Meta, Google and Addictive Algorithms | Cloud Native Nears 20 Million | Robot Umpires Reach MLB

Hey everybody, welcome to the Techstrong gang. We got everything today. A little bit of legal action involving big tech.

We got software development issues from Amsterdam, and we got baseball. Hey, spring is here. Baseball season's started.

Wouldn't be a gang if we didn't start talking about baseball. Want to thank our guests for joining us today. We have, of course, Tracy Reagan joining us.

Tracy, how you doing? Great. Glad to be here.

All right. John Schwartz is out in California. John, how are you doing, buddy?

I'm doing good. It's good to be here. It's good to see everybody.

And Jack Palmer, fresh off a red-eye from San Francisco, is back home in his normal environment. But Jack, good to see you. Good to see you all.

And then finally, we have Guy Currier. Guy, how you doing in the world? I'm doing all right.

No red-eyes here. All right. " And I don't know how this court case is going to play out.

There's probably going to be appeals to be coming through. But Guy, is this the first of many of these lawsuits in your mind? It depends on how long your time window is.

I have to compliment Mark Lanier, the attorney on behalf of the plaintiffs, for finally coming up with the right argument here, which is it's not about the content, it's about the algorithms. It's about the, let's say, the auto-curation, instead of just saying algorithms. Algorithm is a broadly used word.

So to answer your question, ultimately, I think yes, even in the United States. In the short term, I think that probably TikTok and Snap did the right thing and paid up and got out, because this is finally the core and the crux of the whole issue. So why?

Section 230 of the Internet Amazingness Act, or whatever it was called back in the 1990s- Oh, it's Communications Decency Act. Yeah Yeah. Yeah, that was my little joke.

But- Sorry ... Section 230 gave roughly the equivalent of carte blanche to the social platforms and other platforms hosting content not created by the staff directly of those companies. And that protective shield has only strengthened over the years as these tech giants have become giants and gained influence, Meta being one of them, Google being another, TikTok being a recent one, such that they're basically saying they're not held responsible in any way for content posted by users.

They get to make all kinds of money off of them, but they're not responsible. And that has turned out, quite obviously, to be a super reductive and commercially interested argument. In this case, it's not so much that people are posting content.

Actually, in all cases, across all these platforms, it's not that people are posting content, it's that the platforms are selecting and serving up certain content to users. With what goal? Is it to inform users?

Is it to educate them? Is it to delight them? Is it to do any of these things?

No, not really. The goal is simple, to keep the users glued to the given social platform. TikTok has probably been the most successful at this.

So the question resolved by the jury in this case was not was bad content or unhelpful or damaging content, at least for teens, posted at Meta and Google, originally also TikTok and Snap. That was not the question. The question posed to them was, did Meta and Google, for their own commercial interest, post stuff that was harmful?

And the jury agreed, yes. So I suppose there's going to be a lot of commentary about how $3 million is a drop in the bucket. It's not even a drop.

It's like a molecule in the bucket to Meta and Google. That's all true. But the implications of this cannot be more profound in the long term.

I am not just skeptical, I'm pessimistic that in the short term, say within the next few years, that even this particular judgment is going to stand. I think the power here is far too great on the side of the tech giants who have successfully prevented Section 230 from being updated to the modern, actual way these work. I'll just close with this.

It's obvious what's going on here. It's obvious to anybody who spends five minutes thinking about it objectively, which is that the commercial interests of the companies publishing this stuff run counter, in many cases, to the privacy and interests of the users. And that's wrong, and it has to be fixed.

John, what are they saying about this in the Valley? I think they're scared. I actually think this is a major issue.

And it's not just this case. There was a case in New Mexico where there was a $375 million judgment against Meta just a couple of days earlier. And both plaintiffs' attorneys followed basically the same legal playbook.

They looked at this kind of big tobacco similarity between what's happening with social media platforms and going back a century. And they basically proved, without hitting their heads against the wall with Section 230, that these companies are liable under product liability theories. These are defectively designed products that failed, and the companies failed to warn...

parents and children about the addictive nature of we want to call them algorithms, whatever the case may be. There are a lot of other cases actually in the pipeline, at least three of them that I know of, that are going to go perhaps to jury. There's one in Oakland set for this summer, among others.

And on top of that, there are about 20 states that have enacted laws to regulate social media use. I bring that up with the caveat that the White House and big tech is doing everything it can to eliminate those types of laws. But I think these companies, and particularly Meta, are worried.

And you know what really hurt them was this notion that a lot of employees, disgruntled employees within these companies, came forward and told us exactly what Meta and Google were doing. In particular, Meta. They were telling us that they were knowingly reaching out to kids who were barely teenagers, knowing that their product internally was damaging, but all in the pursuit of more usage, advertising, revenue.

So I do think so. And for months, other executives, and Meta has a lot of detractors out here. Salesforce's CEO, Marc Benioff, referred to them as the big tobacco company.

He has some skin in the game. But I do think it's significant. And another thing that I think is very significant is during this LA trial, Marc Benioff spoke-- Excuse me, Marc Benioff.

Mark Zuckerberg spoke, and I don't think he was a very effective witness. And for years, he's gone before congressional panels telling them, "We're going to self-regulate. " And I think people finally just got tired of it.

And I think there finally was a legal strategy that looked around the 230, which has been around for 30 years. So as Guy mentioned, Mark Lanier, who is the attorney in this case in LA, and other attorneys are probably going to follow this playbook. I think we're going to see more of this.

Hey, John, there's another aspect to this, and I'm glad you brought up the analogy to big tobacco. And I think what, to me, this shows is really the hubris and arrogance of Silicon Valley, where- Indeed ... not having, more importantly, not having looked at and ignored what happened in the big tobacco lawsuit cases, they committed the same sins all over again, which is not only did they do it, but they documented it in email.

" So he was completely aware of it. It wasn't like it was a side effect of-- There is an argument that could have been made, and I think they tried to make, that we're just trying to keep users engaged because we want to sell ads, and that's how you run an ad business. Like the newspapers, right?

They do the same thing. The old story about newspapers was if it bleeds, it leads. Bleeds, it leads.

Yeah. You have all the death and destruction on the front page of the newspaper to get readers. But Meta internally, from low-level product people all the way up to the CEO, were aware of not only what they were doing but the impact on their users.

And that arrogance of we understand exactly what's going on, we're documenting it, we don't care. Nobody will ever know. Right.

Until they had conscientious employees who came forward. And you're right. Yeah.

They refer to these as the smoking gun emails. It was plain. It was black and white.

" But this jury did a damn good job, I think, in just getting parts into the truth. So let me put this down here then. Would not another jury in a different time conclude that maybe, I don't know, that, well, the parents gave the kids the phone, and it was the parents' responsibility to stay on top of what the kids had on their phone and how they were using it?

And I don't know. Guy, I'll throw this back to you, but might there not be competing findings in a different case here? And then what?

Well, that's the thing. At a higher level, ultimately, this is a case or a situation that goes to the United States Supreme Court, ultimately. There's two elements of this.

That's one. The other one being the company's own reaction to this situation. They're going to hold this off as long as they possibly can and argue about their protections under Section 230 regardless.

So I think you're just making really good sense, common sense, Mike, and unfortunately, that just doesn't have a play here, at least not for a while. The one advantage that the public has, and the safety and health of the public has in this situation is that everyone's real suspicious of these tech giants. Everyone.

The mass of the public is. That wasn't true about tobacco. It took a lot longer for that to develop.

There was lots of messaging out there and a lot of confusion about the-- Science knew of the bad effects of smoking tobacco going back to the early 20th century. This has been a lot shorter of a buildup and quicker of a buildup. And all the negativity around AI, for example, despite all the use.

I think that's the main thing in its favor. Trying to blame parents? Sure.

They'll try to blame lots of things, lots of victims, or lots of ancillary They're going to throw everything they can against the wall to prevent this. And I think it'll largely work because at an elite level, there's too much interest in all the money being made. Mike, one thing that's interesting too is this week, which is a crazy week for news, there was a White House, this new AI committee of executives, big tech executives, of course, like Jensen, Mark Zuckerberg, what a coincidence, Lisa Su, I believe Michael Dell.

And their big push is against state regulations of any kind against AI. They want some sort of national standard, which basically means they're not going to do anything because there is no national standard. Yeah.

They call it empowering parents. So they believe what the solution is, is to protect children by empowering parents to have parental controls on these tools, on these games and platforms. But it's so shortsighted.

We can ask Congress to try to address this and start writing laws that says you have to build behavioral security and safe by design software, but I don't think that's going to happen anytime soon. So I think that the only path to solving a problem here are civil lawsuits. If we have our government saying, "Hey, you know what?

" Which means, well, okay, but you can't put parental controls on algorithms . And what do we do then, as parents, what do you do? Just take the video...

Because this is broad. This is going to cover big sections. We're talking about co-pilots and AI agents and adapted learning platforms, and any kind of gamified system.

It's broad. It's super broad. What this kind of lawsuit does, it kind of narrows Section 230 in a way that says we have to have safe by design engineering practices and behavioral safety is paramount.

And we're not having that discussion. Well, I guess we're going to have to see how all this plays out, and I'm going to move on, but I would just point out one thing. Early on, they had all those ads for kids or aimed at kids with Joe Camel, that camel with sunglasses smoking a cigarette.

So maybe somebody's going to start tagging Meta with the phrase, hey, Mark Meta. Fresh in your mind. Does that mean Joe Camel?

It could be the same thing. So one way or another- Can I just mention one thing, just really quick. Something that Jack mentioned about the arrogance and hubris within Facebook at the time.

Frances Hagen, who was one of the whistleblowers, I remember interviewing her, and she told me they knew about this. She would go to them, she would try to talk to them, try to influence them in any way, and she was always ignored, and this was the MO within that company. Sorry.

It's a bad look for big tech no matter how you sort it out from here. All right. Let's shift a gear here.

This week was the big giant KubeCon + CloudNativeCon conference in Europe, and some, let me think, I think they said it was 13,350 folks attended this conference to talk about the latest and greatest in cloud native software and where all that was going. And one of the things that keeps coming up is, well, just how many developers are there? A survey suggests that maybe we're closing in on 19,000 cloud native developers, but that's kind of a guesstimate based on some funny math as to how many- Thousand or million?

Sorry, million. Sorry. Versus how many folks are actually using particular projects, and other surveys suggest that people aren't using quite as many, and yet maybe folks are using software without realizing that it's cloud native because, well, it's all hidden behind abstraction layers.

Tracy, what's going on here? And I'll add a couple other things to this mix. " Well, first of all, I think that we can say that Kubernetes is now a baseline.

If they're looking at, I think the number was like 40% of companies or developers are using Kubernetes, that's a baseline. And developers shouldn't have to know what Kubernetes is doing, to be quite honest. It was years before I decided to look under the hood and start understanding.

I first started programming on the mainframe. I didn't care what ZOS was doing. But when OS2 came along, I was very curious and started looking at how OS2 managed memory.

And I think we saw a similar progression with Kubernetes. Developers were curious about Kubernetes. What can we do with Kubernetes?

How can we make it work better? But now it's a baseline, and I think that there were a couple of things interesting in what happened at KubeCon this year in Europe. But on this topic, we are seeing now that Kubernetes is a platform of choice, and platform engineers are going to be in demand.

We need more of them, and I think it's good news for people who are in DevOps and platform engineering that more developers, and it's all about the developer experience. It's DevX at this point. How easy are these tools to use?

Does the developer need to know anything about Kubernetes to write good code on it? Now, let's go to one of the other announcements that they made, and this is, I think they called it Cars. It was their kind of a standard for containers for AI runtime infrastructure.

And it looks like they're going to do two things that I think are pretty big. One was what they referred to asstable, like in-place pod resizing. So if you dug into Kubernetes, you realize that there are aspects of it that don't support AI very well.

And being able to change memory allocations without restarting a pod is really, really important, and they're going to be offering that. The second one was workload-aware scheduling, which means that in the past, one workflow would fail because the default scheduler wasn't really designed to keep track of the topology. So workflow-aware scheduling means that we're going to get rid of the deadlocks that we've been seeing.

Now why is it important for AI? It's really important for AI infrastructure that these particular problems with Kubernetes are addressed. And I don't remember what they called it.

It was Kubernetes AI something. It was like cars or something like that. I don't remember now.

When I read the article, I can't remember what the R stood for. But the two areas that got me happy was the pod resizing and the workflow scheduling. So I think what we're seeing is that Kubernetes now is a baseline.

Developers should not have to understand it to write applications for it. Platform engineering is going to continue to be a very big area and probably an area where there's going to be demand. So developers may be moving from being developers to platform engineers to support these platforms, and Kubernetes is getting ready for AI.

All of which are good. Guy, Jonathan Bryce over at the CNCF believes, to Tracy's point, that it won't be too long now maybe before the total AI workloads running on Kubernetes will exceed all other workloads combined, and basically he's looking at it and saying AI workloads, especially inference workloads, are the killer app for Kubernetes. What do you think?

Well, I think that when you have a hammer... Let me finish that sentence for you. So there's tremendous pressure to run AI inference and AI agents and AI services on Kubernetes, on containers, cloud natively.

They are not suited whatsoever, Tracy, apologies, it's getting there, but Kubernetes is not suited whatsoever for this. A platform designed for microservices and service-based overall infrastructure and statelessness and all that other sort of stuff now finds itself on many fronts trying to handle stateful, trying to handle huge amounts of memory. You didn't even mention the KV cache, which is critical for performance in inference.

Concurrency. Shoot, maybe we should start calling it something like a virtual machine or something like that. Maybe that's what we need in order to share resources in a compliant and continuity-laden way or whatever.

My words are tripping over themselves. Snark aside, I've been doing a lot of research and investigation here into where does, as a model, as a philosophy, where does containers for cloud native and virtual machine overlap, if at all? Because I just felt, and that's not reasoning, it's feeling, that we're accomplishing two different things.

How does this relate to this? So I think Bryce is correct, if I slightly restate what he's saying, which is that AI is going to drive development of Kubernetes so that it stops trying to impose quite so much on the architecture of applications and starts accommodating both many types of architecture and needs in the better, that is to say, in the applications on the one hand, and on the other hand, a much better developer experience, and to which I would add data scientist experience and so forth. Let me just qualify all this by saying I'm really talking about inference.

When it comes to training, as the boundaries of frontier models keep getting pushed, that is practically a bare metal scenario. It's not really, but you kind of have to think of it that way. That's a totally different nail for the hammer to try and bang in.

I don't disagree with you, and in a lot of cases, it feels a little bit like the proverbial round peg in a square hole kind of thing, but it looks like they are trying to figure out how to extend Kubernetes. But Jack, does this just feel like there isn't another choice out there anyway, so by default, we're just going with Kubernetes because that's what's at hand? Yes and no.

I think we've spent a lot of time talking about a very specific use case, and I know this is going to be very hard for all of us to do, but if you ignore for the moment the world of AI. Now realize I just spent four days at RSAC where we were all talking about security, but the only thing we were talking about was securing AI, right? Now over the last three or four years, of the 600, 700 vendors and 40,000 people that attended every year, all the vendors were talking about how they were moving their on-premise stuff to the cloud, and it was cloud data, but it was Kubernetes-based and some form or flavor of container-based, right?

There's a couple different things in there that doesn't really matter, and that's ignoring all of AI. So Guy made the allusion earlier that 10 years ago, the de facto way we built our infrastructure was based on virtualization And in fact, we still use a tremendous amount of virtualization in the world, and it's not uncommon to walk into- We're not going to stop. We're not even going to start anytime soon.

We're not going to stop. Absolutely. " Right?

And so the reality is these things are tools, and you apply the right tool for the job. And if Kubernetes is the right tool for AI inferencing, apply it. If it's not, go find or build the right tool for the job.

I will say one thing- And I think that's part of the discussion here. They're starting- Right ... to build requirements, which was the R.

Kubernetes AI requirements for what we need to do to take Kubernetes to that level. I guess I get excited about any new major change in platform engineering and that- Well- ... and Kubernetes was big, cloud native has been big.

Oh, yeah. And it was never going to go away. And let's not leave out edge and edge-based inference.

K0s at least is almost a requirement for these types of implementations, or at least the more interesting of them right now. My assertion is that containers are absolutely fantastic for mobility and for highly distributed architectures. And that's going to hold true for AI as well, no question.

So, and if I just might point out that both Guy and I have participated in Tech Field Day, AI Infrastructure Field Day. I just recently did a Cloud Field Day. And over the past six months, we've had a couple of different vendors show up and basically say, "We're introducing another caching layer," because somewhere in the environment, they didn't build caching in in the first place.

And so now we need to do that and add that in. And whether that's inside the container environment or inside the virtualization environment, people are realizing that we built something, and we missed the fact that we actually need another cache layer in the middle there to go from super fast memory to regular memory to NVMe-attached SSDs, to finally slow stuff. And somewhere in there we missed something, and so now we're sort of shoehorning caching back into it.

And I think now what we're seeing is with Kubernetes and some of the other environments, we also have to put in some other features that we dropped along the way for simplicity. Now we find we really need them. I think Jack has it spot on.

That's exactly what's going on here. And I would just point out, it's taken us 10 years to get Kubernetes to the point where it runs a database reasonably well. So you got to figure maybe halftime with the AI is going to be, what, five more years of development work on the requirements?

We'll see, but maybe they'll get it faster, but I don't think any of this stuff is going to happen overnight, that's for sure. I'm going to shift a gear to a lighter subject. Baseball is back.

Opening Day has come along, and there's been a lot of different games, and the one that of course leapt out to me was that Yankees-Giants game, John. Yep. My Yankees, looking pretty good on that first game anyway, but we'll see how it goes from there.

But the more important thing in my mind was this was the first time in the Majors, during a game, where somebody appealed a call, and they had to go to the robotic service to see whether or not it was an actual strike or not. Did the crowd harp? Did they know about it at the time, or is this something we're- Oh, yeah ...

going to see more of? They've been anticipating, yeah. You and I, and baseball fans, have been anticipating this moment for several years because the automated ball-strike system has been tested in the minor leagues for a couple of years, and the first instance was this year.

This is the first year officially where the ABS system is in place. And it was the top of the fourth, and now forgive me if I mispronounce the New York Yankees shortstop because I'm not familiar with him. Is it Jose Caballero?

Correct. Yeah. Yes.

He was up, and he disputed a strike call. It went to the robo-ump. Robo-ump actually upheld the human ump.

So the human ump, as infallible as that strike zone usually has been to the annoyance of many baseball fans over the years, it was upheld. But it's interesting- So does that show that the human ump was a good ump, or does that show that the robo-ump- Yes ... is trained to curry favor in order to gain credibility?

Oh, yeah. Well- That's what I want to know ... Guy, yes.

Perhaps you're right. No, but that's interesting you say that, though, because, I think Mike will agree with this, in baseball, there are umps who specialize in calling balls and strikes who are considered the best, and there are others who have varying strike zones. So they have their high pitch or low pitch or inside on the corners, and it was maddening and frustrating to watch, especially if you're a hitter.

The WBC, the US eliminated the Dominican Republic with a pitch that was clearly not a strike. It was called a strike, and they ended up winning the game. But this is part of this whole kind of, in a bigger picture, the technological advancement of baseball as it tries to curry favor with younger viewers.

We've shortened the game. There is an element of replay. At this game, it was streamed on Netflix.

For the first time, Netflix televised a game. There were a lot of drones in the pregame. There was some autonomous vehicles that were showed off.

It's basically trying to pretty up baseball. And I guess one of the elements of this system, and the system in part is based on this kind of system that's also used in soccer called the Semi-Automated Offside Technology, SAOT, I had to look at that. And that looks at offsides and other types of things, goals, whether a ball's clearly over the line for a goal But now there are subtle rules at play here.

So, as it turns out, baseball's been testing this stuff in the minors for a few years now, and the way the majors have decided to implement it is that only the pitcher or the catcher or the batter can make the appeal. Right. The manager cannot do this from the dugout, but I imagine that eventually, the manager will have a sign telling the batter to appeal.

But, and then as they look through the records for the last few years using this, it looks like the catchers do a better job of understanding- Yes ... when to appeal. The pitchers are probably the next best, and the worst is the batter, who probably didn't see the pitch in the first place, because otherwise they probably would've swung and hit it, right?

So, okay. You said something really interesting about the catcher, because the idea was that this was going to minimize the human umpire, and it was going to minimize the ability of catchers who are very good at framing pitches. I think it's just going to be the opposite.

I think actually, the catchers who are really good at framing pitches are going to know exactly how close the ball is, and they have better eyesight than the umpires. So... Well, but hang on a second, Mike.

Let's be fair, there's a little bit of statistics that you're ignoring there, okay? And baseball is, if nothing, it is all about statistics. Hey, Tracy, let's leave.

Let's go get a drink. This is... My God.

I'm not interested in shooting myself. No, go ahead, Mike. Go ahead, Jack.

I'm just wondering when the umpire's going to be gone, and it's just going to be a robot back there making the calls. Well, look- What I want is human umpires and robot players. But anyway, go ahead, Jack.

I know two things about baseball, and I know just a little something about statistics, which is if your job is to sit there and throw every pitch, or more importantly, catch every single pitch, versus your job is to hit once every inning or twice every inning, you see a lot more pitches, and therefore, you have a lot more experience judging them. Mm-hmm. So it's very obvious that the catchers and the pitchers will be much more accurate than the hitters.

The umpires, however, are sitting there watching every pitch just like the catcher is, and therefore, they should be hopefully as accurate. But if I recall correctly, two, three years ago, sometime in the very recent past, Major League Baseball brought in a pitch clock to speed up the game, and I think, John, you mentioned this. Correct.

Right? And one of the reasons is it's really hard to get people to go into a stadium because you have day games in the middle of the week, and most people are working and don't want to show up for a noon game to buy a $12 hot dog and a $15 beer and watch a four-hour baseball game or a six-hour baseball game anymore. So you've got to cater to the TV audience, and you got to keep them watching the ads, so you have to have time places for ads, just like football does, and just like basketball and hockey do.

And I think this is all part and parcel of putting a rhythm to the game and taking out the long arguments of this and that and the other. Thee- But that's just my naive, non-baseball fanatic view of the world. I think that makes it.

A couple of things about that. Look, first of all, the batter is lucky to get up four times a game, depending on how many opportunities there are, so your statistics- Or unlucky, as the case may be, yeah ... yeah, it has to be.

But in defense of the afternoon weekday game, it's also known as the businessman's holiday game because everybody there is bringing their smartphone and is basically sitting there doing their email and whatever else they can get going remotely while they're watching the game. So when you go to those midday games, it's mostly guys working white-collar jobs who are just kind of hanging out for the day, doing their thing. Not that I know anything about that, but I do.

No. Absolutely. I've never been to a hockey game.

So do you know what I find interesting about this whole thing, other than talking about statistics? Thanks for that, Jack. Mm-hmm.

5 came out in November 2022, okay? So how long before that were you looking at studying and learning about AI? All of us were doing it for years before that, right?

Yeah. Okay. Anybody remember this really old-fashioned AI term, explainability?

Mm-hmm. Yes. I remember feeling so encouraged when explainability started.

" Everybody needs to see where these AI results came from. If there's a better example of the need for human involvement and human oversight of AI that's also fun, I can't think of one. Because an AI that sits there and says, "Look, I'm using my computer vision," personifying the AI, "I'm using my computer vision to show you where the strike zone is, okay?

" And then a human can go in and say, "Wait a minute, AI, you screwed up. The ump was right," or, "The ump was wrong," or whatever. That I could see.

But instead, it just becomes this new, far more impenetrable black box than the one you were describing, Mike, the ump that always has a low strike zone or a high strike zone or a shifting strike zone or whatever. So I have nothing against this ABS system as a concept, but like freaking everything in AI, we're just presented with some thingamabob that produces something that's designed to look correct, whether it is or not, and everybody just goes, "Oh, great. Great.

" So... So can I make two- What they're saying, guys, the ump actually should have glasses that actually help them make the calls, right? Oh, right.

Yeah. That's something Amazon might need to go crack a few eggs and show to you. There's a tool that allows the ump to explain his call.

In front of his eyes. Let's re-engineer the entire thing right here, right now. I am a fan of a standard strike zone, but the standard strike zone is supposed to be based- There is none ...

on the height of the player. There is none. And that's hard to assess, especially when each player has a different stance at the plate.

So, somebody who's 6'2" is going to look crap from the outside line, right? Yes. Sorry, Mike.

One of the things that players, have you heard this, that they're listing their height as shorter because they think it will work to their advantage in terms of how the box interprets whether it pitches a ball or a strike? So they're saying they're shorter because it plays better to their advantage. But I was going to mention, Jack, there are two things.

You talked about stats. They were keeping stats in spring training about which teams were most effective at challenging balls and strikes. So yes, when you talk about baseball being a game of statistics, it reaches that granular level.

I don't know. Could it end up being that baseball is just a game played by people who are roughly 5'8" to 5'10" versus all those 6 foot guys that have bigger strike zones, right? Who knows?

Mm-hmm. Yeah. And also the irony is that- How do you assess a strike zone on a robot?

That's what we're going to have to figure out. Oh, God. But the irony is that by installing this system, it might kind of counteract the pitch clock.

I mean, are there going to be challenges? We don't know when they're going to occur. It's a work in progress.

They're probably going to be backloaded when the game's really closer to being on the line. So you're going to see more challenges probably seventh inning on, I would think. Mm-hmm.

Hey, Mike, can I ask you a question? Is this maybe just a gimmick? MLB wants to get in on the AI excitement and action?

Well, the minute it says sponsored by so and so, yeah, then it'll be a gimmick. I'm sure they already have that planned, so- Let your inner cynic out, Mike ... you've answered my question.

Didn't we have similar discussions when instant replay came around? I bet we did. I imagine we did.

Now they claim that humans know how to have these meetings and make these calls without lengthening the game any further than they have, but that remains to be seen in my mind. That remains to be seen. Well, I know Alan was a big proponent of instant replay on penalties like pass interference in the NFL, and you're probably going to see that happen, but I'm all for that.

NFL replay, I'm all for Jim. Being a naive person or a naive baseball person because I don't really pay attention to the game much at all. I get it through osmosis with friends or talking to Mike and other people, but I imagine part of this is to bring more consistency to the game.

Yes. Now whether that's an improvement or not an improvement to the game, it can be, as I know in other arenas, when you have lack of consistency in applying rules and judging of rules, that can be extremely frustrating both for the participants of the game and for the audience watching the game is- So- ... why was that call there, right?

Right. And in fact, John, you brought that up of another game, a critical game that was lost because a call went wrong, right? So I'm going to bring that.

So Mike will agree with me on this one. 1997, Game 7 of the World Series, there was a guy named Livan Hernandez who pitched for the Florida Marlins. His strike zone was gifted to him by an umpire named Eric Gregg, who since I think he's passed away.

There are replays where balls are clearly six to seven inches outside the strike zone, which is a big, big difference. And that influenced the game. They actually won a championship because of that.

And there are other instances, but that was the ball strike controversy. And here's another thing. There's a guy who's one of the color commentators for the Giants, and before each game, he talks about the umpire and their strike zone.

" It's like, wait, shouldn't there just be a uniform strike zone? Right. Also, remember, there are north of 160 games in a year, so humans, I can have a great day when it's 75 degrees out there, but come August and it's 105 in the shade and I'm trying to call a game, maybe I'm not on my full balls and strike calling capability, so.

Mike, I'll just point out that nine years later, John knows which game, which player, which inning, which ball was the issue at hand. Nine years later, this was so important. That inconsistency still sticks in his craw.

That's true. He's wearing Google Contacts and asking the AI. That's- Right.

You know who's actually, I was going to say Barry Bonds was doing the color on this game. Now I have a lot of mixed emotions about Barry Bonds, but you talk about, Jack, you talk about people who remember balls and strike counts from eight years ago. He has like a computer for a mind when it comes to that at least.

Well, every sport complains about the refs, right? Every single sport. True.

" Your family- So consistency in the game is important. Well, and Tracy, as I said, I come from a different world, the world of motorsports, internal combustion engines, and in that world, pushing boundaries, the rules, and seeing not just how far you can push the boundaries, but how big of a leap over the boundaries you can go is, that's part of the sport. That's what makes it fun, so.

Right. All right. Folks, we got to wrap it up there.

It's Friday. We're going to go a little short and give you guys some extra time back to yourselves. But, hey, I love baseball.

I'll still be going to the games, and no, I'm not really interested in watching robots run around and kill each other on a baseball field. That's just not going to get there for me. Hey, guys.

Thanks for everybody being on the show, and thank you all for watching. Please stay tuned for the lineup of the Techstrong TV replay, and we'll see you guys all Monday.