Welcome to HumAIn. My name is David Yakobovitch and I will be your host throughout this series. Together we will explore AI through fireside conversations with industry experts. From business executives and AI researchers, to leaders who advance AI for all. HumAIn is the channel to release new AI products, to learn about industry trends and to bridge the gap between humans and machines in the fourth Industrial Revolution. If you like this episode, remember to subscribe and leave a review.
Welcome back listeners to the HumAIn podcast. Today, our guest speaker is Jakub Krcmar, who is the founder of Veracity Protocol. I had the opportunity to meet Jakub digitally within this past month through a few mutual connections involved with startups and business between New York City and Taiwan.
In fact, the ecosystem between New York and Taiwan is thriving. There are a lot of events happening with consulates as well as ecosystems building cross border innovation, both inward investment and foreign direct investment. So super excited to have you on the podcast. Welcome Jakub.
Thank you. Good to be here, David.
I’m really excited about the product you’re working on. Myself, being someone in the data science and AI space, I’m always looking at new research and new technology. And as we’ve discussed offline, the Veracity Protocol, which you’re working on, is this new security standard for physical objects using AI and other technologies. So let us know for the listeners, what are you up to?
Definitely would love to. So at Veracity, we developed an algorithm based on computer vision and artificial intelligence, which basically enables any camera, be it a smartphone camera or industrial camera to be able to analyze any object’s physical structure, and create something that we call tamper-proof physical code.
This basically really works on the basic idea that any physical object itself is unique. It has its own unique material structure, same as a human fingerprint, which is immutable. You cannot really replicate it down to the level of microns. And also, mass produced items release a certain signature of how they were manufactured. The material used, the sewing, the printing technique, all these factors, again, create a unique signature, which is something we can pick up on.
And we are doing this and using this physical code to basically solve the issues of authenticity of manipulation to help detect any anomalies or tampering and to really protect human lives and brands and national security and allow for a fully automated, digitized, tokenized future world and industry 4.0. So this would be, in a nutshell, what we are focusing on.
So this is so cool because when we think about identity, whether it’s in real life identity or digital identity, it’s very challenging to track, to verify, to authenticate what is real and what is fake.
I just recently watched an exposé on the Wall Street Journal, one of their tech briefings was talking about a company in Japan that works with knives. And basically, what they said is, if you want one of these really sharp knives, you’re going to get your identity stamped into the knife, as a QR code, that’s laser etched.
And I found that so fascinating because otherwise, how do you know who it belongs to, or is it even authentic from the manufacturer? So it sounds like this is a problem. And you’re talking about use cases around identity and authenticity. Why is this a problem we’re experiencing today?
That’s a good question. We see the upcoming era of blockchain and a lot of companies using blockchain as something to track provenance, for example. By my point of view, that’s another duplication of how we are using physical certificates of authenticity. Even blockchain itself is a very cool technology of decentralized databases.
I cannot really secure what is essential, which is the way how you connect the physical object itself to any digital entry, be it a database or blockchain. How do you create this secure connection? Most of the time you are relying on a unique ID, RFID code, QR code, NFC chips. Things like that.
Even things like invisible inks, DNA markers, diamond dust, other things like that, and all these are security elements which you need to embed, which can be tampered with. Change can be de-attached and you require, most of the time, some hardware to work with them or special skills, but really, the idea here is still working with the physical object itself, the way it is. How it was produced leaves enough fingerprints and enough ways of connecting the object itself to the digital ledger through its unique material structure. You cannot really replicate a sheet of paper down to each fiber or the same way.
So it’s basically a solution to this issue, which then allows you to truly tokenize physical objects, to work with them. To assure they are setting a truth from all the data connected to those objects as they go down the supply chain, to allow for real transparency in the supply chain for reliable track and trace.
When you buy something and once you verify it originally, it executes the smart contract. So this is a key technology to not only some issues of today, like the counterfeiting and transparency, but also of tomorrow to enable the optimized industry.
It’s so fascinating that you mentioned all these technologies have been used today. RFID, QR codes, NFC chips, invisible inks, diamond dust. It gets me thinking about the show on HBO that’s been all the rage in the past year, His Dark Materials. They talk about all objects, whether we’re thinking of the real world or an HBO, having a track or having some system or dust or material.
So it seems that there is this trace, there’s this thing that can be tracked and you’re working on that, but you didn’t just get started on it today while you’re working in Taiwan. But in fact, you went through an accelerator in New York City, and that’s the Techstars Accelerator, which, if we’re in New York, we know Techstars is one of the leading accelerators to take startups from idea to execution. Want to hear about how your experience was in Techstars?
Definitely. That was one of the best and toughest experiences of my life. To be honest, it was three months. A lot of sweat and blood and really the best 6% of equity we ever invested. And since our big round originally, we were starting in the space of art and collectibles.
We were building a company to exchange artworks before. And one of the issues we came across was counterfeiting. And we saw how all these technologies are not really solving the issues. And we were thinking there has to be a simpler way. And it was the beginning of the process three years ago, when we started developing the technology, the algorithm, and creating the data sets and getting everything moving.
Techstars was a milestone, really, which through their process, pushed us to literally get our stuff together and really be precise about how we want to use this technology, what market we want to target and get everything together. So that’s been really amazing. I cannot recommend it enough.
And as you’re taking all that success, you’ve pivoted and grown from Techstars in New York now as split or focused between New York and Taiwan. That presents a lot of opportunities and interesting notes.
I was mentioning at the start of our episode that New York has a strong ecosystem in Taiwan. In fact, there’s a venture called Anchor Taiwan that builds partnerships with supply chain and hardware and startups all across the border. And so now, you’re splitting your time between New York and Taiwan. What was the reason to scale the venture across the globe?
So we originally started from Czech Republic, from the middle of Europe, going to New York, where so far there this day, we have most clients, most engagements. And it really has been the best place for us to be, another thing I can’t recommend enough time on.
Honestly, the reason I’m here is also that my wife is Taiwanese. That’s first of all. Second of all, I’m combining this with also the business opportunities, because one of the verticals we are focusing on is semiconductors.
And, obviously, in that area, Taiwan is the place where you want to be. Because I don’t know if in remote nosy, but Taiwan is the place where most of the semiconductors are produced. And that’s why it’s also become strategically important globally.
That’s excellent. And one of the companies that, if listeners did not know semiconductors were produced in Taiwan, if you say the company, Foxconn, traditionally thoughts are China, but it couldn’t be farther from the truth. Foxconn has a huge, or perhaps their biggest presence in Taiwan.
That’s correct. And also, TSMC would be another major, big one and there are several others. Taiwan is a leader in the industry.
And it sounds that it has given you a spark to grow and scale, evolving from, as you mentioned, focusing on the art and collectibles industry now into security for physical objects, primarily in semiconductors and manufacturing.
This is such an interesting use case, at least in the States. We think often about a lot of intellectual property and tech transfer. And a lot of the conversations have been around what’s been in the news the last few months and that’s focused on one of the top four, if not even the number one global cell phone manufacturer in the world, Huawei. Huawei out of China.
Huawei is a great company. They’re building 5G signal networks in the European Union and Britain. They’re building smartphones you use throughout South America and Africa. They’re even creating laptops and technology to bridge the gap in emerging frontiers and developing markets, but it seems whether truth or rumor, there has been a lot in the news about Huawei and these microchips, and let’s call them, the veracity or the traceability of how secure some of these networks are.
And I’m bringing this story to light because you’ve mentioned about the focus on semiconductors. I want to know what has inspired that shift or transformation into a semiconductor focus? Has it, any of it, been from the Huawei story? Has it been from the growth of IOT and 5G? Everything is a smart device now, as we’re gaining acceleration in 2020, what’s some of that inspiration?
So there’s several angles to this. One of them is, we obviously are not focusing only on semiconductors, there’s two other areas, collectibles and sports memorabilia is a big one, luxury goods and also security documents like IDs and passports. The overarching team is really like objects or items of high value or high security threat. For example, if you take a several months warrant, it’s very easy in the supply chain to introduce fake components to it, which creates a lot of issues regarding their life expectancy, but also, it’s possible to introduce hidden chips which can provide route access. For example, to the server, the motherboard it’s installed with.
So there’s security issues of software angle and hardware angle. And software angles you can always fix. But the hardware angle is unfixable until you actually replace and fix that thing physically. And in some areas really where we are coming from, in order to have tight, secure, transparent supply chains.
You want to trace every component all the way when they are assembled to motherboards, be able to trace those motherboards, as well as verify the authenticity of the components and also of the motherboards to have a way, how to, then, trace if anything happens where actually these compliance and the motherboard comes from and what happens.
When the parties that you need, and also have the ability, we have a simple scan to see if it has been tampered with, if some integrated circuitry has been replaced or moved, and there is a chip hiding underneath. The cost of planting tiny spreadsheets is just $200. It’s been a great article in Wired about it. So this is the angle we are coming from.
It’s fascinating that as you’ve mentioned, as multiple verticals you’re looking at, although one of them is semiconductors. You mentioned as well, Jakub, about collectibles and sports memorabilia and luxury goods. And one of those use cases that comes to my mind might have been the early works of Veracity Protocol, which is about art exchange and art networks. We think of a Rothko painting or a Pablo Picasso painting that you would see in Tate Modern or in MoMA or in the Whitney. And when you see these different art works, you often wonder, are they authentic? Are they real? And the reason this comes to mind for me is there’s been a lot in the news about fake paintings and the trade market and creating all these lithographs and extra copies.
And you often wonder, whether it’s art for an investment or art for pleasure, Are you getting the real thing? And today the solution doesn’t seem that clear. With art, it’s what you might have, as you mentioned, some RFID tags, QR codes, but there’s a lot of ways to forge and mimic and redevelop these or plant these. So it sounds like, if I understand correctly, that this is a perfect trap in the market for a solution like yours.
Yes, our position was never to have the role of saying this is authentic and this is not. Our position is to provide you with a technology where you can protect that painting. And we can always guarantee this is that painting, which has been protected there in that time. We won’t be able to tell this is the original Rothko. That’s up to a person who’s protecting him. For some reason you may be wanting to even protect the fake. That’s really not our position to judge. This really brings you the identity to be sure.
For example, when an expensive painting is being sent for exhibition and comes back, it’s easy to swap it with a perfect copy and to make that copy in Chinese market. What’s impossible is to replicate it right down to a level of detail smaller than a hundred microns, which we already can pick up with just a smartphone.
And so it just takes the smartphone picture and knowing that this is really that original item or not. But to be honest, we are not really focusing on the art sector anymore. Our focus is much more on the sectors where we can actually bring value, by really saving lives or improving security. Or do a lot of improvements or setting foundations for future work.
That outward is a market, which is very resistant to innovation, to changes or other people want to keep the status quo. And that’s not really a market where we would like to grow a company and innovate.
That makes a lot of sense. With the art market, there’s been so much occurring, especially with outsourcing and development of paintings. There’ve been stories about how one city in China makes over 60% of the world’s paintings today, and how that was focused on in the last 10 years is now a vastly declining market, perhaps for the reason of authenticity or protection. And as you’re speaking of these use cases, Jakub, there’s so much we can think of.
We can think of Louis Vuitton bags, Rolex watches, these Picasso, Rothko paintings. But even manufacturing cases, like you mentioned, maybe a USB lightning port, something that you would charge your phone with, that you buy on Amazon or Alibaba or AliExpress.
I don’t know about you, but for me, I have gotten so many fake products from Amazon. It is mind-blowing. Just in the last few weeks, Amazon started a new process where they email people who purchase products on Amazon, if they were fake. And if the FDA and FTC are investigating claims on fake products. You never used to think this was something that was real. But today on Amazon, I got three of those emails on food goods, consumer food, grade foods, like protein powders and supplements that I bought back in 2016 and 2017 that now I’ve been told they’re fake. Who knows what could have happened to me? I could have been sent to a hospital or even worse.
So I feel like this fake product system is an evolving marketplace. That’s not just impacting Amazon, but also Alibaba. And it’s tough as a consumer to know what’s real and what’s fake, especially when reviews can’t even be trusted on these platforms. There was a research finding, recently, that said between TripAdvisor, Yelp and Amazon, in the US, almost 80% of these reviews are fake or paid for. So it is impossible to know what you’re buying. If it’s genuine. Unless you’re physically going into a store in Manhattan or a major city, or you get the manufacturer to sign off that it’s real.
So I want to hear your take. Are there certain products that you think are going to be good use cases, even for consumers with Veracity Protocol?
That’s a very good question and good points. There’s so much to unpack here. Just starting with multinational marketplaces.
Amazon is a huge marketplace. It’s not really motivated to crack down on counterfeits, because there will be most of their revenue probably. And there’s cultural factors. You will think that Amazon will be more motivated but actually Alibaba is, because chinese audience is more status-oriented. They care more about having the actual, authentic thing. They are afraid their friends notice, Oh, it’s actually fake. That will hurt their social status.
Another very good question is, where should this authentication actually happen? The example you mentioned is very scary and I know that feeling. Once I bought a screen on it, and then I got a call from the police that it was stolen. And that’s definitely not the way forward. I don’t even believe you should be doing the authentication as the customer. I believe that should be the role of the marketplace protecting you and doing this authentication for you. And right now we see many marketplaces like that. The RealReal, Gold, Stock X and others, where authentication is really their business model.
And I believe these will really start to cut market share of Amazon and grow by really providing an environment where we can trust that someone actually physically authenticated the things you are about to buy.
And then, really, the question is, would you like to buy directly from the producer? Would you like a marketplace which authenticates it, or do you want to have your own way to authenticate it as the original?
Some companies we work with, they see the use case of critical life, critical hardware, which is critical in a way, let’s say hardware wallets, cryptocurrencies, or which are counterfeits made in China. And you don’t really have a way of telling if this is the original or not, or if it hasn’t been tampered or not. So it will be definitely a use case where we would then help. But personally, I would much rather engage the level higher and bring technology to marketplaces, to work with brands, to work with manufacturers and to be able to totally mitigate this and solve this issue.I believe it’s possible.
It makes sense to focus on the marketplace. If I think back over the last 10 years, most of the purchases I used to do were directly from brands. In the past I purchased Nike shoes directly from Nike, but today I might purchase Nike shoes from Amazon. And one wonders if those new Nike vapor fly shoes are going to be real or they are counterfeits for the 2020 Olympic games?
The same happens if I purchase an item that was real from a platform like eBay or Amazon, where I purchased an authentic Hèrmes belt. And I believe that H belt is so authentic. I go to the Hermes store on Madison, in New York City. I say, Oh, I like to have this refurbished. And they tell me, no, it’s not real. It’s not authentic. And I say, how do you know this? It looks real. It matches the photo on the website. It has the same engravings of the number of your SKU and your model. How do you know is not real, you don’t give the consumer the benefit of the doubt and there’s no way to even verify or have the veracity on that.
So there is a lot of he said-she said going on there in the world of authentic and verified goods. And it sounds like, as you mentioned, Jakub, we see today some platforms like The Real Real, and StockX, which are beginning to help alleviate these problems by building new marketplaces that verify the goods.
Where do you think StockX and The RealReal have been successful to date? And where do you think there’s room for improvement for perhaps Veracity Protocol to partner with platforms like them?
So before I answer that question, one very good point is this is the era of when we start having deep fakes in video and audio, it’s kind of a subject going on. The same thing happens with physical objects. You have super fakes that are really impossible, even by a naked eye, to distinguish the details. You really need to be drawing, train your nets and computer vision to really be able to go down and recognize the difference between the fake and original.
It’s the same automation we see elsewhere in healthcare to recognize cancer cells. We always get much bigger, better accuracy and precision than the human eye. And coming back to your question, this is exactly the area where, then, these marketplaces have to rely again on the physical menu verification. Instant we would offer them a way to build data sets to automate this and to be able to scale without the need of relying on manual labor and this experience, which then fades as people get out of the company, or sometimes it’s not even enough. And really, through optimization and AI, increase the accuracy of the process and be able to scale it.
That makes sense. Because today you have gemologists and leading experts with jewelry who could look very fine under microscopes to say this is authentic, or this is not authentic for diamonds and products like that. But the challenge, as you just mentioned, Jakub, is then the individuals exit, or they get retired from a company as they age, and then that knowledge goes with them. It fades. And so this begs the question on how can we evolve from a manual process to an automated process?
And thinking about Veracity Protocol and these unique fingerprints that are being manufactured, all types of physical goods, something struck with me that you shared earlier in our conversation. I used to think that I needed a super high resolution camera. Without a massive Sony, Nikon, Canon camera costs thousands of dollars, there is no way I am going to be able to do effective computer vision or effective photography that can see deep within these photos. But what you’ve shared with me offline is that it cannot be further from the truth.
Perhaps the cameras we have today can see in-depth detail. How in-depth is that detail that we can see from our smartphone cameras? Like the iPhone 11 pro, the Samsung Galaxy S10, the Huawei P30 pro, the Google Pixel 4 and so forth.
So to give you a comparison of what’s the level of detail we can work with, it’s just a smartphone with basically a 12 megapixel camera going all the way back to two iPhone5, what we can really recognize is a sheet of white paper against each other.
You can buy five postcards, you can with a smartphone, and protect one, let’s say this is my original. And you would see that we would always recognize the original and reject all the other postcards you would ever present in front of the camera.
So there’s a level of detail we can work with and with use cases where you need more detail, let’s say smaller semiconductor components, of course go deeper, use higher resolution. And now we will be talking more about specialized hardware or ultimate-list manufacturing or assembly line where we had the camera installed and this is done automatically.
It is absolutely incredible to think that the iPhone 5 from 2013 has the technology capable of understanding the difference between two white postcards at this 100 microns or smaller level. So in fact, if I understand correctly, everyone in the world, most likely has a smart phone capable of recognizing something that the human eye cannot be trained on.
And that’s incredible. But one of the challenges as well is that the hardware’s finally catching up. We have seen that it’s been around since 2013, but now basically every consumer in the world has a smartphone that can look at photos at a 100 microns level or smaller to detect these anomalies and tamper at the different angles and materials that you’re building at Veracity Protocol.
And the software is only catching up to the hardware now, with new algorithms and packages like PyTorch. But one of the biggest challenges that I see present in the machine learning world is data and how to get real data and synthetic data. How did you explore the data problem for the work you’ve been doing?
So let me just start by saying you mentioned a very good point. Everyone in their pockets already has a device, which we can empower them with our algorithm. This is something of a breakthrough on our side as well, that the ability to design an algorithm, which can work really true with various lenses, various cameras in various conditions and light.
Even when the item gets damaged, you’re wearing term or clothing stretches and stuff like that. The idea is really, you can take your phone, snap a picture and know immediately. All this is truly based, not only a lack of building this computation algorithm, but also the data sets and the data, and all this data we had to acquire by ourselves.
There’s no more data sets or databases of camera lenses and material structures and different types of objects. So this is something we’ve been building for over two years. We have our own extensive data sense. And the first data set ever created was really neat. Painting a canvas white and painting another canvas. And basically dissecting it into hundreds of small pieces and taking pictures of it still really creates the first data to run tests on it, to see if the hypothesis holds.
Moving beyond the canvas. So segwaying back to manufacturing and material products. Another interesting arena that’s grown a lot in the last few years has been 3D printing.
I have a good friend in New York who has their own 3D printer, and we’ve made a lot of cool items there using different CAD models online. And I wonder, how unique is the structure and the materials of items? Can a 3D printer replicate or can a 3D printer’s print be identified?
It’s a good question. So every object, the 3D printer prints will be different. Its structure will always be different. It just depends how deep you need to look, how much resolution you need.
What’s cool about 3D printing actually is, and which links or artists in general that you mentioned in the moment of 3D printing an object you can already protect. You can always know this one belongs to me, not only based on the temp design or the 3D printer we are using, but actually based on the object structure, which you will never be able to recreate on the 3D printer.
So that’s good to know that no one, any time soon, will be printing forgery on 3D printers. It should hopefully be a more hobbyist activity for fun at home, but as we’ve learned the identities and the data that drives our digital-first society today, is anything but real and anything but certain.
There was recently a big movie out on Netflix called The Great Hack, all about how you can change hardware, how you can change software. And that was very much focused on the 2016 election in the United States. And we thought that was gone and done, that what was uncovered was about, potentially, if there was interference, and all these topics have been so focused on in the news, we thought it was only 2016.
But thinking back to just February 2020, what we’ve seen is that the Iowa Caucus in the US for voting with the democratic candidates, was mired in issues around their software company named Shadow Of All Things, as well as the hardware and integrations and mobile apps and testing to see about their viability of voting and these digital certificates. It’s amazing to think that we are only just now uncovering the surface of what’s possible in verifying and validating data such as what you’re doing at Veracity Protocol.
But I know all news sounds bad, but not necessarily that’s the case. It’s just what’s trendy at the moment. But thinking about global trends and the movement in the industry, there’s a lot of good things happening. And it sounds like what you’re doing a Veracity Protocol is only just the beginning. What’s new in the pipeline? Are there any new products, releases, announcements that are coming out?
So one of the big ones, obviously, is right now in the RSA constraints, presenting a solution together with Intel of securing the transparent supply chain for the critical hardware, like several motherboards, showing the solution, how we are able to fingerprint individual components, several motherboards, to be able to allow anyone down the supply chain to verify this is the authentic motherboard, what’s actually its history. All these components authentic have something being replaced. So this is a solution we’ve just unveiled and we are very excited about, as this really presents the cutting edge of what our technology can do.
And that’s not the only thing. Another area of our interest is really working with big manufacturing companies and companies and supply chains to really build the industry 4.0, with automated factories, where components and final products are tokenized and everything is settled on blockchain, as it moves down through the supply chain through individual parties in the process.
That’s definitely something extremely interesting. You may think like why not use just a QR code or ID or barcode there? And I was actually asking this question, as well, to one of those companies, and they just believe it’s not enough, it’s not a secure enough solution, as it can be very easily
copied or manipulated with. So this is an area of longer term interest in something they really believe our technology can enable.
It’s amazing to think that at the RSA Conference you’re presenting this new solution with Intel and, as you mentioned, you’re enabling secure hardware. For our listeners on the HumAIn podcast, one thing that may not be common knowledge is that Intel powers the chips of most of the computers in the world, and most of the Apple iPhones in the world. Who would think that Intel was such a big player in the space for power from processing and that they’re also looking at security? What are some of the trends that you think are going to emerge from the conference or emerge further out in 2020 with security?
From the point where I’m standing, I’ve seen security is definitely moving more into the physical world and getting much more attention there, because, thanks to the democratization of technology these times, 3D printing, the availability of chips and stuff you can already do after a few tutorials on YouTube is really amazing. And this brings further pressure to create barriers of entry to bad actors in the supply chains, which you really need to counter with more advanced technology.
So it’s definitely something we see, not only concerning counterfeits, but exactly as you said, concerning several motherboards, which are running critical government or defense servers, for example, or your smartphone. So that’s definitely the trend we see.
The final trend that we don’t know just yet, but there’s been a lot of talk lately from all the major reports and trend setters about what is the next device we’re moving into. We know that the internet has spawned that emergence of portable workstations, like laptops, as well as iPads and smartphones and all these mobile devices. But I’ve been seeing a lot of research lately about smart glasses. Even Amazon came out with their own new one to compete with Google Glass 2.0 all about having laser prints of data etched on our eyes or in these contact lenses. So I wonder if that’ll be the next smart device. Any take on which device you think will be more helpful to do some of this verifiability in the future? Do you think maybe smart glasses could be that device?
That’s an interesting view. Definitely. What I see regarding the trends is really, ever since moving into the cloud with upcoming 5G, everything will be in some instant. The verification down with a smartphone with our technology will be instantaneous. You don’t really need to carry much hardware, hard days, all this processing power and stuff. If you can rely just on the 5G and good coverage of that, that would bring a lot of gadgets and devices to further easily, embed you as a person to be able to wear them, it will be glasses. It will be airpods without an iPhone. It will be just screens, not really, computers, which only continues the trend of enhancing ourselves right now. It’s items we put on the table, we hold in our hands. Then it’s wearables, then it will move into your skull. And that’s definitely the direction where it’s going.
I’m looking forward to that world, whether we’re called enhanced humans, whether we’re embedded people with these personalized experiences, it sounds like we’re continuing to move into a world where it’s humans and machines, and that’s only going to continue to be defined in 2020.
It’s great to hear everything you’re working on at Veracity. And Jakub, thanks so much for joining us on the HumAIn podcast.
That was exciting. Thank you, David.
Hey humans. Thanks for listening to this episode of HumAIn. My name is David Yakobovitch, and if you like HumAIn remember to click subscribe on Apple podcasts, Spotify or Luminary. Thanks for tuning in, and join us for our next episode. New releases are every Tuesday.