Resources / Blog

NVIDIA’s Adam Scraba on AI That’s “Sprinkled Into Our Lives”

Join Infinia ML’s Ongoing Conversation about AI

Episode Highlights from Machine Meets World

This week’s guest is Adam Scraba, Director of Product Marketing at NVIDIA. Highlights from our conversation include:

“Now we have tools that effectively allow nearly any company in the world to be an AI expert, but they don’t need the data science expertise.”

“I think we’re on a path where AI is actually going to become kind of like a utility. It’s going to be accessible to pretty much everyone.”

“AI won’t be necessarily a technology that you really need to spend a lot of time thinking about because it’s just going to be built into everything.”

Photo by abigail morris on Unsplash

Audio + Transcript

Adam Scraba:
I think we’re on a path where AI is actually going to become kind of like a utility. It’s going to be accessible to pretty much everyone.

James Kotecki:
This is Machine Meets World, Infinia ML’s ongoing conversation about artificial intelligence. I am James Kotecki and my guest today is Adam Scraba, Director of Product Marketing at NVIDIA. Welcome to the show.

Adam Scraba:
Thanks, James, for having me.

James Kotecki:
So Adam, NVIDIA, which I should say is not Infinia, these are two totally separate companies that we work for here. NVIDIA makes computer hardware that makes machine learning and AI possible. So when we say machine learning, you guys are the machine. Is that fair enough?

Adam Scraba:
It is fair enough, yeah. We do a lot of machines. Certainly our business is very much in the hardware space, but we actually are much more of a software company at this point. We invest a huge amount in software development tools and so we really build a stack of computing and that incorporates hardware and software as well. You know, in the old days, you would have to understand intimately how to create, say, edge detection, edge detectors to identify, say a car in an image. And it was an expert writing software. It was an expertise being codified into some software. Right now with AI, what’s incredible is AI is actually software writing software.

Adam Scraba:
So if you look at a piece of AI that was trained on millions of images of cat faces or cars, AI actually taught itself, which are those important aspects, which are those important edges. And it turns out if you actually open up the black box and you look inside of it, it actually looks like an expert has written it. It actually looks like an old piece of software that maybe I might’ve written 15 years ago, but it was actually software that wrote it. And that’s the magic. The magic of AI is just so incredible. And it turns out that image and video analysis is actually one of the more mature modalities of AI. And it’s actually one of the more concrete, it’s tangible, you can understand it. So that’s made it easy for us to communicate this to our ecosystem.

James Kotecki:
So you’re working in product marketing. What are the products that you are marketing right now? What are you actually working on to apply this in the real world?

Adam Scraba:
Yeah, so our team works on a really interesting area of applying AI. It turns out that video and video cameras are incredibly powerful IOT device and arguably perhaps the most widely deployed IOT device the world has ever seen. There’s probably about a billion video cameras deployed in the world right now, and it’s growing at an incredible rate. These IOT devices are actually incredible sensors. We can answer a lot of questions about our physical space or real estate with cameras. And so our team really works on effectively making spaces smarter. So there are a broad array of industries, as you can imagine, whether it’s airports or factory automation, retail, where there’s really high value assets — that could be people, it could be things — that we want to increase the operational efficiency of or increase safety. And so we use cameras to do it, and it’s really becoming an incredible business opportunity. It solves really important problems. And with AI, for the first time we can solve previously unsolvable problems with video cameras. And like I said, increase operational efficiency and safety to a great degree.

James Kotecki:
And what’s the state of the intuition of the clients that you work with for how to apply and use this stuff? The metaphor is often used that AI is the new electricity. When they were first rolling out electricity, people didn’t even really have a good idea of how to apply it, at least in the ways that it eventually became applied in useful ways. Are business people becoming just more attuned to like, “Oh, that’s a problem that I have and I can solve it with AI,” or even now being able to define problems in new ways maybe that they didn’t even define as a problem before because they know that AI is able of solving that problem.

Adam Scraba:
Yeah, we’re in such an interesting time right now. A few years ago, a lot of people were very focused on AI and early stage AI development. What we found actually was we were finding, in particular in our industry, which was in the image analysis and video analytics space, we found that there was a lot of companies or partners that were computer vision or AI experts first and industry specialists after. Maybe they were experts at computer vision, and then they realized they could apply it to retail. And so they would go and build solutions for retail. What’s really incredible right now, and it’s interesting that you say, you use the utility analogy is that the tools have evolved so much right now that with things like transfer learning in the AI space where you can start with a pre-trained model and apply very limited amount of retraining to customize a model, you now have potentially retailers that are — or retail experts, industry experts that can come at a problem from an industry specialist standpoint, as opposed to needing to be an AI expert.

Adam Scraba:
And so what that allows us, because now we have tools that effectively allow nearly any company in the world to be an AI expert, but they don’t need the data science expertise. And they don’t need to invest years or have an army of data scientists to make really magical applications happen. So I think we’re on a path where AI is actually going to become kind of like a utility. It’s going to be accessible to pretty much everyone.

James Kotecki:
What’s a use case that you’re excited about that’s either about to come out or you can see it on the near horizon?

Adam Scraba:
I think that one area that I’m really excited about is, we talk a little bit about, we talk about AI being an automation tool, we talk a little bit about robotics. I think what’s really exciting right now is we’re actually starting to see that some of our physical spaces, whether it’s our office buildings, it could be our airports, it could be our transit stations. With AI and actually video as a really important unstructured sensor data, we’re actually turning them into robots. So if you think about robotics, it’s perception, reasoning and action-taking. Perception, that first layer, is really going to be so suited to AI. It’s such a perfect space just to take unstructured… There’s way too many people on level three of a building, or occupancy’s too high, or an aircraft has just come in to the gate and we need to refill it much [more] quickly or we have to increase operational efficiency on the tarmac of an airport. These are all things that we’re starting to see AI take shape. And it’s very, very exciting.

James Kotecki:
What is something about AI that you think business leaders just still don’t really get?

Adam Scraba:
I think that we have . . . as an industry, we could do better at packaging up some of these solutions. If I maybe read into your question a little bit deeper, and I think about maybe in some cases, why isn’t AI being more adopted just everywhere? If there’s critical mass of capabilities and industry specialists building such incredible tools, which there is, why isn’t it just everywhere? And I think we can do a lot better at packaging these solutions up. I think that over time, AI won’t be necessarily a technology that you really need to spend a lot of time thinking about because it’s just going to be built into everything. And that building into everything, that packaging, I think is something that we’re going to see a lot more of, and it’s going to be a lot more seamless. And that’s what is going to increase adoption of this technology to a huge degree.

James Kotecki:
I often wonder if this glorious AI future that we’re all driving towards is going to look less like something fundamentally different and just look more like a world where everything just works better and everything just works more often as you would expect it to work.

Adam Scraba:
I agree. And even think about where we are today — we talk to our speakers, we talk to our cars. There’s an intuition — Netflix knows what we’d like to watch. It’s actually very accurate at sharing with us what we’d like. AI is being sprinkled around us all over the place. And I feel like we’re already getting that sense and we’re seeing what that future already looks like. And I totally agree. I think it’s just going to be sprinkled into our lives and we might not necessarily need to think much about it.

James Kotecki:
Well, Adam Scraba, Director of Product Marketing at NVIDIA. Thank you so much for joining us today here on Machine Meets World.

Adam Scraba:
Thank you so much for having me.

James Kotecki:
And thank you so much for watching and/or listening. You can always email the show. It’s mmw@infiniml.com. Please like this, share this, give the algorithms what they so desperately want. I am James Kotecki, and that is what happens when Machines Meets World.