opportunities for new funding sources. Right now, current venture capital is all based on a need for a large amount of money upfront to burn in order to get to a point where you can then succeed. But what if we flipped that model where there’s much fewer resources wanted early, smaller teams, better opportunity, less sales and marketing, so that we can focus on revenue earlier, and funding that mechanics to then share in that revenue. There’s a new term for investment called Shared Earnings. And this is a combination of the earnings that the company has and the earnings that the entrepreneur has. So as the company is successful along the way, you can share those earnings with investors. And they can then build that into their models so that we don’t have to have an all or nothing network like we do today. The venture model is really messed up. I mean, I can say that I’m in the space. But the irony in all of this is if any startup were were to come to a venture capitalist with our business model, we would kick them out immediately. Right? Here’s the deal. I’m going to take your money, I’m going to give it to other people. And that’s my job. These people are going to be so good at what they do with your money that they’re going to be able to give me more of that money back and And then I’m gonna give you a cut of that, and I’m gonna keep a copy of it for myself. That’s just not the way that we do business. But it is the model that we have today. I think with this new infrastructure, we’ll be able to align those interests much better.
So, taking all of this, where do we go next? What do we need to what are what are those next steps we need to be thinking about? First concept to keep in mind is our initial knee jerk reaction to any new technology is to apply that technology to the old solution, and call it new. And that’s what we’re seeing right now. Right, you can use AI to create content and to create images. And to do this and to do that. And it’s different, and it’s cool. But we’re really struggling to define why it’s better. AI is super expensive. There was a study that came out a couple of months ago, looking at generative AI for creating images. And with everything that’s involved in doing that, the way that they helped us understand it is the final conclusion was that every image that is created, would use up about 13% of the average smartphone battery. That’s how much energy is going into every one of those images. At that point, we’re getting better, it’s getting easier to do. But just knowing that it’s a little bit of foresight, why Stability.AI is probably having so much, so many challenges, why their team has left, why their CEO was gone, and why they will probably be the next to go away. And that’s a company that’s been around for a couple of years now and burned through hundreds of millions of dollars. It’s cool, it’s different. But we’re not sure why why it’s better.
The other important thing is just adapting our thinking. And this is something we’ve had to continue to do for a very long time. But it becomes more and more important today, both with the capabilities of AI. And then the parts of it that we focus on. So one of those, again, is measuring the quality. This is one that we’re struggling with all inside of the enterprise. Why are companies concerned that employees who are doing remote work have three full time jobs or working for two other companies. It’s because they don’t know how to measure quality. We’re coming from and agrarian to industrial environment where butts and seats were our main measurement of the quality of work that someone is doing. And we haven’t gotten to the point where we do a really good job of measuring, identifying that and measuring it. And this is one that we have to be really careful with. Because we’ve also heard stories of organizations that are really good at measuring quality like Amazon, where employees talk about feeling like they’re just robots, and every piece of their entire life has been measured and filmed in control. And we’ve taken the human out of that. But it’s still important for us to know, as we look at the changing in the infrastructure, being able to measure that quality will allow us to get to the point where we won’t care if a human does it, or a bot or a human managing bots, I know that this is what I need done, I need it at this level of quality. And here’s how I define it, I’m willing to pay this amount to do it. And I need it done and need it done in this amount of time. And with those three criteria, as we get better and better at measuring and understanding that, then that’s where this component exchange will be able to put that together for us and create those solutions so that we have the SaaSless, Appless future that we’re talking about. Another thing it’s important to do and adapting your thinking is, to some degree, we need to assume that the infrastructure will be there. Every one of these technologies and people working within them have the same limitations. And by only developing a solution that fits within those limitations, you limit your potential for where things can grow. Now, it doesn’t mean that we can just assume everything’s going to be perfect. You’ll never release, but you have to know. And so that will be an interesting one for the human pin that is underperforming right now because of various reasons, many of which I don’t think are being addressed right now in the media as far as why it doesn’t work or why it sucks. But it’s it’s working towards that and within their model is the assumption that by highlighting the limitations, many more people will start working those limitations and then they will come to pass. And it’s a it’s a difficult balance. You can’t be too early maybe right on time. You certainly don’t want to be late. But we’ve seen it time and time again, when we look at dial up modems to broadband to mobile to 5g, everyone’s having those challenges, we now don’t even think about video streaming on a device that you’re carrying around, you know, as you’re walking, and making video calls to your family across the world, that future is coming to AI. And AI will never get any worse than it is today. And it’s getting better and better on a weekly basis. It’s pretty crazy. And then finally, this intersection, the understanding of what are we the best at, and when are we at our best, they get this piece of kind of review, critical thinking, measuring quality, that is very human, there are other aspects of that creativity, we need to focus on fostering the growth of those skills so that we continue to be even better at it. And we don’t give that up. And then also tuning the technology to augment that and facilitate our use of it, instead of hoping that it will replace it. And that is going to be one of our greatest challenges is there’s no easy answer to that. And we’re not quite sure what that means or what it looks like. Because we’re so used to the interfaces we have today. And kind of the general the competency of technology as a whole. That that is going to be, you know, a big concern.
And then finally, when are we at our best, I think the answer to that is we’re at our best when we’re working together. And so all of this remote work and future work elements, all of it is being tied back to the opportunity for us to get back together in person. And maximizing that time. Understanding that we can be at our most creative when we’re feeding off of each other that that is what drives the relationships and the future opportunities and our ability to then create the next version of whatever this may be. And it’s one that we have to be really focused on on not losing. So whether it’s a conference like this, or how you conduct things in a classroom or at home, that human connection side of things is one that again, it’s gonna be really easy for us to start rationalizing why we don’t need it. But we have to keep it as a part of what we’re doing.
I’m gonna leave you with one final thought. Last month, Sam Altman was talking to Alexis Ohanian at a conference. And he noted that in the very near future, we’re going to see the first set of companies that have 10 or fewer employees, but are still billion dollar companies. We haven’t seen that happen too often. There’s been a couple of cases of that number that’s in close, we will WhatsApp at 16 or 17. But this is talking full employees, not just founders, like No Pay Pal, they have like two dozen founders or something crazy, but actual total number employees. And the other part of the quote here is that they’re already placing bets internally, on what year do they think we will have the first billion dollar company that only has one employee. And that’s something we’ve never even talked about. Now, hopefully, it’s not so far out that inflation is just so high that having a billion dollars now what it once meant. But I can’t imagine being an entrepreneur going into this marketplace with this type of opportunity, but also pressure of being the first to do this. And looking at that. What does that environment have to look like where an individual or a team that’s that small, would be able to accomplish that goal. A lot of things have to change. We still have, you know the future to make for ourselves and see what that what that will be.
So that ends the prepared part of the presentation. Ai worked out pretty well for me. I appreciate the help. Here’s my additional contact information. And then happy to open things up for questions if we still have time.
We’ll open it up for q&a.
QnA
Remember, they gain value out of it, and they will keep coming. One of the ways that I’ve measured the success of that group is we have five different organizers in that group that also organize other AI meetups. So this is the one that they come to, to connect with each other. And then they take that information to go out to their community. It can be that simple. And again, it started with five people. And that’s it. There’s no big broad broadcasting of it. I’m not trying to market it. So when people show up, this is coming to you are very informal, in that. I do have some podcasts, I listen to mostly general tech, just to give me some some direction. There’s the pivot podcast with Scott Galloway and Kara Swisher, that I really like the verge helps to keep up on tech talks. There’s a few AI ones, they tend to kind of ebb and flow. I’m I’m much more interested in ones where you have a presenter with an expertise that they’re sharing with you instead of one that just has speakers come on, because those tend to be a little salesy. So that’s just kind of my personal preference on that. But yeah, I do listen every week and, and, you know, just kind of gives you some directional knowledge of where to go. But that that local group of interested individuals is a real opportunity right now. I don’t know where the mic set? No, unfortunately, it looks like we’re out of time. By five minutes, okay. I try to keep
trying keep my, my answers more concise. Okay, with
which percentage of AI startups coming up next, whatever, 123 years will really be disruptive. I had someone who FinTech startups, it turns out 90% are just standing sitting there has to do something better and underwriting customer service everything. So what’s your thing, which are potentially disruptive? And what percentage will just improve something somewhere?
I think we’re following kind of typical startup metrics on on that, you know, most are going to fail. I think some of the best near term opportunities are with teams that have actually been out in market for a while for a couple of years, usually SAS, but they’ve been collecting this data. And they also have access to that network of individuals or organizations. And so they’re able to apply the new technology to their process, and, and have greater opportunities within efficiencies, but then also providing that last piece of quality on top of it, anyone that’s just doing a wrapper around open AI, that’s the same thing that anyone else can do. And that’s why the GPT store or you know, app marketplace hasn’t done anything, it’s just mostly trash. And so you really have to have access to that data so that you can train off of the last kind of last mile last portion of the success criteria for your client. And that’s where we’re seeing the most success right now. We see a lot of great ideas on how we can have AI for blah, and I’m just not interested. Love to ask you questions.
It’s not a question, I just want to validate you and leave so you can relax. I want to go back to the previous point that you made about how you learn about AI and our faculty. And last semester, last year, actually, a year ago, I started using AI in my class and teaching with AI. And a colleague of mine asked me how do you know how to teach with AI? And I said, I don’t know anything about it. And he said, How can you teach something you don’t know? And I said, Isn’t that the base of learning? Teach? If you want to learn something, you teach it. But what really validated my sort of my fears was, since none of us really know much about it, the collective intelligence seems to help. So what I do, I have a group of students who know I’m interested in AI every day, they come in late, they give a little prompts. And they say, look, what I brought to you when I learned about this, do you know about it, and sometimes I pretend that I do that I don’t always want to look at the dumbest person in the room. And turns out accidentally, I became an expert in AI in the last three months. And this is going to be my sales pitch. I’m gonna give you a little book that we’ve been working on with our students. But what I was going to say is, you said that you don’t want to talk to educators. Because you know, we have a different perspective. I think we have very, very common perspectives. When it comes to AI we all want to learn and we have to learn from each other and I really love this idea of the meetup and I’m gonna start with the for me that was in Boston. So thank you so much. I really appreciated that.
All right, make it a good one.
Enjoy the presentation. One of the things that I found was a little troublesome was this idea of a single owner billion dollar company and social implications of and particularly when we think about innovation We think about innovation, whether it comes from diversity in diversity of inputs and changes that occur within groups. I’ll cite James March 91, talking about that as a very important thing for growth. So what do you see as the downstream effects of that?
I think the number one is a billion dollars isn’t what it used to be. You’re talking about a billion dollar company that has a market that the big tech trillion dollar companies won’t even touch, that’s not worth their time. And so there will be opportunities for much smaller companies, organizations, individuals to be successful at owning a part of this. And then yes, some of them will be even even more successful. But I think the lines are going to blurred between startup and entrepreneurship and employee and earnings and income and how that all works. You know, if you’re managing a team of technology, that’s that’s solving a problem accomplishing a goal, that might be something you do inside of a larger organization, because they’re providing you some infrastructure. But with the tools available, you might be able to just do that on your own. And we might not know the difference between, you know, what is own your own company and employment. So I think that the opportunity is that more people will have access to do more things. Some of them will be able to be more successful or come together in a way that we don’t recognize now. But billion a trillion is a big margin. And there’s also a lot of multimillion in between there that are opportunities that the big companies just won’t focus on it all. From Washington. Surely. But thank you so much.
Would like to thank you
okay, thanks everyone. We have a few few more exciting sessions left. I know they haven’t found it yet. But the next session in this role is
