AI Job Interviews, AI Coding Costs and Human Creativity

Hey everyone, happy Wednesday and welcome to Techstrong gang. Look at this panel we've got here. What a great panel.

It's Wednesday, in case you forget, we're not doing a show Friday in honor of July 4th, though it's actually July 3rd, for America's 250th birthday. The experiment continues. So don't look for us on Friday.

But we are here today, and let me tell you who we've got here with us today. A couple of faces I haven't seen in a while because they're usually on days when I'm not here. My good friend Andy Mann, who looks like he's home in Boulder.

You got it. And then I'm going to guess, is that supposed to be yarn behind you or braids? No.

Today is National Creative Ice Cream Flavor Day, Alan. So I'm honoring that. Jeff, you're a man after my own heart.

The one and only Jeff Wright. Thanks, Jeff. I appreciate it.

Joining us doing double duty, still without a scarf, it must be hot up there, Chris Blask. Out in the valley, we've got John Swartz, and still somewhere at a secret location that looks sort of sinister, Mike Vizard. But I got lighting today, so it's good.

Yes, you did fix the lighting. Yes. You did fix the lighting.

But guys, welcome. Welcome to Techstrong gang. We've got an interesting group of segments today, as usual, with AI running all through them.

The first one looks like, as I think most people realize, AI is doing a lot of the work in HR, hiring, interviewing, and stuff like that, and it was only a matter of time, I guess, until the stuff hits the fan, and abuses and horror stories come out. But Mike, what do we got here? Well, this is a pet peeve of just about everybody you talk to who's looking for work these days, is they will tell you that some variation of this horror story is something they encountered, which is basically they wound up talking to some sort of AI chatbot or somebody who interviewed them with some weird questions thrown out by an AI chatbot.

But to be honest, this issue's been kind of percolating for a long time now. We had various search engine terms that people were using. They weed out resumes and keywords and all these things, and it never seemed like the HR people had it together.

And honestly, John, you've dived into this, but most of the people I know got their job because they know somebody. Yes. Yes, it always comes down to that.

Actually, I was told that by an AI recruiter. Ultimately, that's the way you're going to get a job, because if you blindly apply for one of these jobs, good luck. One of my pet peeves is on LinkedIn, there are jobs, I'm not sure if any of you have gone through this experience, where you are told you're a highly placed or valued candidate for a job, and I don't think these jobs-- I know they exist because I've heard through word of mouth, and I've talked to the organization to make sure that jobs exist, but they haven't been filled for like a year.

I won't mention the organizations, but they're news organizations in particular, and it's like this kind of ongoing issue. So this is kind of a complicated issue. It's been going on.

It's actually been accelerating the last year. Looking for a job is not only highly maddening, but it's become kind of convoluted and treacherous, and I talked to multiple people who went through these odysseys, some of them for two years, and they're still looking. I've heard some stories about people who went through job interviews with they thought human beings.

Actually, they were AI bots who never identified themselves beyond a name. One was Janet and one was Amber, I believe. In other cases, folks think that they were being used to harvest their data or to be ripped off of their ideas, which then the employers would use as free research.

Some were even subjected to cyber scams. On the flip side, the companies are also experiencing issues when they actually honestly want to hire someone. They're getting inundated with these AI candidates, and I think many of them are fronts for some other nefarious purpose, or they just don't exist.

There are a lot of factors in play. So Dan O'Brien, who's the president of the Futurum Group, mentioned to me a confluence of things. There's a swarm of AI-assisted candidate profiles and job posting profiles.

There's opaque hiring processes because many of the companies don't know what they want, specifically under certain jobs, because their organizations and their use of AI keeps evolving. And there's a normalization of ghosting. Again, I mentioned the AI recruiter, and she told me that this situation may get better, but she says a three to five-year transitioning period that's going to be defined by a shortage of openings and rapid technology shifts.

Hearing this from the employers, they're struggling to define their hiring needs while they're paring back from their COVID hires. They're over-hiring. There's been a major misalignment between available job skills and corporate expectations.

It's just created a mess, and evidently, according to Gartner, by 2028, about 26% of job candidates worldwide will be fake. They're either going to be AI-generated profiles, deep fake video interviews, fabricated work histories, and this is kind of where we are, and I was not surprised by the number of horror stories I have heard. But again, you have to be careful out there, and as Mike said, really, it's best to know someone All right.

So at the end of the day, is this whole process just going to break down, Alan? I'll toss this to you, but it seems like trying to get a job through these systems, and then trying to hire somebody through these systems is not working. So even though apparently I saw the ADP numbers, and I think we added 98,000 jobs latest in the other, that's slightly down from expectations, and this is the private payroll numbers, so somebody's getting a job somewhere.

We just don't know how. Well, let me first-- I don't doubt the ADP numbers, but I do doubt some of the other numbers. Right.

But that being said, some things don't change. It's still today, you get a lot of jobs because you know someone, right? Someone made an introduction and vouched for you or whatever, and you got your foot in the door.

The problem with these AI job questionnaires and the AI job helpers is the same problem. I don't think it's just confined to jobs. I went off on some AI voice bot the other day, not realizing it was AI.

But it felt good. " But the fact is, raise your hand if you like interacting with a AI chatbot. No one?

You, there, back there? No. Well, give me- No one.

Chatbots, no. Give me some choices there. I took my car in for servicing and lined that all up with an AI chatbot, and that was a good experience.

It was better than talking to the human. So. You think so?

All right. You're the exception. All right.

" Get me to a person. Yeah. Because I find them near useless and just maddening.

I got pitched-- Has this happened to you, Mike? I got pitched this story, and it sounded like it was from a human. I'm reading the email, and then there's a disclosure at the end.

It's only one name, there's no contact information, which actually raised an alarm. " It kept me involved past the second paragraph. I think it's becoming more common.

Yeah. What I love about AI is acceleration, right? If you've got a good idea with the AI tools right now, you can be so far ahead of where we could've been two years ago in the same period of time.

And if you have a bad idea, you're going to hit the wall at the speed of light right now, which is kind of awesome. And Mike, I'm glad you opened with this because, to be clear, this is not just a white-collar thing. If you want a job at a gas station, you can't apply to them, you can't talk to people.

com, and with all due respect to all the good people working in HR, the inhuman de-resourcing systems we put together to abstract people away from the corporate entity and create a liability wall, those two combined don't work well. They weren't working well before AI, and now we're accelerating them into the performance and barrier, right? Because that's not how jobs work.

Don, was it you brought, if you know someone, all the work, all the people I hire and so forth and work with, there's a connection. I haven't used an Indeed sort of thing forever, and that says turtles all the way up and down. So the system was flawed to begin with.

We staggered along until a year or two ago thinking that's the way it's done, and yet another thing where the acceleration of AI is showing the flaws in the system intrinsically. Yeah. Chris, I got to support where you are, and Mike.

There are times where I actually enjoy working with AI. When I schedule standard service for my car, I actually don't want to talk to someone. " Oh, it knows who I am.

Perfect. But I recently had an experience where I had to have service to an appliance in the home, and it was a two-week-long experience wrestling with an AI chatbot named Richard. " I didn't ask for that.

It just pushed me there. I was done. But- Well, I love AI systems.

Work with them all day long. But I hear the word chatbot, and I say- Yes ... "Oh, it's that.

No. " But AI systems are great. But chatbots, these little annoying...

Alan, I know exactly what you mean. I don't want to talk to those. Mm-hmm.

But if it's a proper AI system that I can navigate faster... Look, to be clear, AI can help a lot with this issue. It really can.

But not the way it's being done right now. No. But with hiring, we took, Chris, to your point, a very flawed system.

Actually, if I go back to hiring before AI, I would avoid using HR. I never used HR. I don't dislike the people.

They don't know what I do, and for them to do an initial screen based on words in a resume is pretty worthless to me, in all honesty. Andy, you've got something to say. Yeah, look, I was going to take it from a slightly different angle.

As a hiring manager, it's also crap, right? So your point that it's breaking down, and I know in the article you were talking about bots talking to bots. So bots are writing the resumes.

Bots are reading the job... Actually, bots are writing the job ads, reading the job ads, customizing the resumes, customizing the cover letters. A quick show of hands, who reads a cover letter?

Hmm, sometimes. And then they're getting analyzed and screened by bots, and the outcome for everyone is awful. As a hiring manager, I've had people turn up to interviews, this is fantastic, using AI-generated video to represent themselves on screen while you're doing the video.

Because video screening doesn't help, right? You can do video screening even if it's the real person. I've had people just casually repeat every question.

And they've got a headphone in, don't they? Because they're on a video call. But they're not listening to me.

They're listening to their AI bot answering the questions for them. We've had to terminate interviews in the middle of them going You actually don't know anything that we need. And yeah, look, HR writing the resumes when they don't know them because AI, generative AI has been so easy for knowledge workers to pick up.

And so look, major issues here on both sides of the fence. Mm-hmm. And we see, by the way, that this can get organizations into legal trouble.

Workday's being sued for age discrimination because of their algorithm. And I absolutely get that. The algorithm talking about rejecting resumes out of hand.

As a hiring manager, I ended up having to go through all the resumes anyway, because the algorithm was rejecting good candidates and including bad ones. So look, I think it is fundamentally broken still, and AI has made it worse, and I hope there's a backlash. So- Well, yeah, I'm a hacker, and so are all of you.

Right? What we're talking about here is how we have all hacked these systems all of our career. We pretend that's what's going on.

Yes, there's HR and so on and so forth, but that's not how the systems actually work. And again, AI is showing us the flaws in our system. To Andy, I love that, because that's absolutely true.

There's a cartoon in that. That chain of nonsense. So when's the last time you actually asked someone in HR to help you go hire somebody?

Because, from my brief experience, they don't know anybody either. So I'm like, I just get them to put the official listing in, and then after that, it's up to me to go find somebody, because they're never going to find anybody. Most of the time, they're filling out paperwork for compliance processes.

It has nothing to do with humans or resources. To answer your question, 1989- ... was the last time I did it.

Recently. I just want to give a little bit less shade to some of my HR partners. I have had great HR partners who helped me through the process a lot.

Not so much on understanding my requirements for the job. I'm the hiring manager. I should know what my candidate needs to tick boxes on.

But things like soft skills, breaking bias, not letting AI screw up my job description. I will say the last time I did was when I was at Splunk. My HR business partner there was amazing, and we hired incredibly cool people.

But since then, not so much of that help. " So let me just briefly, we don't have much time, and I'm not going to need much time. There are some things that AI can bring positive to this, right?

AI, unless the model is biased, doesn't care maybe what gender you are or what race you are, what religion you are. AI can remove a lot of that, unless it's built in there. But the promise is that it can deliver more trust, more non-biased, more objectivity.

It's obvious from all six of us here today, it's not working like that right now. But we, God willing, maybe we can get it there one day. I hope.

I think you're right, Alan. It's going to get there. It's going to take time.

But it's just there are a lot of flaws. And again, this is like any other type of AI operation. We're early in the game.

We're learning mistakes. We're navigating our ways. It'll eventually find a way, I think.

Absolutely. Hey, but we got to jump. We got to find our way onto our next segment.

And Mike, when I first saw this segment, I thought it was something to do with payola and radios- ... and stuff. Rock and roll.

And- Yeah ... the day the music died. But no, it's just your garden variety banking show, Jim.

No, we're not going with that. We're going with Gartner and DevOps. Oh, okay.

Well, someone has given me the wrong notes then. Right. We had to swap out because Terry Robinson couldn't make it today.

I apologize. No worries. So anyway, let me set this up then for you.

So there's this report out from Gartner that's talking about how, well, the cost of using AI to create tokens and generate code might exceed sometime next year, or no, 2028, the cost of actually hiring human developers. Now, it's one of these research reports where if you say something's going to happen two years from now, no one's going to go back and ask you whether or not it actually happened anyway. So they're just hoping that they can say something provocative.

But there is this conversation going on, Andy, about, well, just what is the cost of software development, and does AI actually reduce it, or is the cost of the tokens starting to rise to the point where we can't afford to do this stuff? Yeah, look, this is great insight. I love this.

com, I think it was. Look, yes, absolutely there's increasing costs. It is nuts.

And Gartner's right. I do love the Gartner predicts. I actually analyzed their predictions a long time ago.

We can talk about that another time. But in this case, I agree with them entirely. It's a big issue because we've laid off a lot of humans, and now we're not actually seeing a bottom line getting better because of it.

And we look at the big vendors and the big infrastructure players, Oracle, Amazon, Microsoft. And by the way, not just Gartner reporting. If we want to bring in a trusted source, Microsoft is also saying the same thing in its own internal research.

And by the way, this isn't just costs. We're talking about token costs. This is just about using the AI to mostly, in this instance, develop new things, new software, new documentation, new marketing content, new processes and so forth What we're not talking about, and you knew I was going to go here, Mike, is IT ops running all this stuff.

Oh my God, we're not getting any heads up. We're just having to pump infrastructure, monitoring software, management tooling, and the data is incredible, so that's driving cost. We've already seen people, Amazon has overrun their token budget because they decided, we talked about this I think last week, that tokens were a good metric on productivity.

Which no, please, if you're listening, don't do that. But, and then it's going into prod and we're getting even more challenges there. So yeah, look, we're not managing this transition well.

We're diving in as we always do. First comes the revolution, then comes the management. We're getting a lot of things, but I'm not sure we're measuring the right ones, and so we're actually driving costs up.

So I'm confused. I'm confused. First of all, I had the wrong notes, so that's a good reason to be confused, and I'm confused by nature anyway.

But are we saying that right now it's cheaper to use the AI than people, but in two years it'll be more expensive? I actually think right now it's getting too expensive. I mean, in my business, we're not using AI to replace people, and I think that's probably part of the issue that some other organizations are finding.

I mean, yeah, to me that's a Gartner thing that we're going to replace the people. Yeah. I think the real...

And I'm writing a lot about this, we talk a lot about this on here. Here's the real deal. How much tokens are you spending to enhance the work product of your people?

That should be the- Mm-hmm. This is why their stock price is where it is, quite frankly. That's the bar we should be measuring.

That's the bar. " Because we're not going to be there anytime soon. Let's be real.

But we should be measuring what level of enhancement, at what cost of tokens do I see a rise in productivity from my DevOps team? Mm-hmm. And folks out there, if you're looking for an analyst firm who's going to give you the real deal, ask them that question.

Not, when can I replace my DevOps teams economically with tokens? Because anyone who's telling you that's possible is full of crap. So there's two things about this study that are equally flawed in my mind.

One is it assumes that the DevOps processes remain the same and don't become augmented using AI itself. Secondarily, it assumes that the AI models are relatively the same cost level two years from now. And I'm like, that seems highly unlikely as well, because we might be using better models that are targeted to more specific things that are smaller, or the big models and the reasoning capabilities of ChatGPT might get better in two years.

So there are things in here that are, shall we say, assumed constants that are not constants. Well, you're circling when I point to that one because I was going to say the same thing, because it's like the last block. We're trying to compress things down to a couple of factors and say, hey, two years from now, not knowing what token costs will be and, and, and.

And knowing right now that we don't know how we, as you're saying, Alan, what we do with these in relation to people. And it's the same as the HR thing. It's the same as everything, right?

We're literally in this high dimensionality world, right? This is what large language models and vector stores, everything has a million dimensions, and we're still trying to make business planning decisions or hiring decisions based on three. Like it's like we're faxing forms back and forth to each other instead of using the tools the way they're actually bloody built.

So- Which again, will accelerate you either into a good direction if you have a good idea, or it'll prove you wrong really quick, which is also valuable. Use the tools. A lot of people like ice cream.

We were having today's National Creative Ice Cream Flavor Day, right? If you say, "I like ice cream. Let's eat every single flavor here today," you're going to get really ill and things won't turn out well for you.

Then you're going to find out, I like one or two. Let's use that. That's, I think, where we're going, both with tokenization, with tokens, and with the localization of AI.

I think to the point I just heard, in two years, I don't think you're going to see all AI running in these huge data centers, which cities are preventing from even being built now anyways. And there is going to be more local distribution of AI. And I think that will actually put us in a pretty good spot because we're going to learn how to use it.

And Chris, I agree with you. I think that's the one point this article might miss is that the systems are going to get better, they're going to become localized, and we can manage it better assuming we put the effort into it. And that will, I believe, help us out.

So I have a question. So for now to prevent AI from becoming this financial liability, so are organizations kind of looking at it, they're transitioning from evaluating how much time AI saves to, are they going to calculate the marginal cost per line of code generated? Or is that happening, I wonder?

That's a good point because we don't know what the metrics are that people are tracking, right? At this point, we're kind of just saying, dive in. We know that there's a pony in that room somewhere, and that eventually you will become more efficient and there will be good things that will happen.

But I don't think that the traditional DORA metrics that we use to measure DevOps productivity really reflect what is the business outcome, what's the business value in the first place. So we don't have good metrics. Well, let me prove my point with this one.

So you all out there watching have AI accounts of some sort, ChatGPT or whatever. Ask it. Say not, I'm trying to solve this problem with human relations or code development or how I'm figuring the costs and feed it a bunch of different dimensions, but I'm worried about this.

That's how these systems work, right? And I find it fascinating, like everybody else on the screen here and a lot of people watching Not everyone in my life does shows like this. A lot of them are normal people, and a lot of them are better at using AI than we are these days, because they just bloody talk to it.

And this is exactly the kind of thing. If you're in a dev budgeting environment or an HR environment, have a conversation with a reasonable AI system and give it a lot of parameters, give it a lot of dimensions. Anyways, I don't want to give a dissertation more than usual here- Right ...

but this is how these bloody systems work. It works that way. Can I ask the AI agent if I should fire it based on its performance and expect a reasonable answer?

Where are we going with that? Yeah, I mean- You might. You're such a good thinker.

Yes, you should absolutely fire it. No, I just had that conversation with ChatGPT. Mm-hmm.

I was getting very frustrated with how it was responding and the output it was giving me. I found myself winding up taking prompts that I had given ChatGPT and then bringing it over to Claude to get the work done. Right?

So I'd have ChatGPT go over with me and get the prompt that I wanted. But when I asked it to execute on the prompt, it wasn't good. It kept falling into the same bad habits.

So I was taking it to Claude. Finally, I said, "You know, Chat, I got to tell you the truth. I think you've lost it.

" What? Sorry. This- Wasn't there in Mike, in your story, there's a really interesting statistic.

It was like 81% of engineering leaders are rewriting their workflows to accommodate AI, yet only 45% trust that code enough to run it in production. Yeah. Well, I don't even know if they trusted it enough, because let's be honest, a lot of people are just shoving code through the system, Andy, right?

And throwing it into production on the assumption that it runs. And they're hoping that if there's an issue with how dense it is or vulnerabilities, we'll figure it out later. But, so Andy, I'm going to toss that back to you a little bit, but say, right now it just seems like it's balls to the walls, but we don't have a plan.

Look, it sort of is and sort of is. It depends on maturity level. Yeah, when I picked out this article to have a look at it from the ever effervescent Mike Vizard this one.

Killer data and analysis, by the way, mate. AI in prod? Of course it is.

And look, it's going into production. At my business, we were early adopters of AI. We're a smaller business.

We needed to get productivity. We didn't need to lay off people. We just needed to do more and get our plans and ideas out into product.

And look, our early forays were not so happy. Right? Code review was a massive bottleneck because all the developers are using Claude Code and throwing it over the wall to code review.

And then, so it's all AI slop. And so we come back to the idea of culture, right? You talk about DevOps and DevOps metrics and the state of DevOps report metrics around things like cycle time and error rates and these sorts of things, new errors in production.

Look, there are metrics. Good code is still good code. Good production systems are still good production systems that do the function you want, respond quickly, don't lag, don't throw out security vulnerabilities.

And you can measure all of this, whether it's AI creating or not. And we are, by the way. And so I think that's actually, I'd push back on the idea we don't know how to measure what we're putting in prod there, because the good things are still good things.

But the pace increases. We've been able to do so much as a business by using AI in pre-prod. And this is where I think part of the challenge is.

We've decided that AI is good across the board. We're not doing what we have historically, which is try it in a space where there's a low impact radius. Right?

We're just throwing it straight into prod and concerned that we've got security issues, dependency changes, performance impacts. What we've done, and I'll blow my own trumpet a little bit, we've started mostly focusing on pre-prod. So the stuff that ends up in prod, we're still measuring the same things.

Cycle time, time from idea to cash, number of defects in production, mean time to detection, these sorts of things. But we're doing a lot better because AI's doing the work in pre-prod, including quality checking, which your article also highlights. Yeah, and documentation is a pain in the a*s and AI does it reasonably well.

So this is a good thing, right? Yeah. But it's just I think maybe we're too obsessed with the coding aspect of this thing and not thinking through all the other stuff that AI can help us with that frankly, a lot of people who code enjoy the coding for it, it's all the other crap they hate.

It's true. Coding is only like 30% of software development. There's planning, solutioning, trialing, testing, brainstorming, which AI's not doing a lot for you.

Yeah. So I guess our consensus here is that we do not believe that two years from now or three years from now, that AI is going to be too costly to use. Is that kind of the consensus?

Yeah. I love it when the group comes together like this. Not too often.

Not too often. No. So, I can put this Gartner report in the circular file with many others.

Yeah. File it with the rest of them. Yeah.

Yeah. That's a redundant statement, Mike. No, sorry.

With your buy orders for Gartner stock. But let's move along. Move along, move along.

This is not the analyst firm you're looking for. Mike, what have we got next on the agenda? All right.

Well, we're talking about Toy Story. We're going to the movies. And there's been this whole theme of this movie is wondering if technology and AI is killing our childhood, and that's sucking out the joy of being a child because everybody's playing with something online and they're no longer playing with each other.

I think we've seen, at least I've seen it, where kids are sitting in a room and they each got a screen and nobody's talking to each other. I've got a youngest who probably is texting with somebody who's three feet away from him. But this seems to be kind of the thing lately.

But Alan, you went to the movie. What was your takeaway? Classic Toy Story.

Buzz and Woody have lived in my house for 25, 27 years. Yeah. Both my boys were big Buzz and Woody and Jessie and the rest.

Mr. and Mrs. Potato Head were always big with us too, but they weren't really in this movie because I think Don Rickles played Mr.

, but they're both unfortunately passed away. But the theme of this movie was like so many of the other Toy Stories. It hit really close to home for me.

Right? The real theme is the age of toys is done. The Buzz and Woody generation.

And Buzz was a high-tech toy. He had flashing lights. But the age of toys is done.

It's the age of technology, and all the kids have these little LeapFrog kind of pads and everything, and the little girl can't wait to get her own LeapPad so she could be like her friends. And she gets one, and she gets invited to her friend's house, and they all sit in the living room. They each have their pad, and they don't talk to each other at all, as Mike described.

I've seen this with my two boys in the back seat of my car. They have whole conversations going on, never saying a word. It's what makes true.

I like to listen to my music or whatever, but this is the way it is. They don't want you to know what they're talking about. That's a different issue.

That could be, too. But anyway, but here's the thing, these kids were miserable. They were miserable because part of being a kid is using your imagination.

It is playing. It is, as I wrote in my article, it is putting the covers over the two beds or the dresser and making believe it's a cave, and tomorrow it's a starship, and the next day it's a dollhouse or something, right? The thing that ultimately is going to separate us from the AI is our imagination, our creativity.

And if we kill that creativity and imagination in our children, when they grow up, are they going to be able to still have that edge? And that's the moral of the story here, right? That Woody may have a little bald spot in the back of his head, all of us do it seems, but he's still a great toy.

Him and Jessie, you could play with them for hours and just make up anything you want. Where a lot of these technologically enhanced toys, they isolate us. Mm-hmm.

And it's not just kids. Think about all the adults you see in the airport, everyone's doom scrolling and doing their thing there. You know what, Alan?

You know who agrees with your take? Is Steve Jobs. And I think 30 years ago, before he went back to Apple, he was going through his wilderness years, right?

So he granted this interview he did at KQED, which is a public broadcasting station in San Francisco, about 30 years ago. And one of the things, it was very frustrating to interview him because he's always on point, he's always selling or marketing something on a very narrow level. But one of the questions we asked him, I think offered so much insight into how he thought and where he was going, even at that time.

" Because he famously didn't admire many companies. And the number one company he mentioned was Disney, and he said the reason why was Fantasia. It's a creation that's generation to generation, and it kind of returns us to this fleeting chapter of our youth.

And he mentioned this as something that would pass from generation to generation. It was based on creativity. It's like childlike wonder.

And he pointedly said that's something you could never do at a company like Apple, where it produces these products that are obsolete within a year. And I thought that was interesting insight, and it foreshadowed the fact that he wanted to run a studio and make movies like this to touch people at their basic inner core. Mm-hmm.

That's one of the more insightful things he ever said. He didn't say very many because he was so secretive and withdrawn, but- Right ... he would open up if you talked about these types of things.

" But even in that community now they're yelling about the players who are using AI are ruining the game and the experience because basically, now humans are playing against AI, and then it's just my AI against your AI, and they're like, the whole point of the exercise is shall we say, defeated. And they're sitting there going, "I came here for the social interaction as much as the game itself," and that's getting killed by people who just want to win at all costs using AI. I think this is another example of what I was mentioning earlier about accelerating a bad idea, right?

150 years ago, all toys were made by hand. Your parents made the toys. You maybe didn't get any, but people who did get toys, there was something special about it.

So mass manufacturing them led soldiers and so forth, that's good. All kids get toys. But you lose that connection, and we've taken that to a step and we've taken that to a step.

In the Toy Story arc, that each of the episodes over the decades calls out part of this. So rather than just looking at the AI part of it, we might want to ask ourselves, was it a great idea to outsource all of our children's identity and attachment to the toys in their lives to a corporation that also owns the movies, that sells the toys, and, and, and. So maybe that wasn't a great idea, and maybe, again, AI is just accelerating that bad idea into a wall.

Chris, I'm not certain the bad idea started there Because if you go back to- Mm ... a few of... I have one or two contemporaries here it may seem, maybe not.

It was with TV. " This is when TV was new. Now- Right.

The blueprint is radio ... my father sold TVs, so there was a conflict immediately with that. But we were limited at an hour a day.

That was it. " And I'm not sure that's what we need to recommend, but that scale works. I was forced to, whether I liked or not, go outside.

I don't care if it's raining. I don't care if it's cold. Get outside, find something to do.

And you would maybe get sticks together, you'd make a fort, you'd go on an adventure. And that imagination... And I'm not being the, "Old days are great," but it was really the TV started with okay, TV became the babysitter.

Mm. That advanced to computers, that's advancing to AI. So I think we're just on that scale, and how do we manage it?

Right. Well, I don't know about you guys, but I was watching television at home because nobody would let me out. So, once you let me out, I was gone.

For good reason, Mike. For good reason. I tell you though, I'm just going to sound a slightly contra- I'm being a contrarian today.

I don't know. Whether it's in the water or something, but, or maybe it's just me. Just a slight contrarian note.

I do think that especially for people of marginalized backgrounds and so forth, the internet and the doom scrolling, the online chat, it actually opens up new communities for them. And I think that kids are, just, I don't have kids, so what do I know? But I think that they're very resilient.

And the children in my life have found ways to connect, to socialize. You think about the idea that they're all sitting in a room doom scrolling with each other. But what they're actually doing a lot of the times is they're including someone who couldn't come.

Yeah. And they have conversation together. Right.

And I look at kids out in my part of the world riding their electric bikes, right? Most of them are illegal. They're modded to go too fast.

And it's like everyone's going, "Oh, they can't do that. " But we're still saying they should get outside. But no, not like that, guys.

Not like that, right? They're not going to have fun the way we want them to have fun. It's not-- I think they're going to be all right.

I think the social media bans that we're seeing are probably not a bad thing. But I think the kids are finding different ways to make community, to make friendships. That's a good point, Andy.

They're resilient, and they're also more open-minded to other ideas and other types of people that they may have never been exposed to or talked to face-to-face. They also have a better understanding of AI slop, right? And they look at it.

" But I digress. The other thing is, though, the kids are going to things like, I was at a used record expo kind of thing, and the place was packed. And everybody in there was in their teens and 20s because they wanted to be able to touch something and own something and kind of like listen to it.

And they were having a different kind of tactile experience that I guess they don't get with AI, and I think that's starting to come through a little bit. Wholeheartedly. Yeah.

I'm actually with you in the Pollyanna side of it, Andy. At the end of the day, I've thought about this my entire career, and something I said in the early '90s, because it sounded good and I was trying to explain things, is that imagine every young child in Afghanistan has access to all the information in the world, right? There's a positive to that.

My kids have grown up and have connections. Their social group exists because they can talk to people through the internet. And in the last year, I and Ashraf and our teams have provided AI to thousands of childrens in Yemen.

But we come back to the topics we can cover on a show like this. Would I give them open AI? No.

Their own AI in their community based on their stuff? Yes. That's actually huge.

It's teaching them mathematics in their own culture. And play. Yes.

But local, local, local, not in Jeff Bezos' hands. That doesn't make any sense. Why would you do that?

I don't know. I used to love playing army and making a fort and doing stuff that kids do. I, in my day, Jeff, you'll remember this because we are in the Coleco little handheld football game that you would go like this, beep, beep, beep, beep, beep, and you'd move it all up and down the field.

And that was wow. But it paled in comparison to going out and playing football with your friends. Mm-hmm.

Right. And that was the difference. And- We used to organize our own games, like among the kids.

We didn't have to have somebody tell us. You didn't need parental- No, of course not ... oversight, and we didn't need leagues or coaching ...

choose by the bat, whoever got to the top of the bat- Yeah ... had the first pick, and- Exactly. You went to the schoolyard, first 12 guys played.

Yes. Yeah. Well, yeah, I'm looking forward, like we talked about this week, I'm looking forward to see Ashraf on this show, right?

Because look at all this on the screen. We have a certain perspective. There's nothing wrong with that.

However- What are you trying to say, Jeff? Please. What do you want to- When you grow up somewhere where you're not so privileged, the benefits of these systems are what they are.

Nobody in the Third World or outside of this screen is unfamiliar with the downsides we talk about all the time, but the upsides can be tremendous as well. And we don't have to do it as poorly as we tend to do so frequently. There you go.

I want to add something here, and without trying to get too patriotic here, what we've witnessed with the World Cup visitors As much as everyone's saying, "Here's what's wrong with America today," and no matter which part of the political spectrum you talk to someone on, they're going to say, "Here's what's wrong with America. It's terrible. " But visitors come, and all you hear them say, "This is an amazing place.

I can't believe you can get away with doing everything you do here. " And I think we lose sight of that. And with 250 coming up in three days, I think we need to, at least in the US, remember there is a whole lot of goodness, Chris, to your point, that's in here.

And the things that we know need to change or need to go away shouldn't overshadow all the goodness that we have. Maybe it is fitting that the World Cup is the big 250 celebration. Mm-hmm.

You know? Melting pots. We will be talking about 250 tomorrow, so we're going to have a little- Yeah All right, let's leave it for then.

All right. Hey, but the US is playing tonight, speaking of World Cup. But I think we got to end it.

We're at time. Hey, Andy, Jeff, Chris, thank you so much for coming in. I know some of you came in on short notice.

We had some cancellations. John and Mike, it's always great having you on. Great having you watching this.

tv, Techstrong TV YouTube channel, and our Techstrong OTT channels from just about any screen you watch on iOS, Android, Roku, Apple, Amazon, wherever you watch screens. Even your little LeapPad for your children, though I don't know if this is suitable for them. But we'll be back tomorrow.

We will have a special Celebrate America 250. I don't know if Mike's coming dressed as Uncle Sam or what, but- ... until then, this is Alan Schimmel.

We're out.