Skip to main content Skip to footer site map

Artificial Intelligence (AI): Chatbots, Metaverse, Robots, and More



View Transcript

Speaker 1 (00:03)

Welcome to Paychex THRIVE, a Business Podcast, where you'll hear timely insights to help you navigate marketplace dynamics and propel your business forward. Here's your host, Gene Marks.

Gene Marks (00:19):

Hey, everybody and welcome back to another episode of Paychex THRIVE. My name is Gene Marks. Thank you so much for joining me today. I am speaking with Brett Winton. He is a Chief Futurist. Brett, the Chief Futurist, I'm assuming, Brett. I can't imagine there's like a team of futurists, but maybe there is, at ARK Investment Management. So, Brett, you're out in California. Is ARK also based in California as well?

Brett Winton (00:44):

No, ARK is based in Florida and we have team members all over the country and I've been in California since inception of the firm. We were in New York, we've been in Florida, but really we access talent anywhere in the country. And so, the transition to remote work was very easy for us because we were already doing everything we needed to kind of asynchronously work and exchange documents and forecast.

Gene Marks (01:07):

I believe it. Yeah, no, I believe it. And is, tell me a little bit about ARK. It's an investment management firm, but is ARK proactively involved in making investments in companies that are of interest through their shareholders and their clients?

Brett Winton (01:25):

Yeah. Ark Invest is a company, we invest in disruptive technologies in the public and private markets. We're probably most well-known for our actively managed ETF suites. We actually kind of define the category of, hey, we are going to transparently show you everything that we hold and we're going to actively manage these ETFs and we're going to invest in what we see is the five innovation platforms that are defining this technological economic moment.


Brett Winton (01:55):

So, between public blockchains, artificial intelligence, energy storage, including how it's transforming the mobility space, what we call multiomic sequencing. So everything from being able to edit the genome to understand exactly what's going on in the proteins in your body, as well as robotics. All five of those innovation platforms hitting the marketplace at the same time, we think creates a unique moment in history. And that's why we founded ARK and that's what we've been doing since inception is investing and forecasting those.

Gene Marks (02:25):

It's very, very cool. A lot of it is long term and a lot of it is kind of shorter term actually, particularly when you see how fast some of these technologies are advancing. So, as Chief Futurist, what does that mean?

Brett Winton (02:41):

Well, so we have always differentiated ourself from other market participants by taking an intentionally longer term point of view when we invest in individual companies and when we forecast technology. Often people will invest in a stock based on what they expect to happen over the next year. We underwrite securities over five years and our technology forecasts extend out to 10 years and beyond. And so, there's always this kind of marriage between the top-down work that we do, the forecasting, the technology level, and then how that kind of fits into our expectations for the individual companies.


Brett Winton (03:25):

And so, as the Chief Futurist, I'm responsible for steering the ship in the right direction on the top-down research and figuring out how many electric vehicles are going to be sold, what's the future of AI, and really helping our analysts to forecast and model those events. So, I don't do all the modeling myself, but it's making sure that we have diligent processes for forecasting the future and understanding how big and valuable these technologies are going to be.

Gene Marks (03:58):

So, our audience are business owners. I went through your Twitter account, I pulled out some fun tweets, things that you have your point of view on. And I'm hoping we can get to that because it's AI and it's robotics and those kinds of things that really do affect our audience. But as you're talking, it strikes me that you're making bets on some longer term technologies. Over the past few years you must have had some disappointments. Like some technologies that you were betting on that seemed like they were going to really be something. And not that they won't, but either they're fizzling or they don't seem to be progressing as fast.


Gene Marks (04:44):

So let me give you some examples. There's been a lot of talk right now about the Metaverse and all the investments that Facebook has made. I mean they changed their name to Meta. I know for a fact because I have friends that are right now involved in the building of server farms that are bigger than the state of Utah in preparation for this metaverse. But people seem to have, not lost interest, but they're becoming a little bit more bearish about it. Give me your thoughts on that because that has a big impact on businesses and how they're going to be behaving in the future.

Brett Winton (05:22):

Yeah, I mean I think generally, part of the value of doing the actual work of forecasting is it helps you to separate the wheat from the chaff. And, so when Facebook bought Oculus in, I think that was 2015 or thereabouts, we did work on VR headsets and said, "Hey, how big is this going to be? What is the unit economics going to look like?" And we really couldn't get aggressive on our expectations, relative to what other people were forecasting because the chicken and egg issue between you need quality content specifically designed for the headset for people to buy the headset. But you need enough people to have the headset in order to generate the flywheel of developers willing to develop to it. It just didn't seem solvable given the state of the technology. And frankly it's still not.


Brett Winton (06:13):

I think that Apple's going to launch a $3,000 headset reputedly. That's way too high a price point for it to be a mass market product. I think that even they're ostensibly expecting to sell up to a million units and that would be a remarkable business success given the price point they're at. And it seems as if they're hoping that third party developers will figure out what this thing is going to be useful for. Which, to attract developers, to get people to do that, you need enough people that are willing to put the goggles on their face.

Gene Marks (06:48):

Get into the market. Yes.

Brett Winton (06:49):

Yeah. Yes. So, Meta's strategy I have always seen as just off the mark in that they were so subject to platform risk with Apple and Android over the course of their business life that Mark Zuckerberg has always been focused on, "I need to be in front of whatever the next platform and operating system is and assumed that that was going to get attached to hardware." I don't think that's happening this cycle.


Brett Winton (07:14):

I think that generative AI and all of the ways in which AI is going to change software is actually the platform that's going to change things. And so, by devoting all these resources to trying to build goggles that people will put on their face, he in some ways missed what was actually the technological transition that we're in the throes of right now. Which is, if you can just speak to a computer and it can respond in a natural language way, that changes the way in which we can interact with software and computation. And those kinds of user interface transitions are often where platform economics emerge. And I think that's where we're at.

Gene Marks (08:00):

So, it's funny that you say that. You made the right call on the metaverse only because you mentioned about hardware. I do agree with you that if Apple were to sell a million of these headsets, there is a market for people using these headsets. They're doing training with it, they're real estate agents touring people through a virtual home or whatever. But I don't know, man, unless I'm wearing a pair of glasses that looks like any other pair of glasses, and I can give it a command and switch into some virtual meeting and I don't look like a complete dope wearing some giant space, it's not going to happen.


Gene Marks (08:34):

And even the pair of glasses, the amount of hardware, the processing that has to happen in it, it's ginormous without it looking like... I mean even Google Glass made people look ridiculous. And that wasn't even as crazy looking as the VR headsets that are out there. It's interesting to hear you have that perspective because it's definitely a hardware problem. I'm going to tell you another hardware problem that I see, because you just mentioned with AI and you see it progressing much faster. Right now, when I talk about Chat GPT and AI to clients, it's great. And it's cool. And I realize it needs to improve its accuracy, it's still has bias involved, there's all that going on. But it's doing research for you, it's answering questions and it's doing a little bit of stuff for you. It can write out the email or write a blog or whatever. I'm curious your thoughts on when AI actually gets incorporated when it actually merges with IoT, with internet of things.

Brett Winton (09:35):


Gene Marks (09:36):

You actually have AI software on machines and factories that businesses can afford, not just Toyota or Amazon, that are able to adjust the situations, warrant of maintenance, take certain actions, trigger another machine to start because this machine has finished. Do you know what I mean? Even call an autonomous machine to come over to remove some products from it, to bring it somewhere else. How far off into the future... That's where I see my clients really, really embracing AI when it has that impact on how they make stuff. Does that make sense?

Brett Winton (10:14):

Yeah, yeah. I think that one, the rate pace of capability increase is so rapid right now that just taking the capability of AI systems today and then spinning up a bunch of businesses to apply that to all of these use cases, is happening right now. It's happening and it's not that kind of these generative AI systems are stopping, they're still advancing in terms of their capability. So, on the research side, academics have demonstrated, "Hey, if I give a robot the ability to interface with people with natural language, not only does that make it easier for the user to say, 'Hey, go pick up that screwdriver for me, but actually you can use that natural language ability to help the robot figure out how to pick up the screwdriver.'" Where by having the robot translate that into, "Oh I need to walk across the room, oh I need to look in this drawer for a screwdriver. Oh this is what a screwdriver looks like."


Brett Winton (11:15):

And so, actually the advance in natural language understanding, at first blush you'd think, "Hey, this will help me interface with the robot," But actually these same advances that are driving ChatGPT are actually helping robotic systems to interpret and navigate through the world. I think you'll see it in every endpoint device that has sufficient computation in it, you'll have some kind of AI model pushed against it to make it more performant and more easily customizable by the end user. And then at a very high level the cost of developing custom software and in the manufacturing space, that's a lot of the cost of installing an automated system is actually, "Oh, I have to hire in an engineer to figure out exactly how this robot arm needs to work and my factory..." That cost is going to totally collapse, because really your ability to develop software distills to your ability to be precise with your language now.


Brett Winton (12:22):

Software is really... The irony is people think of these things as, "Hey, this is a way that it allows computers to speak," but actually these GPT-4 allows us to speak to computers and that's what coding really is. And so I think that the ability of enterprises to take digital devices and actually have them do exactly precisely what they want them to do, is going to explode here. And so, in some ways it's ironic to me that in the markets it seems like everybody's paying attention to, "Oh, it's Microsoft versus Google." Whereas I think actually the more interesting exposures to AI from an investor's perspective are the companies that have these endpoint devices, they have the distribution in place and they can massively upgrade their capability by pushing AI models against them just with the existing capabilities today.

Gene Marks (13:19):

It's funny, first company that comes to mind and their CEO's on my list to interview as well, is Boston Dynamics. For example, I think of them and some of the robotics they have, they just released their Atlas and it's a robot that's walking around and you think, "Holy mackerel." I mean AI gives humans the ability to talk to that robot and for that robot to process and understand what they're saying and then do more advanced functionality. It's just that the hardware's got to catch up to it. But that can't take too long to do. And I'm assuming there are other companies you're looking at that are doing very similar stuff.

Brett Winton (13:54):

Yeah. And Boston Dynamics is an interesting case where my understanding of how... They have a super capable dynamic balancing and, think of it as very, very tactical, movement system. And yet over-

Gene Marks (14:08):

By the way, I don't know if you've ever seen their videos of the robots falling over each other and failing on their tests. It's very funny.

Brett Winton (14:15):

Right, right. And my understanding of the backend of their system is it's a lot more kind of marionetting along the way, or to date has been. Certainly for all of the demo videos they do where it's also almost custom engineering the system to do a somersault, so it looks really impressive. And it's clear the direction we're going is instead very unstructured, like instructions given to a robot and it's able to actually prosecute those instructions. Right now Tesla, Elon Musk is claiming that their humanoid robot is going to be more valuable than their entire enterprise combined. And they're throwing resources at it within the context of Tesla as in they're building prototypes, they're developing software.


Brett Winton (15:06):

There's a few other humanoid robot startups that have come out of stealth over the past few months. And the reason to believe it should or could work now is because you actually can get around mechanical limitations with better software. If you think about within the robotic arm context, if you have to design the robotic arm so it precisely picks up something with millimeter precision, that requires a lot of expense in that robotic arm. You need an actuator that is extremely precise and can get right to the right spot. And then on the other side you need the part it's supposed to pick up to get in the right spot. And so, it makes for a very, very rigid system.


Brett Winton (15:47):

But that's not how we as humans operate. We get close to what we're going to pick up and then we kind of wiggle our head around to see. So, the having good software, good perception and fine motor software rather than hardware actually allows you to do a lot more with a lot less expensive components and then that makes manufacturability much easier. And so, you should expect that you'll see much more performant hardware automation systems that are more flexible, not because there's a massive advance on the hardware side, though people are working on that, but instead because the software is just advancing so so quickly that you can get over hardware limitations that were otherwise essentially capability killers within existing robotic systems.

Gene Marks (16:37):

Interesting that you say that. Yeah, so I heard Elon... Elon Musk was interviewed by Lex Friedman on his podcast. If you've ever listened to Lex Friedman's... which is fantastic. And this was like a year and a half ago and even Musk was admitting then that for Tesla to have their self-driving vehicles, it was a lot harder than they expected it to be. That there was a lot going on to make a decision when a self-driving car is going down the street and a child kicks a ball in front of the car and which direction of the car... There was a lot of processing that had to go on. And he said the biggest challenge that they had at the point was processing powers, hardware, you know what I mean? It had to figure all that stuff and process it that much faster. It wasn't as much software.


Gene Marks (17:22):

But now it seems that that problem is quickly getting solved because AI itself is becoming that much more efficient at running, that the hardware challenges are becoming a little bit less. I guess the same example is... Do you remember Microsoft would release new versions of its operating system back in the 90s, way ahead of the hardware and you would have to buy new computers just so the hardware could catch up to it. And today it seems to have caught up. My question for you is, with all the hubris around AI, do you ever sense that there is still a hardware challenge that's still we're facing? I mean we're not making enough chips in this country where...

Brett Winton (18:08):


Gene Marks (18:09):

You know what I mean? We can almost be stopped because we don't have the hardware to process this stuff the way it needs to be. Does that concern you at all?

Brett Winton (18:19):

It doesn't concern me, but it's definitely true that right now open AI would prefer their users to use ChatGPT less because they just don't have enough GPUs on the backend to run it. And so, there's clearly an opportunity for providing just the chips to run these models. And to give you a sense, the total IT spend is between four and five trillion dollars a year total across all enterprises, all their comms spend-

Gene Marks (18:54):

All and that is just fixing printers by the way. I'm just kidding.

Brett Winton (18:59):

But we think that hardware to support AI software, just the compute chips to support AI software will be $1.4 trillion in spend by 2030. And so, there's a clear need for a lot of AI specific chips out there. And what's kind of on the language front and with what Tesla's doing, there's huge innovations in just taking very big models and figuring out how to get the same performance out of a much smaller model. It turns out you can actually overstuff these models with data and allow the model to run on smaller hardware than you would otherwise expect. And the open sourcing of these language models means some people are managing to compress them enough that it could run on a laptop.


Brett Winton (19:49):

And what's really remarkable about it, if you think about how much data is compressed into these models, the fact that I can ask GPT-4 to tell me about Roman dodecahedrons and what they were useful for. And they're these little metal dodecahedrons that people don't know what they're used for and ask it to speculate on that, not only will it know what I'm talking about, but it will have a structured set of these are the things that they could be used for. Some of which have been hypothesized by scientists, some of which haven't. And that is embedded in this trillion parameter model or we don't know exactly how big it is, but actually it's probably less than that, is really remarkable and indicative that you could run that on a smaller system. It would just take you a long time per word.


Brett Winton (20:44):

And so I think, in some ways what these systems will do, they will let less powerful hardware appear more powerful. That's it. At the end stage, Tesla has its chips, it's installing on cars. It really needs those to not be water cooled. It needs them to run in an automotive environment. So, it has to be a very small chip that they're getting the performance that they're getting out of their system. In my Tesla in the garage that's four years old now, it's really amazing and that it's still improving. So it's the opposite of the Microsoft example where they upgrade the operating system and it breaks the entire hardware. Here it's like I have a car that's four years old that's still getting better based on operating system advances. It's a very different paradigm.

Gene Marks (21:33):

That's amazing. That's amazing. Okay, another question on ChatGPT. My understanding of ChatGPT is, as fantastic as it is and we've got obviously more versions coming, people are getting super excited about Auto-GPT. And I don't know if a lot of business owners or people listening to this understand what it is, but it sort of takes the chat out of the ChatGPT and starts doing stuff. It's like GPT doing stuff on its own, which has scary implications but also incredible implication for productivity. Can you give me your thoughts on Auto-GPT? Do you think it's going to be as fantastic as everybody says it is? Do you think that it's going to replace ChatGPT, that type of conversation?

Brett Winton (22:14):

Well replace it? So the general idea-

Gene Marks (22:14):

It's different.

Brett Winton (22:18):

... here is, hey, can I turn a chatbot into an agent where I can say, "Hey, can you find a new business opportunity for me and spin it up and begin executing on it?" And so then the chatbot goes off and we'll begin accessioning APIs and you give it a budget and it maybe spins up additional chatbots that you serve as customer support and it builds the business for you. And I think it's clearly a direction that we're going where you are managing AI agents working on your behalf to fulfill business functions. And so, there's a reason people are excited about it because then it kind of, not only are you no longer writing the software, you're actually not even product managing the business. It's finding enterprise value and extracting it.


Brett Winton (23:11):

And it's early in terms of the capability. We actually have one of our analysts try to spin one up internally, he called it BrettGPT. And it was lost in a loop of trying to get company information and then looking at stuff that's totally extraneous for what it should do. And so, we pulled the plug on it. But I think there's clearly a kernel of an idea that when it works at scale, it becomes very large, which is why people are interested in it. And just on the capability of language AI agents today, in terms of how they impact education and customer service and all of these functions that are really actually big important economic functions, we're going to see a lot of innovation before you even get into the agent out there operating in the world independently staging.

Gene Marks (24:06):

You guys make investments in disruptive technologies. You are taking a long-term angle on this, but let me ask you, I mean I guess without divulging any secret sauce here, our audience or business owners that are listening to this, they want to know what's coming down the pike. What should they be aware of? I guess my question to you is where are you investing your money from a business perspective that would impact businesses, that you think will have an impact shorter term? And by shorter term, I mean somewhere in the next three to five years so that if I'm running a business, you're basically telling me, "You need to pay attention to this, we're investing in it, it's going to have an impact." Can you think of an example or two?

Brett Winton (24:50):

Yeah, I think that internally, every one, certainly our analysts and on the operations side, has been empowered and encouraged to play with the AI tools that are coming to market and figure out ways to do things more efficiently within the business, based on the AI tools that are available today. And to think carefully about how we're going to use the data that we generate to further customize the way in which we interface with the world using AI. And so, I really think that enterprises that don't adopt AI aggressively, will get out competed and die. As in this is a incredible capability upgrade. One way to think about it, if I need copy or something for marketing copy and I wanted to contract out to a writer to do that marketing copy for me, previously I would've paid around 10 cents a word and you-

Gene Marks (25:52):

You just tweeted, this is very true. Keep going.

Brett Winton (25:56):

Yeah, but using GPT-4 or Anthropic Claude, which is an API competitor, I can do it for 2 cents per thousand words. So it's like a 3000 x cost decline in the ability to generate copy. So on the one hand that means, hey, the customer service function, the marketing function, all of that changes. The way in which you advertise to people can totally transform because you can literally write ad copy for the specific end customer and Facebook or Meta will help you do that and everything else. It also means in so far as your business that has some kind of business function that implicitly assumes that somebody sending a letter to you is the equivalent of $20 in sweat equity or $40 in sweat equity, if your inbound channels are designed in some way around the fact that, "Hey, this is somebody that spent time writing this, therefore it must be something I need to pay attention to," you're going to get totally overwhelmed and swamped and that as a signal of importance to you, you have to figure out a way to tune it differently.


Brett Winton (27:07):

And so, I think there're going to be a lot of businesses, and that includes all the security risks that that entails, but also just even... If you imagine you At&T and customers can come to you and be like, "Hey, I'm thinking of switching plans," and AT&T turns around and says, "Hey, we'll give you a $20 discount on your monthly payment, will you stay?" And right now that's the optimum business decision for them because that customer who's agitated enough to go to them to say, "I'm thinking about leaving," is a customer who's also active enough that they actually are going to leave if you don't give them a discount.


Brett Winton (27:42):

But now if everybody gets that kind of complaint feature for free, they have to change the way in which they manage that interaction. Because previously, giving the discount was wise because you reduced churn and it was a small subset of customers that are doing it. Suddenly it's the whole economics of that interaction change. And so, I think it's true across businesses and society that there's an implicit connection between written material and implied human thought. And that's going to get severed and that means that the written generated material won't be a good signal as to whether or not there's action implied behind it to the same degree.

Gene Marks (28:28):

A lot of times clients are asking me... I'm a CPA if you haven't figured it out already by the way I look, but we do technology services as well in my company and as our clients are asking about ChatGPT and AI in general, my advice to them is... These are smaller companies or midsize companies and my advice has generally been to, "Listen, don't drive yourself... You don't have to develop your own AI, you don't have to become some kind of an AI expert. You need to talk to your software vendors that you're relying on your business systems."


Gene Marks (29:01):

I mean they're all using QuickBooks or Sage or Salesforce or Dynamics or Zoho or Asana or whatever. And my belief is, and this is what I want to ask you if you think my belief is right, is that the major business software application providers, they're making the investments to incorporate AI and AI tools in their current and future iterations on their platform, so that they can get their customers to use them. And I think the smartest way to go if you're a smaller company is to pay close attention to what they're doing and to ask and to embrace it. But developing this stuff on their own seems to be... I don't know if it's cost effective because they're not in the software business. Am I giving them the right advice? Do I need anything else?

Brett Winton (29:52):

I think it's true that every software as a service company will have to incorporate AI. And so, all these systems, they will spin up kind of AI... Their software will-

Gene Marks (30:05):

They'll be embedded.

Brett Winton (30:06):

... increasingly be defined by what AI can do. And I think that one, that software doesn't touch all of a company's business. And I think as a CEO or a manager, you need to think about what are the parts of our business that aren't softwarized today that actually should be and will be. And I think that there's a... From our perspective, we invest in technology, we're in the financial services industry. Within the research team, everybody we hire now has to have some degree of coding familiarity and background. And it's not that they're going to come in and write code, but now actually maybe they are. I have associates who are writing code because GPT-4 makes it so much easier to do it and ship it and you can use Replit and you don't even need to worry about the backend. If you have a pain point inside your business, even at the interface between two software as a service products, previously your options where you'd go to some kind of-

Gene Marks (31:12):

Zapier or...

Brett Winton (31:15):

Zapier or if you're bigger, maybe you go to Infosys, go offshore and be like, "Develop me some middleware to stitch these two things together." And that's brittle and it breaks and it's expensive and it seems cheap, but it really is a terrible process.

Gene Marks (31:28):

Takes a lot of TLC.

Brett Winton (31:29):

Yeah, yeah. You are, I think, better off funding that on your own balance sheet and developing a custom piece of kit to stitch things together and really being closer to the metal of what's going on with your business. And I think that's really... The direction that things are going is maybe there's monolithic softwares or service suites like Microsoft and then a lot of micro custom services that are enabled by AI. And so, I think that to the degree that as a business owner, you can get in front of that or at least play with it. Think about customer service function or how do you deal with inbound email volume? Even just using the consumer facing ChatGPT interface, you can increase your productivity and you can have your entire team increase their productivity if you tell them to use it and you say, "Hey, listen, try to use this to make yourself a better, more efficient employee." And what's amazing about the technology, it's so easy to play with and use.

Gene Marks (32:42):

All right, great answer. We're almost done. Before I let you go, I've asked you 10% of the questions I was planning on asking you, as usual. We never you even got a chance to go through some of your tweets. But this is your job, is to look for opportunities to invest. People that are doing stuff that are really cool things to do and can be really great for your company and your clients.


Gene Marks (33:04):

For a business owner that wants to keep up on all the things going on in AI, which is what you do, I'm curious if you've got a couple of any recommendations. I follow a guy on Twitter, I don't know if you know Rowan Cheung, I never met the guy before. He's a great Twitter account. He's always listing out all sorts of different stuff going on, new applications and new developments, whatever. And I think he's great. I'm curious, where do you go and who do you read? Just a couple, if you can share with us, a couple sites or a couple names that might keep our audience better informed, without giving away any proprietary secret.

Brett Winton (33:43):

Well, I follow all of our analysts and so there's a list of the whole ARK team. You can follow the whole ARK team and that's where, in all of these technology areas-

Gene Marks (33:53):

Can the public do this too, by the way?

Brett Winton (33:55):

Oh yeah. But that's actually the real answer is from the beginning of ARK, one of the things that we've done differently from other asset managers is to publish our research and to say what we believe is going to happen and to put the information out there in the world. And the reason we've done it and the reason we're more transparent than anybody else is because that actually makes us smarter.


Brett Winton (34:19):

We publish on the intersection between bitcoin mining and energy storage, not because we're showing off bit because then lots of Bitcoin mining and renewable energy companies and oil companies come to us and say, "Hey, this is how we're using that. Have you considered that?" So actually the way in which we tune the world's information engine to give us material, is by publishing material into the world. And so, I think particularly on Twitter, which is a very useful platform for it, people that put out really tangible specific information about what they're doing or struggling with on Twitter, actually attract a lot of information about that thing.


Brett Winton (35:04):

I think it's more than just following someone and there's plenty of accounts that they list. There's also lots of viral or people who are chasing traffic or "These are the 20 things you're doing wrong with ChatGPT." And I think that a lot of that is kind of marginal utility. But instead saying, "Hey, I'm trying to use ChatGPT to do this, this is how I'm doing it." And that'll attract a lot more attention and engagement. And in tune specifically to who you are and what you're doing. And by the way, you can ask ChatGPT to help you as you're doing it. So, that's like the other kind of dialogue that can enter into the chat is from the artifacts themselves. So, that's how-

Gene Marks (35:57):

So, some Twitter people and then also the analysts at ARK Investments. And it's Am I getting it? Is that the place to start?

Brett Winton (36:07):

Yes, or just follow me on Twitter and follow everybody I follow. You'd do well with that.

Gene Marks (36:10):

Sounds good. Brett, thank you so much. I've been speaking with Brett Winton. Brett is the chief futurist of ARK Investment Management. He can be found on Twitter at Winton Ark, right? W-I-N-T-O-N A-R-K. Great account. We'll keep you up to date on all things going on in technology in general, not just AI. Brett does not limit himself to that. Hey man, thank you so much for joining. It was really a lot of fun talking. And like I said, I have many more questions for you, so maybe we can talk again in the future. Okay?

Brett Winton (36:39):

Great. Love that. Thanks Gene.

Gene Marks (36:41):

Do you have a topic or a guest that you would like to hear on Thrive? Please let us know. Visit and send us your ideas or matters of interest. Also, if your business is looking to simplify your HR, payroll benefits or insurance services, see how Paychex can help. Visit the resource hub at That's W-O-R-X. Paychex can help manage those complexities while you focus on all the ways you want your business to thrive. I'm your host, Gene Marks, and thanks for joining us. Till next time, take care.

Speaker 1 (37:18):

This podcast is Property of Paychex, Incorporated. 2023. All rights reserved.