In this episode of the Mile Marker podcast, Stacey welcomes Mike Demler, Semiconductor Technology Analyst & Strategic Consultant, for a discussion on Artificial intelligence and the how the evolution of AI may impact the future of fleet management and fleet automations.
https://ridecell.com/
https://www.linkedin.com/company/ridecell/
https://twitter.com/ridecell
In this episode of the Mile Marker podcast, Stacey welcomes Mike Demler, Semiconductor Technology Analyst & Strategic Consultant, for a discussion on Artificial intelligence and the how the evolution of AI may impact the future of fleet management and fleet automations.
https://ridecell.com/
https://www.linkedin.com/company/ridecell/
https://twitter.com/ridecell
Stacey Papp:
Welcome to the Mile Marker Podcast. My name is Stacey Papp, and I will be your guide, taking you on a journey into the world of fleet automation and shared mobility, focusing on innovations for businesses with fleets.
Joining me today is Mike Demler, a longtime semiconductor industry veteran, Technology Analyst, and Strategic Consultant. Over the last 10 years, Mike has authored numerous in-depth analysis of the innovative technologies driving advances in AI, ADAS, and autonomous vehicles. He co-authored five additions of the Linley Group Guide to Processors For Deep Learning, along with The Guide To Processors For Advanced Automotive. He now offers his insights as an advisor to clients across a broad spectrum of the technology industry. Today, Mike joins us to talk about artificial intelligence, a topic on everyone's mind. And how its evolution will impact the future of fleet operations. Mike, welcome, it's so great to have you here.
Mike Demler:
Hi, Stacey. I'm really excited to be here and to do this podcast. Thanks to you for conducting it and to Ridecell for inviting me.
Stacey Papp:
Absolutely. So let's start off with AI, again, a topic that's trending these days heavily. You can't really turn on the news or listen to the radio or even a podcast without that phrase coming up. It's been around for a long time, but it seems new. So can you start us off by explaining the evolution of AI and how it's impacted technology specifically?
Mike Demler:
Yeah, yeah, I'd love to. You're absolutely right. I was in the grocery store and the cashier and her bagger were discussing it, and I was in a clothing store and people there were talking about it. And all of a sudden I'm having AI discussions while I'm checking out with my groceries.
But you're right, it's been around in one form or another for a while. And in fact, if you look at the history of computers, even before my time, of the early computing machines that were mechanical. Their inventors were just so got blown away by some simple capabilities like being able to do, say a mathematical problem, addition, they're adding two numbers together. But they thought, well, that's something that only humans could do. And so before we know it, that in a few years these machines will be able to have all the intelligence of a human. And of course they were way off, way ahead of time in their predictions.
But then as things evolved, we got more into the electronic age. I think one of the pivotal points was in the fifties or even before that, during the World War, you know Alan Turing? You may have seen the movie The Imitation Game. And the reason he called it the Imitation Game, it's now known as the Turing test of whether somebody could be fooled into believing that a computer was actually a human and behaved in the same way that a human would. And that's still the test of the so-called artificial intelligence that you can now have in a computer.
And I think for me, the key was researchers for many years have been studying the human brain and looking at, how do we develop knowledge and intelligence? And the brain is composed of something like 80 some billion neurons and each of those neurons is basically a little computing unit-
Stacey Papp:
Sure.
Mike Demler:
And they're connected, they're connected in massive ways. But it's analog. And so the next key point was you can try and model that in an analog way, but it was the transition to the digital technology. And so for example, cameras went from be film-based to using digital sensors and having pixels. That's a format that lends itself very well to a computer doing analysis on it. Or for audio, for speech. And what we see now with say, ChatGPT of converting speech or text into a digital format. And now the computer can analyze it.
So those are some of the key things that have happened and the AI's been in use now by the average person for years, and they don't even think about it. It's the face unlock on your phone, if you get a recommendation from Netflix or Amazon, those are all AI based applications. And now we're seeing the next pivotal point, this explosion of interest generating by ChatGPT. And we're at the beginning of a new era where it'll be used across all kinds of industries.
Stacey Papp:
Well, you made it sound less scary because I'll tell you, this seems to be a very polarizing topic for people. You and I were talking when we originally met, that AI is now being used in the graphic design space, specifically around things like Photoshop. That you don't have to have a photographer take a picture of a sunset on the beach in Nantucket, you can just ask for that and it's right there. And I was just talking to a coworker not that long ago about when you buy clothing online and just with a few questions about your style and your height and gosh forbid your weight, but all critical things. They can kind of mimic your body shape and you can virtually try on clothes. And while that seems really cool, because I think a lot of us are still thinking it's so much fun to not have to leave the house, that's a little different, that's a little bit of a pivot from how we're used to consuming.
And like you said, in the tech space with face unlock on your phone. I remember when Apple rolled that out, I thought, "Oh, this is going to be so weird." And now it's like, you mean I had a home button at one point in time? It just feels so different. So the evolution kind of picks us up and carries us along with it, rather than us being pulled along.
Mike Demler:
Yeah, yeah. And sure, sometimes issues arise, unexpected consequences and whatever that need to be dealt with. But yeah, you always have to weigh the benefits against some of the other issues that might come up. And we can certainly discuss that for this topic, as well.
Stacey Papp:
Absolutely. So I want to pivot a little bit to how this impacts fleets. So can you share with us some of the benefits that you imagine machine learning and AI would deliver to those businesses who rely on fleets for their daily operations?
Mike Demler:
Yeah, yeah, sure. I think it is an emerging technology for a lot of the fleets, unless ... excuse me, you're one of the big companies. I mean, you can imagine Amazon or FedEx-
Stacey Papp:
Sure.
Mike Demler:
They've been employing AI across their operations for years, there's no way to manage, it's a massive operation. But for smaller businesses, I think we're seeing just some of the functions like say, going back to cameras, advanced driver assistance systems. This is a technology that's increasing safety in passenger vehicles, as well. Having cameras in the vehicle that can detect the blind spots, provide emergency braking functions. Overall, improving the safety by bringing that AI powered camera into the vehicle. And I know there's some issues about privacy, but even a driver monitoring system that would allow you to detect if the driver had a medical issue. Or be able to make some measurements of the productivity of that driver.
So I think use of the technology overall is just beginning, but it has a great amount of potential because not just using it reactively, but using it predictively and using it to optimize operations. That's where I see it going in the future.
Stacey Papp:
So I think that's a really important topic to keep continuing. So a lot of fleet managers rely on that data that their vehicles deliver, in order to navigate their day-to-day with a lot of time spent deciphering those numbers rather than making strategic decisions that impact overall operations and really, help them to grow the business. What are some of the common data points you see AI being able to help fleet managers interpret in order to make those strategic decisions?
Mike Demler:
Sure, sure. Yeah, I mean, again, it goes with, you've got all this data. I mean, AI starts with the data. And going back to say ChatGPT, how did it get so powerful? It's because along with the digitization of information, computing capabilities have advanced tremendously. To be able to connect thousands and thousands of computers to analyze much larger databases. And that's beyond any human's capability to do that.
Stacey Papp:
Sure.
Mike Demler:
And so I imagine with the telematics, sure you can get some information that tells you where the vehicle is or how much fuel it has left or how many miles it's been on the road or whatever. But it has sort of reactive. And I think that one of the ... and I've seen this in other industries, is using the data for machine, in this case, it's a vehicle. To predictively determine when it needs maintenance instead of waiting for something to fail. And it's interesting doing some research on specifically how some of the fleet managers are using this, I see a lot of weariness about doing it that way or that would cost more to bring in this software. But think about the cost of having a truck go down in the field or have an accident because of a part broke.
There are other examples of factory automation where they're using predictive maintenance to save lots of money. But there's other things, as well. The example you were using of being able to be on a shopping site and have basically what the industry calls a digital twin. Well, this is being used in factories. BMW is an example, where before they build a new factory or to optimize one that they have in place, they have a model of it. And so imagine that you have a model of all the routes, whatever you would say that's the city and you're doing waste management or whatever your type of fleet might be. You can have a digital model, and rather than wait for an emergency situation to occur, you can actually simulate it ahead of time and be ready before it ever happens.
Stacey Papp:
So you also touched on another critical point. So you keep transitioning into these topics, which is fantastic, is the human role in all of this. And it's very apparent and crucial that humans play a really big role when it comes to managing fleets. You talked a little bit about driver monitoring, there's still, for right now, drivers powering vehicles. Someone's still getting behind the wheel in most cases, outside of an autonomous vehicle. When it comes to fleets, it makes sense that AI can help manage and interpret that vehicle data. But again, going back to drivers, what about them? Do you think that AI will eventually replace that human component when it comes to managing people versus managing data? Or will we always still need that human touch in this process to not totally lose that role that an actual human being plays in ensuring the success, the safety of fleet based businesses?
Mike Demler:
Yeah. Well, I think we can talk about a few different ways, I mean, we can get into whether it'll replace the human driver in the vehicle. But I'm thinking more for the nearer term, it's not a matter of replacing, but more in augmenting and improving the capabilities of the fleet manager. Take an example of emergency response. I grew up in Buffalo, New York, and it's known for snowstorms. And you may recall this past winter, they had one of the worst ever ... and people died as a result. And I was just reading a news article where they're just now doing the analysis of what happened. And they're saying, well, there was a communication issue that they had the right equipment, but they weren't communicating properly.
Well, if they had a potential twin of the various routes and the type of equipment and snow removal and emergency vehicles, and it was a coordinated effort. You could have probably been more prepared. And in a situation where rather than reacting, I don't know, maybe there's somebody in an emergency call center that's trying to sort through all the information that's coming in. It's just not possible, it's just overwhelming.
Stacey Papp:
Sure.
Mike Demler:
And so as an example where you're a fleet manager, but you have these tools available to you to help you do your job much better. It's not going to replace you, it'll make your job easier, maybe reduce stress. Help you train the drivers because they could be trained on one of these digital models. And rather than have to wait for them to get on the road and be trained in the field. And you can optimize your resource management, you may have different kinds of customers, different kind of delivery situations and some drivers may be meta match for certain situations, certain customers. And so these are all the things that the information that the AI program can get, that you may not be able to discern as just a human being overwhelmed by all this data. There's lots of ways to augment and improve, but I don't think any fleet managers can worry about being replaced anytime soon.
Stacey Papp:
So it sounds like you're saying that AI and the human component can live copacetically together, that they can live in harmony. And if you're using your digital twin, which I love that phrase, I think it has a really nice image to it, yet a little bit weird, at the same time. They can live together in order to maybe be a little bit smarter or more seamless. Especially with the incorporation of smart cities, which is a whole other topic in and of itself.
Mike Demler:
Oh, yeah. Yeah.
Stacey Papp:
But it doesn't sound like something that's to be feared, so to speak.
Mike Demler:
I mean, and used properly, it will improve productivity. And use an example with, there's a lot of concerns with say, ChatGPT of being able to spit out text. Well, that could be used as a template to be used more properly, to help people get started on creating some document. It doesn't have to be a replacement and it shouldn't be. So it's a similar kind of vein here, where it's not going to replace what you do. But imagine that you're bringing on a new driver, bringing on a new route or whatever, how much quicker you can get up to speed if you've got that in a digital form to use.
Stacey Papp:
So that bodes the next question. Is AI something fleet managers should embrace? Or maybe still exercise a little bit of caution when they're thinking about it?
Mike Demler:
I think there is certainly a little bit of caution in just in terms of ... I think like any new technology.
Stacey Papp:
Sure.
Mike Demler:
Not to get caught up in the craze of, oh my God, all of a sudden everybody wants to use AI. But how it can be used to improve productivity. And I think it'll be challenging, especially for people that aren't so technology focused of bringing us in. And it is going to be more of a partnership and a collaboration with the various vendors of the hardware and the software, to be able to demonstrate the real value. I wouldn't just jump on converting to an AI based system overnight, but I think the introduction of it gradually and coming up to speed and seeing what value can have is not something that you need to fear, but you should start to look at adopting it.
And I think the collaboration is the key. And I see more of that going on from the tech companies that I follow, that are developing the hardware or the software. They're taking more of a vertical market focus. There's a concept called a foundation model, and it's what, for example, is used with ChatGPT to make it something that will be used, say in banking. Versus just for, say, financial analysis or some other application. And by working together, the companies that are producing some of this software say, well, you give us your data and we'll help create a model specific to your business. And that's the direction that this needs to go in to foster widespread adoption.
Stacey Papp:
So keeping with the theme of evolution, and this is strictly an opinion based question, and I love your candid input. Has AI evolved enough where fleet managers can rely on this tool to assist in managing fleet operations? I think we've touched on this in pieces and parts throughout this interview, when it comes to the human aspect of this. But the evolution of it over time is something we've seen and experienced. You brought it up earlier when we started talking about face unlock or face ID and even when the first iPhone came out. But has it evolved enough where it's to say, okay, cool, we're going to go ahead and rely on this to manage maybe even just a small part of our fleet's operation?
Mike Demler:
Yeah, I think it is mature to solve real problems. I think the challenge is to be a little bit more discriminating rather than just say AI, in general. Because it has so many different facets to it. And even as a technology analyst with my colleagues, we've had lots of discussions. We've even been reluctant to call it AI, we call it machine learning, call it something that seems less foreboding, that less sci-fi based. And using your data to solve a real problem.
What I would recommend to people is to anticipate that there's going to be a transition phase, but that you can look at perhaps other companies or other industries, where this has been successful. And then, begin to adopt it gradually. Not just say, AI in general. It just gets so, I don't know, it just read it so broadly in the popular media. It's difficult to drill down and look at one specific thing, one specific function and solve that problem, rather than try and solve all your problems at once. That's challenging.
But there are the tools. The tools like we started with ADAS systems, with route optimization, with combining your GPS and your telematics data to more efficiently deploy your fleet. Those things are problems that are ready to be solved now. Don't imagine that all AI is going to take over, I'm not going to need drivers, I'm not going to need humans in the office. That's not the way it works. But go solve some specific problems and see the benefit from there.
Stacey Papp:
So last question for you before we go, and those that listen to us regularly know this is my very favorite question to ask, it's the crystal ball question. So if you had a crystal ball, how do you see the evolution of AI impacting fleets in the next 10 years?
Mike Demler:
Yeah, yeah. I love the question too because for example, in terms of autonomous vehicles, self-driving cars, that's one of the things I've been following from the beginning and seeing how overly bullish some of those early predictions are.
But in 10 years, we're talking in the 2030s, you will see a high level of autonomous capability in the vehicles. And that's not necessarily a bad thing. I know for example, that there's a driver shortage, especially for some of the long haul trucking. And even if you still have the human driver in the vehicle, being able to at least partially automate the driving, reduce fatigue, that would be a benefit. And so I know companies in the space that are already there, already deploying that.
There's platooning, for example, that not only has a benefit potentially for reducing accidents with the drivers, but being able to actually reduce fuel consumption. Because of the truck's ability to draft as they do in, say, NASCAR.
Stacey Papp:
Sure.
Mike Demler:
By following much more closely than you would be safe if a human was always solely in control. So there's that in trucking, there's robo-taxis, if it's that kind of a fleet. We're seeing the first of those with Waymo and with GM Cruise. Currently, mostly in fair-weather cities, Phoenix and other areas like, they are deploying in San Francisco and getting experience there. So in 10 years, you'll see some robo-taxi services in lots of cities.
The digital twins we talked about, I think that will be standard. There'll be more automation in the freight yards, more use of robotics, AI based machinery there. It's just going to be used even more. It is beginning to be used now in construction sites, mining. Again, it not having to put people in the mines and having-
Stacey Papp:
Sure.
Mike Demler:
Robots take care. There's just so many opportunities for autonomous vehicles to aid in efficiency and safety. And that's where I see the biggest technology change that we'll see that people predicted 10 years ago would already be here, but in the next 10 years, I think for sure.
Stacey Papp:
Well think where we were 10 years ago when it comes to just things like our phones and computers, and your mind just starts to get blown. I always see those things on the internet about certain generations not knowing what a rotary phone is.
Mike Demler:
I have a good one for you there. My niece had never seen a vinyl record and she was trying to come up with the terms to describe, she called it a radio disc.
Stacey Papp:
A radio disc. Oh my goodness. Well, I feel like it's always an aging factor when you start to hear these stories. Because I grew up in the era of CDs and Walkman, and that was ... I just actually saw this as a answer to a clue on a game show. And the people that were playing, that were being asked had no idea what a Walkman was. And I thought that didn't seem so long ago and that was about 20, 25 years ago.
So the evolution of technology, just like I said earlier, kind of just picks us up and takes it along right side with us. So it will be interesting to see just how mobility ebbs with that evolution of technology. It's knocking at the door if it's not trying to get in already, and I think it's just one of those things, it's just a matter of time.
So I want to thank you so much for lending your insights today on this topic of artificial intelligence. Like we said earlier, it's pretty polarizing, but you made it a little less scary. So thank you for that. Especially when it comes to impacting fleet operations. Again, your knowledge base is fantastic, and thank you for also doing a little bit of prognosticating with us about the future of fleets and AI. It's always nice to get a glimpse into people's brains of where they think technology and how we move about is going to be in the next 10 years, which really isn't that long. So again, Mike, thank you so much for being here today and lending us your brain power. And we certainly appreciate it. We look forward to hearing more from you soon.
Mike Demler:
Oh, thank you Stacey. I really enjoyed the discussion and happy to talk about this topic anytime. It's really a very exciting space. Not something that people should worry about, but they should be excited about the advances that we'll see in the future.
Stacey Papp:
Well worry no more. Until the next time, keep moving the world better.
Thank you for listening to The Mile Marker Podcast. Stay tuned for another episode full of insights and ideas to keep the mobility industry moving forward. In the meantime, follow us on social media and be sure to like, comment and share today's episode.