10. AI is the New Blockchain – Tech Field Day Podcast

All technology is going to be hyped to the point where it is an inseparable part of our everyday life, no matter whether it's tulips or open flow or blockchain. Which brings us to ai. AI is the new blockchain on this episode of the Tech Field Day podcast.

Welcome to the Tech Field Day podcast, the only podcast that dares to be both on topic and on location. My name is Tom Hollingsworth and I'm a part of Tech Field Day as an event lead for all things related to networking, mobility, and security. And each episode we bring you the perspective and thoughts of a group of experts in the IT field, tech field day delegates, and we get an idea, a premise, if you will, and just kind of work with a little bit because it's a lot more fun that way.

I'd like to take a moment for our guests to introduce themselves before we jump into the premise for today's episode, starting with Evan. Thanks, Tom. Uh, my name's Evan Mintzer.

I am Director of Production Infrastructure for Customers Bank. I'm located in the in Pennsylvania and looking forward to this conversation. Hi, I'm Jody Lewan.

I am, uh, ride run TCO Networks and I'm an independent IT consultant in the Niagara area of Canada. So let's jump into the premise. It's an overhyped technology.

It's the solution to all of your problems. We're gonna find new ways to take old things and integrate it with that technology and do exciting, wonderful stuff with it. But there's a dark side.

We don't know exactly what's going on and the people who are trying to drive it, are they doing it for the good of society or are they doing it to try to make a quick buck? And if you're asking yourself whether or not I was talking about AI or blockchain, well that's the premise for this episode. AI is the new blockchain.

All right. All right. Stop typing everybody, because I know you're about to leave me a hateful comment about that's not the same.

And blockchain isn't bad and AI is great, and the parallels between those two technologies are more than a little striking when you think about it. Blockchain is past now, let's be honest. It it, nobody cares unless you're trading NFTs, you're not still trading NFTs, are you?

Okay, I hope you're not still trading NFTs, but let, let's, I'm gonna throw this out to our experts here because trust me, before we started recording this podcast, we were already riding this bandwagon straight off of a cliff. What is it about these technologies that make people seem to think that it is the hammer for every nail that you could possibly have? I, I think that's a really good way to bring it up, Tom.

And, and the way I see AI and, and it becomes, it is a tool and I think it's a wonderful tool and can be used well when used properly. The problem is you get a lot of executives, and this happens with a lot of other technology technologies in the past. You get a lot of people that are saying, oh, we can use AI here and here and here and here and here, and they're kind of missing it where it probably shouldn't be used everywhere.

So what you're telling me is, is that people spend a whole lot of money developing something and then they wanna make sure that they're getting their money's worth out of it. So they're shoehorning it into places that it probably doesn't belong. This is my shock face.

It's like going to a buffet. You spent money going to the buffet, you gotta eat and eat and eat to get your money's worth. I'm, I'm not disagreeing with that because that absolutely is true.

You feel like you have to get your money's worth out of it. So if you have one more plate of shrimp, really what's that mean in, in the long run? I mean, yeah, you're probably gonna get sick and, and all those other things, but it's worth it because you got your money's worth outta it.

Yeah. Ai, AI is something that, you know, a lot of people like to use. And, and as I said, it, it does a good job when done correctly.

And I think like a lot of technologies, you can use it correctly, but you can also use it incorrectly. And if you're trying to shoehorn it in where it doesn't belong, that's where it starts to fall apart. That's, that's where the similarities between AI and blockchain come into play.

Blockchain is a really interesting technology and has some very cool applications outside of non fungible tokens and cryptocurrency, but it's a tool to a specific solution. AI is a tool to a specific solution. And at the peak of the blockchain hype, it was a tool looking for a solution.

And I think AI is in the same boat. Now. There's a lot of places where AI is perfectly good, but it's not everywhere.

And right now everyone wants AI to be the solution to every single problem. And in doing so, sometimes they make it the problem. Well, what is it about ai, blockchain?

I mean, how far back can we go open flow that we, we develop these tools and then we try to apply them to a problem? But isn't it that most successful startups find a problem that people have and then try to solve it? Like, I feel like if the, if the, there's a market mismatch, like we're, we're trying to jam that star shaped peg into every possible hole that we could find, when in fact there is a star shaped peg over here.

So we have, we have good allocation. 'cause we know that AI is really good for certain things, and we are finding adjacencies every day that they're better at. I think part of the problem is, well, the, the, as at the time of this recording at the end of May, 2024, one of the things that's coming up is, uh, Microsoft just had this big new announcement that they're gonna be doing copilot plus PCs, and they have an NPU that's doing some offloading stuff.

And at first the idea was great. I mean, they're doing the usual AI related stuff where they're like, oh, you can do image generation and image editing. And then there was, um, the rewind feature, I think is what they're calling it, where it's like, oh, we're gonna take screenshots of your PC like every five to 10 seconds and store them in a secure enclave so that if you ever forget what you are doing, you can go back and jog your memory by going through your screenshots.

And I don't know, maybe my grandmother thought that was a great idea, but a whole bunch of us privacy experts went, you're gonna do what? No, because you're applying the wrong part of the tool to a problem that nobody has. Realistically speaking, when is the last time you sat down to your computer and went, I don't remember what I was working on five minutes ago.

I know I'll use this really random feature to remind myself, I know Tom, as we get older, that happens more and more. Eh, we're not bringing age into this, my man. We're, we're not talking about me getting old.

We're talking about the fact that AI is trying to solve a problem that I don't have. I think, But, but that's another aspect of it, the opt out. It's like, maybe we don't want it, but we're not being given that choice.

In a lot of cases, it's, Hey, you know, Facebook, the search is now meta ai. That's what you got. Hope you like it.

Well, I mean, we were already exposing some of the problems that AI has introduced into a mature technology. Again, at the time of this recording, at the end of May, future historians, if you want to go back in the way back machine and see all of the ridiculousness that AI integrated search with Google has been coming up as someone put it, and I think it was the most apt, uh, analogy ever. It's like asking a 4-year-old to solve a problem.

They're gonna give you the dumbest answer possible because they don't understand anything else. Like, what was it, um, uh, pouring glue on a, on a pizza to make the cheese look bigger or something like that. Like, yes, that is something that I expect an untrained person to do, but the hope is according to what everybody that I've talked to is, oh, well, it'll get better.

Which this sounds suspiciously like, oh, blockchain is slow, but it'll get faster. Well, the other thing we run into on that one is that, um, it isn't good now, but it's getting better. The problem is, is that we have too many people who are perfectly willing to accept what isn't very good now and not do the thinking for themselves.

I, I don't think for myself, I have AI to do that for me. Exactly. Like, I'll just get it to spit out this Python code and commit it to production.

Let's make this happen. That, that sounds a lot like agile methodology where you put out an MVP, you know, minimal viable product and then, you know, okay, we'll put it out. Just, just get tires on, on, on the base and we'll get an engine later on before the car works.

So with agile methodology, you're making your mistakes and you're correcting your mistakes. It's a cyclical thing. If you're taking AI output and just grabbing it and, and using it, you, you've now created a much bigger path to fixing your mistake because you never understood the mistake in the first place.

You've got a point there, Jody, and, and this is one of the things that we talk about a lot with other industries that are maybe a little bit more highly regulated or maybe a little bit more, um, intolerant of failure. It's like, uh, NASA is a real good example of this. They, they don't prototype stuff the way that we would prototype stuff like deploy to production in nasa, is it better work every time?

Uh, as, as a, a recent NASA engineer said, the next rocket we blow up will be our last rocket. Um, they, you know, they, they don't make those mistakes. But at the same time, when we make a mistake in development, when we make a mistake in writing or something like that, we go, oh yeah, don't do that again.

AI doesn't have that kind of a memory, like, unless it's programmed to do that. And I think that maybe is one of the biggest problems that we've run into with AI is by shoehorning it in everywhere that it doesn't belong. It's not that it's not having an effect because everyone's hope is, oh, well, if we introduce AI and it doesn't actually make anything better, then we didn't cause any problems.

So then the loss is a net zero except it actually is causing problems when we shove it in places because it's making recommendations to fix things that aren't broken. Like when we tried to take a blockchain and apply it to all of these really hard problems that blockchain was ill suited to solve, and yet everyone was like, oh, well it can't possibly be any worse if we put it on the blockchain. It most certainly can.

The rule is it can always be worse. It's true. They go, oh, it can't be worse.

No, it can always be worse. And again, just to reiterate, I'm not saying that AI is a bad thing. I'm not saying blockchain is a bad thing.

I'm saying understand the tool and apply it to the application. There's a problem to be solved, and the, the problem to be solved with AI is not represented by an asterisk. I mean, I, I think you're right.

I think that a lot of people when they get into this, they, they look at something that they don't wanna put the effort into solving, and they're like, well, I'll just unleash this algorithm on it because it's obviously smarter than me. No, it is as smart as we have programmed it to be. And that's one of the other problems we ran into with OpenFlow years ago.

And I, I know I go back to some of these old technologies and people out there probably laughing in the comments going, oh yeah, he is talking about old stuff. I remember a time when OpenFlow was a solution to every networking problem that you had, except it was too slow. Except it was, it required way too much over-engineering of the problem.

0, Dinesh Gut was on stage saying that, you know, Python is the new open flow because the massive amount of re-engineering that people need to go through in order to make those kinds of things work scares the daylights out of them. And maybe that's why things like AI appeal to people so much is because I'm not re-engineering the problem. At least I'm not, maybe AI is rewriting everything on the back end or giving me a different code or things like that, but I'm not the one who has to shift my mindset.

So could it be that one of the reasons why AI is so hyped, like so many other things have been in the past, is because it reduces our uncomfortability with change. That's definitely possible. I'm a little more cynical and think it reduces our dependency on hiring people.

It streamlines the work process and people see that they're saying, you know, I can, I need to write a, I don't know, write a document could go in there to chat GPT and just ask it to do, write up the document, and you, you get a document. I, I find it funny when you, we use the, these, these third parties, copilot, jet, G, BT, whatever, and there's always that little asterisk there where they say, you know, you should be careful with, with the output. You'd always double check the output.

And how many people are actually double checking the output? I gotta document out of it How many organizations have people who are skilled enough to check that, like half the time you're going at AI so that you can get things that you don't have the local skills to take care of. And that's problematic.

We've already seen it happen in another of other industries. Uh, probably the one that I can think of more than anything else is Sports Illustrated. Um, effectively their writing staff is either left or they're running on minimal skeleton staff and a lot of their articles are being written by ai.

And you can see there's, you, you, it's the uncanny valley problem of you can tell there's something that's not right about this. You can't quite put your finger on it. Um, you know, I've seen it otherwise, you know, like people don't talk like that in movies.

Like they don't have dialogue like that. If we wrote movies the way that people talk, nobody would watch them because they sound dumb. But yet that's what we do.

And I think that we're starting to see this transition in AI to, you know, it's, it's finding its niche, which is great, but unfortunately for people who wanna invest money in ai, the niche is boring. And that's what we ran into with blockchain. Blockchain is great as a distributed, secure, immutable ledger.

It is not the fix for all of the global financial sector problems. And what we're starting to see is that the companies that were betting on that, whether they were, um, you know, defi companies or what have you, are starting to collapse under the weight of their own overinflated valuations, uh, turns out that it's not that important. And this is not, you're not a lynchpin, you're not too big to fail.

But yet I flip around and I see the fact that Nvidia is now as least at the time of this recording, almost a $3 trillion company. Will there come a day when those same valuation resets happen to companies that are doing in massive investment in infrastructure and consuming more power than nuclear can provide for everybody? That's a crystal ball.

I don't have, Come on prognosticate. Tell me if I'm wrong, 'cause I know there are people in the comments that are doing the same thing right now. Punk Phil Prognosticator of prognosticators.

Okay. Uh, I'm gonna leave it to Pennsylvania for that. Sure.

Bring it back to me, right? Just for a minute. I'll, I'll speak in my own time.

Yeah, no, it, it, it, I go back to the tool. It, you know, everything's a tool and, and when used correctly, tools are great. AI's a great tool.

It's just when you don't use it correctly, it becomes a problem. Yep. I'm behind that.

And then Nvidia at least has hedged hedged their bets. Like what they're heavily investing in is raw horsepower. It's necessary for a, for AI applications, but that raw horsepower can be applied to a whole lot of different things.

Coincidentally, blockchain too. Um, so when AI go, if AI right sizes and the next big thing that requires massive horsepower comes down the pipe, Nvidia is gonna be just justified. But we've seen this already from Nvidia because a lot of the technology that powers these AI clusters is actually a growth out of the HPC market.

And they were able to kind of pivot quickly away from, you know, HPC, which never got to the lofty, like expectations of blockchain and open flow, but it was still a hype technology. It just never grew to the point where it was over-hyped. And a lot of that allowed Nvidia to kind of pick up and run where a lot of other companies were still trying to adjust.

And I think we're seeing that now where more and more competitors are starting to come online with, with new things like IBM and Intel. And honestly a lot of companies that we cover on the Geal, it rundown on a weekly basis. But I wonder if that kind of speaks to the fact that blockchain was already this proven solution that people were trying to associate with other things in order to get it to a point where it could be big enough to make them rich.

Because I think that that's one of the things that we're starting to see that has this common thread is you come up with a technology that has potential to be revolutionary, but what generally ends up happening is that there's a certain class of user who winds up there whose first thought is, I'm gonna be filthy stinking rich, I'm gonna do whatever it takes to make this the vehicle for me to get filthy, stinking rich. And that kind of deflates the overall viability of it because rather than letting the product grow into a market space where it's apt to do things, it's let's stick it everywhere we can because one of these slot machine handle pulls is gonna pay off and I can retire. I think the answer's gonna be the same thing as it was for blockchain.

Same thing as it was for open flow. Wait for the hype to fade. The technology will either right size or fade with it and then decide what you wanna do with it.

I mean, I'm just thinking about what you said and, and it sounds like a lot of startups, um, you know, come up with a really good idea and then try and make as much money off it and get out, sell it to a larger company or, or, you know, make the money and go. But you're, and you're right, that is the whole purpose of a startup is to come up with a new idea, marketize it, productize it, and ship it. If we didn't have that culture, we would be in a lot worse place than we are right now.

But I think that the difference there is that if your product isn't marketized, productized, operationalized, eyes, eyes, eyes, there's still value in what you built. There is value in the the product. There's value in the service that you are offering.

And what do we typically see in cases like that that company gets purchased by another company. I don't see a lot of the kind of merging happening in the blockchain space. It's still way too early for ai.

Although we're seeing that with, you know, uh, open AI snapping up some resources from other people and Microsoft making huge investments in open ai like blockchain didn't have these massive blockbuster consolidations. In fact, most of the time what happened was defi companies that went out of business, because it turns out they were fraudsters, allegedly fraudsters, I have to say, allegedly, um, they were getting snapped up for pennies on the dollar and the assets were were being converted because, well, other, it's that, or they go out of business. I mean, I'm not saying that anybody in the AI market space is a fraudster alleged or otherwise, but is if there's not any smoke there to produce fire, what happens to these companies?

Will, will someone take a chance on, well, it's a diamond in the rough and hey, what's, what's an extra a hundred million dollars or are we gonna let these companies start going fallow by the wayside? And it kind of proves out the market that there's really only room for one or two Nvidia out there. It it's that whole find the hype.

You know, the hype cycle is largely a build off of that. Um, and then what, what's left when it's gone? That's, that's kinda where we come.

Uh, it's gonna be, you know, we've got the AMDs and the Nvidia of the world, which it kind of did come down to you look at HPE buying Juniper. Why did they buy Juniper? They bought Juniper for Marvis.

They don't care about the rest of Juniper, which Marvis was a fairly late acquisition for Juniper. That's a whole identity gone right there, all in the name of one AI technology. Now we're gonna see a lot of that.

I think people are gonna buy the AI hype. Someone else is gonna wanna buy them because they bought the AI hype. Whether that's gonna work or not is gonna depend quite a bit.

Well, I don't think we've definitively answered whether or not AI is the next bubble to burst. But of course, if you asked me three years ago, I couldn't have answered the same thing about blockchain and going all the way back to the Dutch and those stupid tulip futures. The problem is not that the technology isn't progressing at an incredible pace, I am completely amazed with the capabilities that a lot of these AI platforms have developed.

My problem comes in the application phase, and that's what happened with blockchain. That's what happened with OpenFlow. I mean, go back in the networking space.

That's what happened with ATM lane. And if you're old enough to remember what that is, you probably need to schedule a colonoscopy. It's more about the fact that people misapply technology to solve problems that it shouldn't be applied for because they think that it's their golden ticket.

Is that an outgrowth of rampant, of ca of late stage capitalism? That's not that kind of podcast. That's for somebody else to answer.

All we can say here is that you need to be cautious as you look at the way that AI is being applied. You need to ask some very intelligent questions when someone tries to tell you that their product is the best because it has AI in it. And you probably need to get better at filling out those captions and Turing tests because if this turns into a bigger problem and we are actually in the beginning stages of trying to solve Arco Baus look that up.

Um, I for one, like to be useful to our new AI overlords. That'll just about do it for this episode of the Tech Field Day podcast. I wanna thank our guests for joining us.

Evan, if people wanna find out more about where to follow your, uh, musings and thoughts, where can they go to do that? io. Alright, well if you wanna follow the things that we do here at Tech Field a, make sure you follow us on your favorite social media platform.

We're at Tech Field a in most places. com. com/podcast.

Don't forget that you sub can subscribe to our YouTube channel. You can also subscribe to us in your favorite podcast catcher no matter how you choose to consume our content, whether it's video, audio, or text. Please make sure you leave us a rating and a review.

We would love to hear your thoughts and we'd love for others to hear your thoughts as well. We'll be back next week with another great episode. Until then, take care and, uh, do some AI development.

You never know what you might get.