Seekr’s Rob Clark on Collaborating with Intel

This is Techron tv. Hey everyone. Welcome back here to Techron tv.

My next guest is Rob Clark. Rob is the president and CTO of Seeker. That's S-E-E-K-R.

I'm gonna tell you all about that, a little bit about himself. We're gonna talk, we're gonna talk some ai, some gen ai, and no one ever talks about Gen ai, it seems. Hey, Rob, welcome to Techstrong tv.

Nice to have you on here, Alan, an absolute pleasure. Thanks. Uh, thanks for having me.

And thanks for spelling Seeker Rod as well. Like I'm beyond impress. Well, you know what, I knew how it was spelled, but I, I always liked to, when something's not readily apparent, I like to spell it out.

'cause a lot of times people will race right now, right? They're watching this on their laptop or something, or a pad, and they'll go to the website, say, okay, what's this guy about? What's this company about?

So that's it. I Love it. Try, try to make it easy for him.

Hey, Rob, what is this guy about? Let's hear about the Rob Clark story. Yeah, fantastic.

The, the Rob Clark story is, uh, is interesting. You may not find it. So, uh, look, I've, I've been in software for whatever over, over 20 years now.

Um, back before AI was, uh, was cool, like ML earlier times, NLP, um, right through to where we are right now. And look, you, you would say AI two years ago, three years ago, and people thought you were talking about the Terminator or something awful. You say it now and everybody's like, yes, I know of it.

I use it. I've seen chat. GPT I've seen the good, I've seen the bad of, of all of these generative as well.

But I've really worked in that for, for a long time on everything from, um, smaller scale web, right through the large scale web search engine. So Seeker, some of its previous technology, almost goes all the way back to nine 11 as well, post nine 11 with the need to look at content, to understand content, to process it and pull out trends to find kind of separate the, the reality from the, uh, from the made up as well. And I think everything we're gonna talk about today, which is really what, what Seeker is doing and what we're bringing to the market around trustworthy LLMs Foundation models, all grounded in trust, transparency and explainability.

And also contestability, which I'll touch on as well. But obviously all that experience I've had across what, four countries I've lived in now I've worked in kind of even more of them, but all of them, A common trend is how do we build large scale systems that can analyze, um, information and give the user, whether that's a concealer, whether that's an enterprise, a business, the right information as as, as we always say, um, being informed and not influenced. I like that, being informed and not influenced.

If only we can make that true. I mean that, you know, as I sit here, of course we're in an election season in the US right? And we've been talking on, on Techstrong TV as well as not on, you know, just privately what influence ai, gen AI is gonna have on this cycle, on this election cycle.

Yeah, no, huge, huge, huge is the easy answer. And really, we, when we founded Secrets, really off the back of what we saw in the last cycle, it's what we seen. I Was just gonna say that.

And the last election, even without Gen ai, we saw what social media and fake news and all that stuff, you know, did to, to credibility and to people's confidence in elections and everything. That's, but frankly, that was baby do, do. Yeah.

This is compared to what can go on this time, a Hundred percent. I, as I was saying, that's why we founded this. That was the incentive.

And like you said, it was being done by people, but it was the scale of people. You think one person sharing something, uh, call it misinformation, something harmless. Grandma saw something on Facebook, thought, thought it was funny, shared it with the family, so nothing too bad.

But then when you start looking at 'em, what we were seeing was really coordinated inauthentic behavior. Like that's the scale of it. That's the problem that we've been seeing.

And thus, we built a system, we built a system that could really look at 50,000 news sources, domestic and international, that can look at all the different formats as well, image, video, web, across the whole set, and start to understand what was going on within them. Was it, was it, um, was it method that talk that, um, kind of attack people's biases, their cognitive biases? Is there a political lie that's hidden inside of there as well?

And then really importantly, the three words that we love are really provenance, lineage, and objectivity. Where did it come from in the first place? Where has it been since?

Has it changed? Has it morphed? Has it mutated?

Has, was it an image and somebody was swapped out, or was somebody put in now with somebody shaking hands with a Putin that was never even in that photo in the first place to begin with? Or somebody got chopped out of it too? And then what was the objectivity?

Is this deliberately to mislead? Was it grandma just popping something up there and, and kind of not realizing what, what she had done, but all of that coming together. And I think that that's why we're so passionate about what we do at sea, is we want people to see more about it.

We, we wanna do the difficult work. We know ai, we know scale of information, we wanna do the hard work to give people the answers, but with all of that transparency, like who are we to say something's credible or not, show the score. And that's really what seeker's been all about, is look at the information, understand everything about it, where it's come from, where it's been, and then show that to the end user.

It's al it's almost like an antivirus when you go back to this file might be bad, click it at your own risk. We're in the same game. We're not censoring, we don't wanna say, do not read this.

Like, how can a world exist when you can't show different points of opinion, but where something is not authentic, where something could have been even more advanced techniques like information launder, and it could have been laundered through to look authentic to you. And I couldn't, I couldn't agree more. Alan, this election cycle is gonna be off the charts compared to where it was generative and what's needed to counter that so important.

Absolutely. Yeah. I was listening to you talk and I was reminded recently in the uk, uh, uh, princess Kate and put out, put a doctored photo sheet.

She's an amateur photographer who plays in Photoshop. So what, what person Photoshop doesn't, you know, take a few blemishes out or, or do a little Photoshopping. They, they made it like it was the most evil thing in the world.

You, she, she's hiding, you know, of course, unfortunately, you know, with the cancer diagnosis and everything and, but this seems so innocent and people, yeah, it's started putting evil connotations to it. It certainly, uh, blew up on a few people what came out of it, but that, that's why the word objectivity there, like what was the intention? It's like she didn't add people, she didn't remove people.

It's all the, there, there was, there was some blurring on the cuff that was out of alignment and stuff like the, the harmlessness of it. But I think where it's important as well is also flagging that information too. Like, I think it's, it's okay to say this, this was modified, this was beautiful, Whatever this is enhanced or whatever you want to Call it.

Yeah, right. It was, it wasn't generated completely from ai. It was tools that people use.

And look, we, we we're all in the grocery store or the supermarket for the Brits. You see that magazine on the shelf right there? It's been photoshopped, it's been touched up.

It's just at this point people are really worried about the things that are generated from scratch as well. And they conflate the two. But the other thing I love that you picked up on it as well, Alan, is really this idea of sensationalism as well, right?

Yes. This is almost the, the, the problem the media can have as well as they want to be objective. But where does your money come from?

It's coming from ads, it's coming from readership. How do you get that? There is some degree of sensationalism in them.

Where we've come at this is flag that as well. Like obviously that's not true journalism, but flag it, put it in there. Was it exaggerated?

Like, and then in the worst case, was it click bait, whatever that was. But that, that all, uh, the World we live in, you know, we, again, we were discussing this on Textron Gang and it played, uh, played on Thursday. For those who want to go back this past Thursday show.

Um, and I'll give you another example is, you know, here they're looking at, and I'm, I'm sure other countries are too. We're looking at, uh, possible criminal statutes, uh, not just civil for people who are using AI and doctoring up stuff. And, and for sensation, I mean, perfect cases, two of the leading women's college basketball players here in the US have been the victims of ai, kind of deep fake Yep.

Nude and, and embarrassing photos, right? That at first weren't so apparent that they were doctored. People thought things were legit and, and it did a tremendous amount of harm.

These are, at the end of the day, these are young girls. They're 19 years old, 18 years old college athletes, right? Maybe 20, 21 even.

But you know, relatively young girls and you know, their parents and families are out here seeing pictures of them reportedly nude and, and engaged in all kinds of lewd activities. You know, it's one thing to say, okay, we can use Seeker to ferre it out, that these are in fact fakes or they are doctored or they are, you know, ai whatever, whatever the verb we're gonna come up with for this type of thing. But it's another thing to say, okay, what should the repercussions be?

If we can find out who did this? Is it a criminal? Does it rise to the level of criminality that someone should go to jail?

Should someone have to pay a lot of money? Um, you know, what, what as a society, what penalty are we putting on that? No, I can couldn't agree more on that in my, from my simplistic view, it's like the level of the crime and the objectivity, right?

Take things like somebody dying. There's different levels. Was it a certain degree of manslaughter in an accident?

Was there malicious intent? Was it murder? Was it absolutely brutal murder and then a cover coverup afterwards?

I think, I think that obviously has to be thought about like intention and what it was. But look, they're not, like you said Ellen, they're not, they're not harmless crimes in those senses as well. You think, No, they're not victimless Like leak nudes and things like that.

Now it's another level. It's take somebody's face, take a cloth shot of them, and all of a sudden there's technologies that these serious harm and like you said, for, for young kids, for teenagers, their images, everything, right? Well, the how they see themselves, how others see them, and like the emotional part is bad.

But look, this, this, this is why it's so important to kind of, to have that adversarial technology when it's being misused. We also need the ability to do it. And the, the other big part of that is, um, building in the trust into that generation as well.

Yes. Like being able to detect it as one thing. The other part is how how do you then have generative systems that that do good, that are trustworthy?

And really that's what we've been focused on too. And we, we just signed a recent, um, deal in collaboration with Intel Corporation to use their latest generation that gaudy three chips or gaudy two and gaudy three chips, their AI processes on their, uh, t developer cloud. And that's really what's powering our systems right now at scale.

And that price performance is so good, we can pass it on to the consumer. So when you're talking about people scoring information, when you're talking about trusted LLMs that can detect this, we even starting to build foundation models now as well with that trust baked into them using that scoring. We know what goes into those models.

We know where the bias is. We know where the maliciousness is and the intent to do those things and then give a better outcome to, uh, the people using them. Whether that's consumer or importantly for businesses and enterprises.

'cause they're, they're, they're in a, they're in a similar boat right now around image and how people see them. So how their generative technologies, how their brand, even when it's being advertised. We have a product called Seq Align that actually looks for harmful things within content as well that you don't wanna put your brand against.

So we're listening to, um, tens of thousands of podcasts of audio to see what, hear what's being said to classified understand it, and then let, let companies place ads within that or not, as the case may be, find more good audience, avoid the bad stuff. Like a lot of this for consumer or business is about your image as well. Great.

You know, you listening and thinking about it, Rob it. So there's really two main parts of my mind. One is to have the technology seeker like technology that can at least recognize that, right?

Yeah. 'cause that that's job one. If you don't realize it's faked or enhanced or Hey, I don't do anything about it Right then, then you're, then you're really left at the mercy of, of these people.

And then the second thing, you know, I learned is I went to law school a hundred years ago and, and one of my professors there told us that, and this is back then before there was a commercial internet or anything, that the pace of technology is such that it takes, you know, two to three sometimes more years for society's, more for society's kind of norms to catch up to the state of technology. I would, I would present that that was before there was an internet, before there was everything, social media. It probably takes society even longer, I think today to catch up to technology.

And I think that that's the whole reason for Seeker to be here today, I think right in, in many ways is we need, we need a tool to help us do this. So important. Look like you, you hit the nail on the head.

You think about the reason change often happens psychology wise is when does it affect me as a person? Then when enough people feel personally affected by it, then they start to take action. But look with the, what's clear right now is this has been happening for a long time.

Whether it was machine generated at that scale or not, it was already happening. Again, one person, you then move forward to troll farms and kind of scaling people to do it. And now every one of those people that's been scaled can do it again.

And that's, that's why it's so important. And I think, um, within AI as well, everybody can see the pace it's moving. Like every day there's something new going on right now in the news around ai.

And look, an enormous amount of money is being spent around kind of partnerships within AI right now. Whether it's Nvidia, Intel, any of the players, we're right in the middle of that war right now. The next 12 months, there'll be a lot that goes on around acquisitions m and a because they're trying to control the infrastructure.

There's a battle over getting the infrastructure in place and obviously why we needed to secure scarce resources. Right now, AI processes are scarce and the availability, and they've gotta be the right cost. Like if we're gonna offer this to people as well, it's gotta be at a low cost to them.

So by us doing this collaboration with Intel, we can drop that price. So when we go out to the market, when the tools are there, so people need them, they can also get them earlier. And I think to your point Alan, some of that lag in the cycle of people doing something is sometimes they don't know how and be, it's expensive as these things.

Yes. It's, you can drive that cost down at the same time as well. I mean, there's a whole discussion over how many of these AI process, you know, if you're going to do the Nvidia stuff, it's the G GPUs and so forth.

But how, you know, the cost of running all of these AI processors, the on the environment, on on infrastructure and everything else. That's, you know, AI can almost be a victim of its own success in that alone. But that's, that's farther for another interview, man.

Rob, we are, We're yeah, that's, I was gonna say that like, you think, you think about investors right now. First it was put money in software, then it was the hardware, then it was the data center. Now it's the power stations.

Like its the power Plant. Now you gotta generate clean power at that, right? That's it.

And let the, the silicon bits the other part of the equation. Like you look at the CHIPS act, what's going on there right now when we talk about us as a society within the US kind of defending against these things and giving people the right information. I think that's where things like the chips are important.

Like people are not going down, Well, they're about 6 billion here, 8 billion there, 9 billion there. Look, that's, that's the investment to keep it, uh, stake, to keep it unsure. And the investment now is what protects not only bolsters right now manufacturing the us but insures within years that we have the things we need.

Because look, China isn't stopping, others are not slowing down right now. That's why they're important. That's why the steps now and again, why, but it's also why, why we get a deal with a manufacturer So important.

Why absolute important, so's important one of two that can do it. Rob. It's also why secret's important, right?

Because that's there when you have China and I'm not, I'm not a a China bomb thrower, right? But it's that. But when you have others that you know, may have different aims and goals you need, you, you, you need that referee.

You need that, uh, you know, clear, clear line, not blurred line. So you wanna make a decision, at least make a informed decision. Hey man, we gotta run onto our next one.

Again, it is seeker. com? Yep.

Dot com. Yep. Go check them out.

This is an issue that every single one of you are thinking about, I'm sure. So these, these are folks who are doing something. Rob, come back and visit us soon, man.

I appreciate you. It will be a pleasure. Thank you so much.

Appreciate it. All right. Bye bye.

Rob Clark, president, CTO Seeker, you're on Tech Drunk tv. We're gonna take a break. We'll be right back.