Vision Podcast #22 – Embedded Vision Summit 2026 w/ Jeff Bier

MVPro Media – The Vision Podcast #22

Guest – Jeff Bier General Chairman, Embedded Vision Summit | Founder, Edge AI and Vision Alliance | President, BDTI

Jeff Bier, General Chairman of the Embedded Vision Summit, Founder of the Edge AI and Vision Alliance, and President of BDTI, breaks down how computer vision is moving from specialist systems into practical, embedded applications, why real-world guidance on trade-offs has been missing, and how emerging technologies like vision language models are shaping what comes next at the Embedded Vision Summit 2026.

---

On this page:

• Podcast player
• Guest information
• Useful links
• Episode chapters
• Episode transcript

---

Listen to the Episode:

---

About our Guest:

---

Useful Links:

• Jeff Bier LinkedIn profile
• Edge AI and Vision Alliance website | LinkedIn page
• Edge AI and Vision Alliance Newsletter
• Embedded Vision Summit 2026 website & registration link – use promo code 26EVSUM-IP-MVPro for 10% off.

---

Episode Chapters

Click onto the chapters to access the relevant sections of the transcript below (coming soon).

1. From Lab to Deployment
Jeff reflects on the moment computer vision became viable beyond research, and how embedded processing unlocked its move into real-world systems.

2. When Deep Learning Changed the Game
From handcrafted algorithms to data-driven models, this chapter explores the shift that redefined how vision systems are built.

3. Beyond the Hype: What Actually Works
Why practical knowledge around trade-offs, limitations, and deployment has been missing, and what engineers really need to succeed.

4. Inside Embedded Vision Summit 2026
Jeff outlines how the event is structured to bridge the gap between technology providers and real-world applications.

5. What Comes Next: Vision Language Models at the Edge
Looking ahead, Jeff discusses how emerging models are pushing the limits of embedded systems and reshaping what’s possible in machine perception.

---

Episode Transcript

Jeff Bier (guest)

What are the actual limitations, the trade-offs, the pitfalls, the proven techniques? None of that was available. And so we decided, let’s try and fill that in so that regular people in industry who are not computer vision PhDs, who have real problems to solve, can get the insights that they need to understand where does it make sense to use this technology and then how to go about doing so.

Josh Eastburn (host)

Welcome to the MVPRO Podcast. For years, computer vision was a technology reserved for the likes of NASA. Expensive, bulky, stuck in elite research labs. But we hit a tipping point. And now we live in a world where your eyeglasses don’t just help you see, they can help you understand what you’re looking at. Today I’m joined by Jeff Bier, founder of the Edge AI and Vision Alliance and the force behind the Embedded Vision Summit. We’re diving deep into the Embedded Vision Summit 2026 agenda to explore the radical shifts happening in the industry right now through embedding, massive intelligence into tiny power-constrained devices. From the jaw-dropping moment when deep learning changed everything to the present frontier of vision language models, Jeff talks about moving from writing manual code to demonstrating tasks to machines. If you’ve ever wondered how a tiny camera module can navigate the physical world without a supercomputer attached to it, you’ll want to stick around for every minute of this conversation.

Josh Eastburn (host)

Let’s dive in. It’s apparent from your track record that you’ve been involved in machine vision, computer vision from pretty early on in your career. What do you feel like is the start of that timeline?

Jeff Bier (guest)

So I’ve been working in embedded processing for over 30 years, but I first got interested in computer vision around 2010.

Josh Eastburn (host)

And was there a kind of a moment that triggered that? What led to that interest?

Jeff Bier (guest)

Yeah, that’s a good question. And it was really the recognition that embedded processors were close to becoming powerful enough that it would be possible to implement useful computer vision on embedded processors. And, you know, once you can embed a technology, that’s where it goes everywhere, right? Like we started with wireless communications has been around for over 100 years, but it’s only become ubiquitous in the last 20 years because now it can be embedded into a computer mouse, a mobile phone, a wristwatch, 1,000 different, 100,000 different things. And so similarly, once you can embed computer vision into things, then its usefulness really is going to explode.

Josh Eastburn (host)

Okay. So you saw that there was a tipping point coming and then this was going to become an option where it hadn’t been before.

Jeff Bier (guest)

Exactly. Yeah.

Josh Eastburn (host)

And is that ultimately what led to founding the Edge AI and Vision Alliance?

Jeff Bier (guest)

Exactly. Yeah. So we realized, oh, hey, we’d seen this kind of transition in other technologies like digital wireless communication, where you have a technology, in this case computer vision, that has been around for a while. Of course, computer vision’s been around for decades, but it was kind of understood to be an expensive, complicated, not super reliable technology. And so most people in most industries would never consider using it. They would think, oh, that’s maybe something for NASA, but not for me. And as a consequence, we saw that it had taken a surprisingly long time, years and years, for earlier technologies to sort of get implemented in lots of products because the reality changed, the technology became practical to implement, but people still had the old idea that, oh, this is not a technology for me. Also, once people realized, oh, maybe I could use this technology, they don’t know how to do it. They’ve never done it before. They need a place to go to learn. How does it work? What is it really capable of? And so we realized like, okay, maybe we can speed things along here if we can create heightened awareness that computer vision isn’t just for NASA anymore.

Jeff Bier (guest)

It’s technology that can go into everyday products and help people learn how to use it successfully. And so that, that was kind of the idea that spurred us to start the Edge AI and Vision Alliance. I think it was 2011. So about 15 years ago.

Josh Eastburn (host)

It’s called an alliance because it is a partnership of companies in the space. Is that right?

Jeff Bier (guest)

That’s right. Yeah. We’ve worked in the electronics industry for decades and we realized this was like a much bigger lift than one company, large or small, was going to be able to do. So we started talking to customers of ours. I run an engineering consulting company and just other companies that we know in the industry. And we were pleasantly surprised that almost every single one of them said, oh yeah, we think there’s a pretty interesting opportunity here. And yeah, we’d like to work with you on this organization. So we quickly gathered, I think maybe 16 initial kind of founding member companies to work with us. And then over time that expanded and it’s around 100 member companies now. The member companies are mostly the providers of the building block technologies, like the processor boxes and boards and the cameras and software tools and so on. And then we work with those companies to reach out to the much larger community of people who are building systems and solutions using this technology to help them understand what’s possible, how it works, and help them connect to the suppliers.

Josh Eastburn (host)

So is that really how the Embedded Vision Summit came about? Like, let’s get these groups together?

Jeff Bier (guest)

Yeah, exactly. And also like, it’s very interesting because we were trying to learn about the technology ourselves 15 years ago about practical computer vision, and we saw that there was very, very little available in terms of practical education. There’s a lot of academic research conferences and journals, but those are not all that useful for like a working engineer in industry. They’re really by specialists for specialists. And then over time, as the technology, the commercial potential has become clear, there’s an awful lot of just like marketing hype about ‘Look at our amazing thing. Look at all the amazing things we’re going to be able to do.’ But the in-between, there’s this kind of vast desert in terms of like, ‘But okay, how does it really work? What are the actual limitations, the trade-offs, the pitfalls, the proven techniques?’ There was none of that was available. And so we decided like, ‘Okay, let’s try and fill that in so that regular people in industry who are not computer vision PhDs, who have real problems to solve, can get the insights they need to understand where does it make sense to use this technology and then how to go about doing so.

Josh Eastburn (host)

How long has the summit been running for?

Jeff Bier (guest)

I think this is the 16th year coming up, if I’m not mistaken.

Josh Eastburn (host)

And how about some of the highlights over the years? You heard it here first moments that you feel proud of?

Jeff Bier (guest)

Yeah, I mean, there are really many. The conference program is something we craft very carefully. We select most of the speakers, our invited speakers, of the talks that come in as proposals that we accept. We collaborate with the speakers to make sure it’s the right topic at the right level. So there really have been hundreds of outstanding presentations over the years, but two that particularly stick in my memory. One is from 2014. We were fortunate to have Yann LeCun give a keynote talk at the Embedded Vision Summit. And most people probably have heard that name. LeCun is considered really one of the godfathers of deep neural networks. And in 2014, deep neural networks were just starting to come out of academic research and be applied in industry. And it wasn’t a widely known technique, was definitely not a widely understood technique. And LeCun gave this talk and people’s jaws were hanging open. And I remember walking out of the talk thinking, wow, If what he just claimed is true, this changes everything we know about computer vision. And of course, everything he claimed did turn out to be true and more, and it has completely changed how we do computer vision.

Jeff Bier (guest)

So that one really stands out in my memory. Another one was a few years later, I think it was 2018, one of my personal heroes, Dean Kamen, who who’s best known as the inventor of the Segway, but has really accomplished an incredible amount in his career, mostly inventing medical devices, but also starting the first robotics program, which millions of school kids have participated in, giving them, you know, really positive experiences working with technology. He gave another keynote talk at the Inbetavision Summit, and he made a surprising to me observation. He said, well, I think of computer vision as kind of just a very flexible software-defined sensor. I design all kinds of systems with all kinds of sensors. And the thing I like about computer vision is I can use it to sense almost anything. And so Dean said, I predict 10 years or some point in the not too distant future, the same way that something like a limit switch is this very basic building block that, you know, design engineers will use in all kinds of systems without thinking twice. I think that embedded computer vision will become that kind of thing that gets embedded into all different kinds of systems and applications.

Jeff Bier (guest)

And people don’t think twice about it. It’s like, oh, of course, I’m just going to pop in a smart camera module here into this system to give me the feedback I need to get the result I want. And it sounded, I wouldn’t say outrageous, but a little bit outlandish at the time. And I remember thinking to myself, you know, normally I would be skeptical of a claim like this, but This man has accomplished so much in his career and has proven to really have vision about, well, here’s where we need to go, here’s how we’re going to get there. I have to take it seriously. And I think, yeah, like LeCun’s keynote a few years earlier, Kamen’s keynote, his prediction has proven to be true. And in fact, his own company has designed an intravenous infusion pump that uses deep inside this pump that’s used in hospitals is a little camera. And the camera is used to monitor the drops of the IV fluid as an input to feedback control. So he’s gone ahead and, you know, implemented the exact thing and kind of predicted almost 10 years ago. So those are two talks that really stick in my mind.

Josh Eastburn (host)

I’m really curious about the LeCun talk in 2014. What were some of the things that he was proposing were possible way back then?

Jeff Bier (guest)

Free deep neural networks. The way we solved computer vision problems was we reasoned about them with kind of human thought processes. And we’d say, okay, well, let’s say, you know, there are strawberries going down a packing conveyor belt at a packing plant, and we want to count the strawberries and check their color and their size and check them for damage. We want to detect foreign objects. So, you know, a bunch of engineers would sit down and kind of reason about Okay, how are we going to do this? Well, maybe there’s enough contrast between the strawberries and the belt under them that if we can find the edges, we can identify like the silhouette of the strawberry. And then maybe from that we can figure out the size. And they’d put these ideas into code. And it was a very labor-intensive process, required a huge amount of expertise. And typically the results weren’t that great. And if anything changed a little bit, like the camera angle or the lighting or the color of the belt, often these algorithms would break. And so engineers would have to go back to the drawing board and maybe come up with a new algorithm or tweak the algorithm.

Jeff Bier (guest)

And because it took so much effort and expertise to come up with an algorithm for each specific application and use case, it just wasn’t practical to solve most of the potential applications. One great example is a lot of people know Mobileye. Mobileye was a real pioneer in computer vision for driver safety. And they put thousands of man-years into developing their algorithms for their early generation products. As you think about, you know, how complex it is for a camera looking through a windshield of a car to understand the scene and identify, is that a pedestrian? Is that a mailbox? Where are the lane markings? How close or far are we from that vehicle? Thousands of person-years. Most companies, of course, could never make that kind of investment. And what LeCun showed is, well, We don’t have to do that anymore because we have this very general learning machine, an algorithm called a convolutional neural network. And if we just show it enough examples, it can learn, we can train it, it can learn to distinguish between, let’s say, a strawberry and a rock or a damaged strawberry and an undamaged one. And that was really the game-changing aspect of convolutional neural networks.

Jeff Bier (guest)

We need a lot of training data now, a lot of examples, but it’s a profound change in how algorithms were developed. The analogy I like to use is it’s like the difference between describing a process to someone in words and demonstrating it to them. Imagine like if you were teaching a child to tie their shoe, imagine trying to do that just through words. Virtually impossible, right? And so nobody would ever even consider doing that. No, you demonstrate. And in fact, you can go a step further. You’ll hold the child’s hand and kind of lead them through. And by repeated examples, the child learns how to do the process. That’s how convolutional neural networks work. Whereas prior to convolutional neural networks, it was more like writing out the procedure step by step and then deploying it and realizing, oh, we forgot to specify, yeah, if the color of the belt changes or the speed of the belt changes or the lighting changes, we have to, you know, adjust our approach in this way or that way. So really a very profound change from each application requires a unique set of algorithms devised by human ingenuity for that specific problem.

Jeff Bier (guest)

We have some very powerful, very general purpose algorithms that we can train with new data to address new applications.

Josh Eastburn (host)

Gosh, that’s fascinating. I think we could probably spend a whole episode just talking about that. Let’s jump forward from there 12 years into the future to our present. What are some of the highlights on the agenda this year?

Jeff Bier (guest)

So the Embedded Vision Summit is really a unique event that’s designed for industry professionals who are solving real-world problems using computer vision and other forms of machine perception. So maybe using LiDAR plus AI as opposed to cameras plus AI, the whole range of machine perception, but especially computer vision, which is, I think, the most powerful, most popular kind of machine perception technology that’s in use in industry today. And we’re trying to do a few different things with the conference program. One is just provide the basic know-how and skills that people need in order to select and utilize this kind of technology effectively. So that’s kind of a core piece of the program. People can come if they’re new to the field, kind of start with very basic introductory presentations and get the practical knowledge that’s hard to get from on the one hand, you know, university courses and academic research, or on the other hand, a lot of these sort of marketing cheerleading kinds of stuff that are out there. The kinds of things you need to hear, you need to either learn from doing yourself, or you need to learn from other people who’ve learned from doing.

Jeff Bier (guest)

So that’s kind of one pillar of the program. Another pillar is, well, what are the commercially available building block technologies, the embedded computers, the camera modules, the software tools, and so on that I can I’m going to use to build the solution implementing computer vision in my application. So that’s another piece of the program is we have many of the leading suppliers giving presentations on kind of technical explanations of here’s what we’re offering, here’s how it works, here’s what problems it solves. And there’s also an exhibit floor where many of those same companies are giving demos and, you know, talking one-on-one with attendees. Another pillar is more advanced techniques, the kind of emerging, helping people get their arms around the emerging techniques, what’s just becoming feasible now in practical applications. And that’s where it’s often hardest to really get a handle on, well, as this stuff actually work, does it really work? Um, you know, and in 2014, 2015, that those were the kinds of questions people had about neural networks. Do these things really work? There’s been a lot of AI fads in the past that have come and gone without much impact. So stuff really working, a lot of people were skeptical and a lot of people remain skeptical for a while.

Jeff Bier (guest)

Took people really learning about the technology, how it works, and seeing what it can actually do in real applications to become convinced like, oh yeah, this time it’s for real. This AI stuff is actually a game changer. And each year there’s a new frontier. Right now I would say probably the most important frontier is what we call vision language models. So everybody knows about large language models like ChatGPT. A vision language model combines both images and language. So it’s a model that can take as inputs images and language. And these are very powerful. They’re also very large, complex models, but they’re very powerful models that are opening up new capabilities and they’re just starting to be deployed in applications and in kind of edge embedded systems, they really push the limits of the hardware because these models are quite large in most cases. Instead of millions or tens of millions or even hundreds of millions of parameters, which is how we usually talk about convolutional neural network sizes, these models are typically billions or sometimes even tens of billions of parameters. Okay. So learning about why would I want to use a vision language model, what would it give me, and Is it actually feasible to use it other than in a big data center somewhere?

Jeff Bier (guest)

And what would that take? What kind of hardware would I need? What kinds of software tools would I need? How would I make that work? That’s one of the core focus areas for this year’s Embedded Vision Summit, because it turns out that much as convolutional neural networks were 10 years ago, today vision language models are just really emerging as a game-changing technology. And people can see enough of that to realize, okay, this looks like it’s going to be important. I need to figure out does it apply to my application? And if so, how do I use it?

Josh Eastburn (host)

And I saw that you’re offering a two half-day training on VLM specifically, right?

Jeff Bier (guest)

Yeah, yeah. So, you know, most of the conference program is, you know, 30 to 60-minute presentations by experienced practitioners. But for some of this stuff to really grasp it, you need to get hands-on. You need to actually train some models and write some code. And so for people who want to get that next level of proficiency, yeah, we have two half-day expert-led training classes on vision language models that the Edge AI and Vision Alliance is presenting with our partners OpenCV. And that’s a great option for people who want to do more than just get a passing familiarity with vision language models, but want to really get started on the path of using them in applications. So there’s a morning session that’s an introductory session for people who are just getting started with vision language models, and then an afternoon session that’s bringing in some more advanced techniques for people who maybe have already started experimenting with vision language models or who took the morning session and now want to get to the next level.

Josh Eastburn (host)

And what do you hope is a realistic outcome that attendees could have in that time period?

Jeff Bier (guest)

Yeah, I think what we’re pretty confident they’ll get based on the results from last year where we ran a similar training class is they’ll get a realistic understanding of, well, first of all, what are these models? Vision language models, how do they work? How do they differ from earlier generation convolutional neural networks? What are they good for? And, you know, where might you want to use one and what type of application? And most importantly, then how do you use them? How do you actually implement them, customize them, integrate them into an application?

Josh Eastburn (host)

If you want to validate the idea, if you think you see some potential for this technology in your own business, this would be a way to get started and stress test that idea.

Jeff Bier (guest)

Exactly. And you know, this is pretty technical hands-on work. So I would say much of the Embedded Vision Summit main conference program is accessible for people who don’t have a lot of relevant technical background. And in fact, we have a fundamentals track. So full 2 days of introductory presentations as part of the main program. This vision language model training, this kind of assumes, you know, you’re comfortable with the basic concepts of deep neural networks and training them and integrating them into applications. And now you want to go to that next level.

Josh Eastburn (host)

Very cool.

Jeff Bier (guest)

Vision language models.

Josh Eastburn (host)

I also noticed there the enabling technologies track, and you mentioned emerging technologies. I’m not sure if that’s the same, if we’re talking about the same thing there.

Jeff Bier (guest)

Exactly.

Josh Eastburn (host)

It is? Okay, great. So I hear from many of the guests on this show that there’s this feeling that there’s kind of a renaissance happening for computer vision right now because of some of the technologies we’ve been talking about. And I’m wondering, what do you expect will be top of mind for presenters in that particular track?

Jeff Bier (guest)

Yeah, no, I think that’s a very apt description. There is renaissance happening in this field. The investment collectively across the industry in new building block technologies for computer vision, I don’t have a number, but it must be thousands of times what it was 10 or 20 years ago when this was a very niche technology. We have these giant companies like Google and Apple and Amazon putting billions of dollars into, for example, new, better, more efficient, more accurate computer vision models. And then in many cases, putting them out for the public for anyone to use, sometimes as open source, sometimes as a paid service. So yeah, there’s a tremendous renaissance happening in the field. It’s a very exciting time. And in this enabling technologies track, that’s part of the Embedded Vision Summit conference program, you’ll be able to hear directly from many of the companies that are on the forefront of developing these building block technologies, new kinds of processors, new kinds of sensors and cameras, new kinds of software development tools about what they’re doing, not what they’re thinking about doing or might have ready, you know, in a year or two, but what they have done and have available right now that people can incorporate into systems and applications.

Jeff Bier (guest)

And it’s very wide ranging. It’s everything from new let’s say lower cost ways to do depth sensing, to new processors that are optimized for these vision language models to run those very efficiently, to tools that are designed to help people who are deploying large numbers of systems. Like let’s say you’ve got a manufacturing inspection system and you’re deploying 1,000 copies of it around multiple factories. When you get up to that kind of scale, it introduces new kinds of challenges. How do you make sure all those systems are functioning correctly? What happens when you find problem with the model and there’s an error? How does that information get brought back and used to update the model? And then the updated model get propagated out to all 1,000 units in the field. So this kind of fleet management is an increasingly important facet of computer vision systems as they become more widely deployed at scale. And that’s another aspect that companies will be talking about in some of these presentations.

Josh Eastburn (host)

Do you reserve the privilege of doing one of the keynote speeches for yourself every year?

Jeff Bier (guest)

Not a keynote, but I make welcome remarks. And in my welcome remarks, you know, it’s covering the whole gamut from where to find the restrooms and get your lunch to, yeah, what do I think is most interesting? In our field today and how can you learn about those things at the conference. I think that I’m very privileged to work in this organization, the Edge AI and Vision Alliance, and to be in charge of this conference. So many very talented, very smart, hardworking people are kind of coming together, bringing their kind of best ideas and their best work. And I get to kind of sit over here on the side and look at it all and say, huh, yeah, Yeah, that’s pretty interesting. I see a pattern. You know, these three things fit together. Or wait, just yesterday somebody came with an idea. I thought, oh yeah, I hadn’t thought of that before. That’s pretty interesting. And so I try and share some of those insights in my introductory remarks, but I’m just the warm-up act for the keynote speakers who are the real heavy hitters. You know, these are some of the most accomplished and influential people in the field.

Jeff Bier (guest)

This year, our keynote speakers, there are two. One is Professor Eric Xing, who is a leading light in an area called world models, which is kind of the next thing maybe after vision language models, models that not only understand images and language, but also understand something about how the physical world works, like the laws of physics. You know, what’s going to happen if these two objects collide, for example. And then our second keynote speaker, Vikas Chandra from Meta, You know, Meta has these Ray-Ban Meta smart glasses, the AR augmented reality glasses, and to get those to work, it’s quite a remarkable product, right? You can wear these glasses. You can go, for example, if you’re a tourist, you know, you can look at a building and you can say to the glasses, like, what’s this building? What’s significant about it? What happened here? Or here on my counter are the ingredients I have. What can I cook with this? And to make that work, these guys have been really pushing the envelope in terms of how do you get these very sophisticated computer vision AI algorithms to run on tiny little processors that can fit in a pair of glasses or fit in the phone in your pocket.

Jeff Bier (guest)

Because if you have to send all the data to the cloud and wait for a big model to crunch on it, send it back, it’s not going to be a satisfying user experience. It’s going to take too long. And also it’s not going to work when you don’t have good connectivity. To the cloud. So Vikas is going to share some of the latest work from his research team at Meta on how do we fit these very sophisticated computer vision models, vision language models, and other types of models into very cost and power and size constrained applications. So they’re really the main acts. I’m just kind of a warmup.

Josh Eastburn (host)

Oh, that’s fascinating. I think we all remember Google Glass 10 years ago, maybe that was, or maybe even a little bit more. So yeah, I’ve been skeptical about this new generation of smart glass. So how fortunate that your attendees get to deep dive into that. Very fun. If someone were listening to our conversation right now, seeing what’s changing in the market, maybe even feeling a little overwhelmed by it, or seeing that there’s an opportunity that they feel like might be slipping out of their grasp and they want to get started, what’s kind of the first lowest barrier that they could try to leap over to start getting educated on this stuff?

Jeff Bier (guest)

Yeah, if you want to learn about kind of embedded edge computer vision, applications, technologies, techniques, come to the Edge AI and Vision Alliance website. It’s edge-ai-vision.com. And you will find just an absolute treasure trove of free information. It’s all free presentations, articles, interviews, demonstrations. It’s all real-world stuff. And it’ll help you really get a handle on, okay, what are people doing with this technology today? And What are the building blocks that I can buy and where are the challenges? What are people having difficulty with? Hundreds and hundreds of hours of presentations, interviews, demo videos, and thousands of articles, all free. Lots of really insightful stuff based on real-world experience.

Josh Eastburn (host)

It sounds like you’ve built something really special for the community. For people who want to go to the conference, it’s happening in May, so I’m hoping that means they still got some time to get a pass if they don’t have one. What’s the process for doing that?

Jeff Bier (guest)

So the Embedded Vision Summit takes place May 11th through 13th in Silicon Valley at the Santa Clara Convention Center. And the conference has its own website. It’s embeddedvisionsummit.com, no punctuation. Go there and there’s still plenty of time to register. There’s discounted hotel rooms available for people who are traveling. And speaking of discounts, there’s a promo code you can use to get 10% off the registration. It’s 26eVSUM-IP-MVPro and use that promo code to get 10% off your registration. And there’s various options, including the VLM training classes we mentioned, one-day passes, two-day passes, et cetera.

Josh Eastburn (host)

Excellent. We’ll definitely put that in the show notes so people don’t have to scribble that down while they’re listening. Anything you feel like we left out you want to touch on?

Jeff Bier (guest)

I think going back to your previous question about how can people learn? And keep abreast of things. Another great free resource is our weekly newsletter. And if you come to the main Edge AI and Vision Alliance website, edge-ai-vision.com, right there on the homepage, you can sign up for this free weekly newsletter. And you know, I would say if you read it really from start to finish, it probably takes 10 minutes a week. And of course then there are pointers to go dive deeply into various topics, but just to get the overview and keep up with what’s happening, in the industry, it’s a great way to do so. And like I said, we’re really aiming for this zone of what is practical today, not so interested in what might be practical in 10 years. You know, what’s emerging as academic research today might be practical in 10 years and not so much interested in kind of the hype and the cheerleading, but really trying to understand what works today. These people are building an agricultural weeding robot and where the hard parts about incorporating computer vision, not just in a production line where everything is bolted into place, but on a moving robot in a farm field.

Jeff Bier (guest)

How’s that different? What are the challenges and how do they solve those challenges? For that kind of content, the weekly newsletter and the website are really fantastic resources.

Josh Eastburn (host)

Excellent. Well, thank you again, Jeff.

Jeff Bier (guest)

Yeah, thanks for having me. It’s been fun.

Josh Eastburn (host)

Whether you are managing a fleet of industrial inspection systems or trying to fit a billion-parameter model into a wearable device, The era of marketing hype-dominated conversation is giving way to the era of practical conversation about embedded applications for computer vision, thanks to folks like Jeff. If you’re ready to get involved in the present renaissance, now’s the time to get plugged in. Join the industry’s best at the Embedded Vision Summit in Silicon Valley, May 11th to 13th at the Santa Clara Convention Center. And grab 10% off your pass at embeddedvisionsummit.com with the promo code 26EVSUM-IP-MVPro. Between now and then, head over to edgeaivision.com for hundreds of hours of free technical talks, demos, and their must-read weekly newsletter to stay ahead of what’s actually working in edge AI and computer vision. If you’d like more interviews like this one, head over to mvpromedia.com and connect with MVPro Media on LinkedIn. If you are a vision professional or integrator with real-world lessons, not just hype, I’d love to talk to you.

Josh Eastburn (host)

Email me at josh.eastburn@mvpromedia.com.

Josh Eastburn (host)

This episode was produced by Rachelle Kondo and Big Robo. For MVProMedia, I’m Josh Eastburn. Be well.