AI Backlash Grows as Technical Debt and Courtroom Chaos Pile Up

Hey everyone, happy Thursday and welcome to our Thursday edition of Techstrong Gang. We opened up the books here at Techstrong Gang and reached out to some of our friends, and I'm happy to see some new faces that I think are going to bring some great expertise, great angles, and great conversations to our gang. So let me introduce you to our gang today.

We've got my friend, David Nicholson. If you don't know Dave, Dave's a treasure. What can I say?

He's a professor at Wharton, an all around well- Not a professor, instructor. Instructor, excuse me. Lower level.

But beyond that, though, that's great, but he's also well-traveled and I appreciate his expertise and what he's gone through in his career. So Dave, welcome to the gang. Returning to the gang after a brief hiatus, my friend, Dr.

Stacy Thayer. Stacey, it's great to see you. Thank you.

Making his, I guess his debut on the gang today, is my friend Dadisi Sanyika. Well, I've been doing podcast stuff with Didisi for years. He's the chairperson of the CDF.

And of course, joining us is our two stalwarts, our man in the West Coast, Jon Swartz, our man in the East Coast, Mike Vizard. We've got a great gang for you today. Mike, let's kick right into it.

Well, it seems like suddenly out of nowhere, it's AI haters day, and all this stuff started piling out. And starting with a survey from the folks at Pew saying most people seem to have issues with AI, and it feels like more people start to say there's more negatives than positives here. And then just for fun, Jason Kelsey showed up with a commercial talking about how well we should collectively start sending all our urine to the data center people because it'll help them cool their data centers.

Then things got even more interesting. Political climate out there is the folks at "The Washington Post" and "Politico" are reporting that the Republicans are starting to freak out about this data center issue and AI because they're worried they're going to lose some elections over this, and particularly in Ohio. There's also been recall movements floating around out there, and there's all kinds of noise in the system.

Jon, you covered a lot of this stuff. Is all this coming home to roost? What's going on here?

Yeah. God, I have to tell you that every story I write seems to have the words backlash or pushback in it. They become fixtures and leads.

I don't know how much time we have for this segment. I know how much time we have for it, but you're right. Increasingly, this has become a topic.

It has gone mainstream. So public backlash against AI is mainstream. So you mentioned the Pew survey which came out.

The really interesting thing about that survey is for the first time it shows significant distrust among people under 30 towards AI. So it goes across most demographics. We have this issue, as you mentioned, where politicians are turning data centers into a major campaign liability or benefit.

And you mentioned there's a case in Ohio, so the GOP had a memo that was circulated. Axios got a hold of it and published it, and basically what the GOP is telling its members, specifically this is geared toward the Senate race in Ohio, John Husted against Sherrod Brown. Husted's a big proponent of data centers, and he, according to the polls at least, is eight points behind Brown and Brown is carpet bombing TV with ads about Husted and his pro-data center stance.

We have another issue in Pennsylvania where Josh Shapiro, who's the governor, in a sense wants to rein in data center projects because he knows it's politically advantageous. But the irony is that he was actually very pro data center just a few months ago. And then finally, it can't be a story unless we have some sort of Kelsey brother angle, as you mentioned.

And it all kind of comes back, I think, where this whole movement or this whole kind of zeitgeist has evolved from a purely technological question into a struggle over ownership, wealth distribution, jobs, this idea or notion that data centers are predominantly in rural areas. And who's going to benefit? The big East and West Coast guys, the bad guys in big tech.

It's all percolating and it's all building momentum, and it has become a major issue, and it's become a major focus, and I don't think it's going away. Alan, are the AI folks losing the hearts and minds here? Is this what's going on?

Well, they're a lovable bunch after all. Who doesn't love Elon Musk or Slippery Sam or the rest of them, right? But the interesting thing, Jon, yes, the Gen Z's are...

But can you blame the Gen Z's? If you were getting hammered, "Hey, you're going to lose your job. You're not going to get a job out of college.

Lower ladder jobs are being taken by AI. " You can only beat the drum so long before people start getting the rhythm. And I think that's what we're seeing here.

Here's what's interesting, though, and Didisi, I'd love to hear your take on this. It seems like developers love AI, right? Because they're using it.

God knows the code's all being generated by AI. So are developers the outliers here, and is it the political class and the people living next door to the data centers that we're hearing from? Just yesterday, I was talking to the ambassadors of the Continuous Delivery Foundation, and this was a topic of discussion, not just how are we using AI, but how is AI complicating their jobs and their expectation to hold onto a job, and then how are you fine-tuning it to actually get real results that you can do something with?

The struggle has been that the way that AI was pitched, it's supposed to deliver something extraordinary, but really it's just a dictionary of things that are already done. And so does that really leverage? Does that really help?

And so no, the programmers are actually having the same problem where the value that they are getting out of it is only benefiting the same big guys. And so there is that worry and stress in those same groups and how am I going to maintain my position? How am I going to continue to perform and excel in my space when AI is right there next to me?

I think that there are use cases for AI where there's clear benefits, but I think most people are experiencing massive amounts of AI crap flowing through their social media feeds, and they're sitting there going, "Really? " And people are going, this doesn't seem like the right trade-off from a cost-benefit analysis in their head. And then the next thing you know, somebody's screaming about let's get some pitchforks, and away we go.

Stacy, is this a psychology issue? It's a people issue. It's a human issue, all this AI stuff, and I stand by it.

No, I think when you see developers are struggling with... We always had issues with third-party code, right? Well, now the third party is AI, and it might be making the code easier, but I don't think it's making it better necessarily or easier to merge in or that there's struggles.

So I live in Loudoun County, which is Data Center Alley row right there, and I actually took a drive past some of those data centers, and it's weird. It's eerie out there. There's big giant buildings with no real parking garages because who's going to park there?

Who's going to go into work in these AI centers the way that you might see at places that we're used to seeing offices at. And then the electricity that's everywhere. You got the Verizon building, the next two is your own electrical grid, right?

And so we're seeing all of this, and yeah, I think there's that question of things are changing, but are they changing for the better? Is it actually better? Making life more complicated and messier in a lot of ways.

Students certainly love AI. Their professors don't. But I think that that's what we're trying to figure out is who's using this and what are we using it for, and is it the best case scenario?

Is it doing good? Dave, you wanted to say something. Yeah.

Look, just to my fellow activists out there, we've successfully driven manufacturing out of the United States. We've successfully prevented the development of nuclear energy for 50 years, and we will win this war against data centers. We will keep them out.

But I don't want anyone to worry about AI because as my blonde, green-eyed, Mandarin-speaking wife would say... Which in English means where there's a will, there's a way. And if you don't have the chips to have your own data centers in short order, flood the market with essentially free AI models, which is what's happening now, and the data centers would be built.

They just won't be built here. And much like semiconductors are no longer built in the United States, we just won't have data centers. That's the direction.

And I would just say follow the money in all of these things because there are people who benefit mightily from not having data centers built in the US. That's my conspiratorial theory on this. But Dave, let's name names.

Who benefits mightily from not having data centers, beyond the Chinese? Yeah. No.

It's anyone selling access to the US economy to manipulate it. So, lobbyists, lawyers, politicians who are in one way or the other being lobbied and having money funneled in. And frankly, it's driven on top of a lot of ignorance.

There are legitimate reasons to not want a ridiculous data center in your backyard, absolutely. But a lot of the environmental concerns just don't hold water factually, except for that concern of, yeah, but I don't want it here. I'm sort of a yes in your backyard kind of data center guy.

I don't want them where I live. So Dave, let me play devil's advocate. Sure.

Let's take Elon Musk. We all know Elon. What an advocate for clean energy.

The man invented Tesla. The man said fossil fuels was a huge mistake, and he's going to correct it almost single-handedly by ushering in an era of electric vehicles. Until it came time for Elon to make a couple more dollars at his Colossus or Hyperon or whichever one is his, and he's building ...

No, it's the one in Texas, the mega factory or whatever. Stargate? Is that Stargate?

I don't think it's the Stargate one. No. It's his factory where he's building chips and everything else.

The Gigafactory, whatever. He's building 28 gas turbine- Correct ... on his property there off the grid, so it won't cost us all theoretically the electricity.

But what happened to green Elon? And how bad fossil fuels are? I'm not against building data centers.

We need the data centers. I don't know if we need data centers that cover 40,000 acres of Utah, but we need data centers. But we need data centers that are using renewable energy, cleaner energy- You can't power them ...

more efficient. You can't power them with solar and wind The only- What? You cannot power them with solar and wind.

So then fine, let's do something with... Why can't you power them with solar and wind? Because a 10 gigawatt data center takes photovoltaic solar the size of Rhode Island with about 100,000 Cybertrucks' worth of batteries to keep it running at a- And where do you want them to put the infrastructure for that?

$10 billion. 10 bill- So, you're saying then we have to go to natural gas turbines and the environment be damned? Correct.

The environment- And you don't mind living in that world because f**k, to hell with it, man. I got my data center and my gas mask- So, so- ... and I'm good ...

as long as CO2 is a pollutant, correct. That is correct. Yeah.

So let me... I'm going to bring this a little bit on a personal level. So I've been exchanging email and I've been talking to a lady who lives in Austin on a farm.

She owns hundreds of acres. She's had this farm in the family for years. SpaceX just bought a big parcel of land around her, and she's leading this NIMBY fight against this.

I'm going to write about this, but she was explaining to me just the absolute... It's almost like a microcosm of the absolute gall of a person like Musk, who's going to ruin the environment and this uprising, and it really is becoming this, I think, Mike, we were talking about this earlier, so it's like an us versus them. It's like the vast majority of us are not benefiting at all, versus the, of course, tech bros, the oligarchy, whatever you want to call them.

Yeah. Dave, you talk about China. Yeah.

We hear about China putting data centers at the bottom of the ocean and sitting on top of it are wind turbines over the sea. Did they not get the memo that that doesn't make enough electricity? Because they're usually pretty smart.

No, a wind turbine can generate maybe in the megawatts range of electricity. So- But how many do you need? Yeah.

Tens of thousands of them. So if you want tens of thousands of turbines in the ocean, that's fine. The point is, China, in this case, takes a longer-term view of things, and they accept that the burning of dirty coal might be the gateway to a cleaner future that's beyond a presidential term away, and that's where they have us beat.

" Fill in the blank. EVs. Well, that's what the farmer's going to do.

She's going to sell her property. She's moving out. Yeah.

So don't get me- She's going to pack up and then they're going to- ... completely pathetic. I certainly have compassion for any of it.

But again, if you take the long-term view, we have history as a precedent. What has our lack of development in nuclear energy gotten us over the last 50 years? What has our driving of the semiconductor industry out of California and then out of the United States?

Where has that left us? So I would just say that- Dave, I think those are great both precedents, and there are other precedents, right? There are other precedents that go the other way, and no one is saying don't build data centers.

5 trillion worth of build-out. Yeah. So we're not going anywhere- And there's an explosion.

There are thousands of data centers. I'm looking at- The wild card here-- Let me just say, the wild card here is this, and- Right. How much AI can we run on our phone that makes- Right now ...

those data centers- Right now, none ... redundant? Right.

Exactly. Right now, four billion parameters on this device. Two generations from now, 80 billion parameters.

My desktop computer, which is now worth three times what I paid for it 11 months ago, 200 billion parameters. So I think inference at the edge, frankly, is going to make this discussion a bit moot. Dave, let me give you another historical context because you mentioned semiconductors and so forth.

Let's talk fiber. You and I were both in our careers already during the dot com days, right? Dude, I'm 32.

What are you talking about? Yeah, I forgot. But, I remember sitting in Enron, trading fiber that I bought from WorldCom for $1,200 a megabit a second, and Enron was going to wheel it around and sell it back to me for $600 or $700 a megabit a second.

Yeah. And by the time the crash came, fiber was $50 a megabit a second or something like that. Couldn't agree with you more.

Couldn't agree with you more. Right? That could be this $8 trillion in data centers.

Yeah. Tell me the reason. Do it.

Do we need to ration the data center access because a lot of this crap that's being used for water and all this other stuff is being used to create crap, and if we have these issues, maybe we got to be smarter about it. Yes. One person's crap is another person's treasure.

Yes. No more anything from Mike Vizard. I'm the arbiter of that.

Yeah. Everything you do is awful. How about we don't use AI so we just can craft a better marketing message, and we use AI for medical research?

There's a thought. Well, this is the Dario Amodei argument, right? We need to cure cancer to give AI a good name.

Mm-hmm. Because we're using it for a bunch of, let's be honest, b******t. Yes, exactly.

Exactly. Predominantly, that's what people use AI for. If you're going to be using it- A preponderance of s**t.

So is it that we need to improve how we use AI? Is that, we need better data to make it more useful? Where is the- You're going to have to take that AI out of my cold, dead, stiff fingers like my gun.

I'm not going to take it out of your hands, but I'm going to make it prohibitively expensive for you to use it. I think it already is. Oh, step on the little guy, Mike.

Understand there's- Yeah. Yeah, man. Pick on Allen.

It's just- Hey, but guys, we're at 18 minutes. We're over this one. We will come back to it But we got to move on to our next segment, Mike.

All right. We are moving on to the next segment, which is somewhat related. But there are new reports out there that are saying that enterprises are having a hard time wrapping their arms around this whole AI agentic workflow thing, and a lot of it has to do with, well, their systems on the back end weren't designed for this stuff.

And the issue is they don't have the controls, they don't know where the data is. There's just a host of this stuff, and it's a loss. At least Deloitte is saying this might be a lost trillion-dollar issue and opportunity.

Dave, you have studied this stuff extensively, so I'm going to defer to you here. But are we not going to realize the promise of agentic AI because, well, frankly, the IT systems that we have in place weren't designed for it, and we need to re-engineer all of them first before we're going to have AI agents that are really meaningful? Yeah.

I'm going to be generous here and say that, as we all know, we tend to overestimate the pace of change in the near term, and we tend to underestimate the scale of change in the long term. And so I think this is about our expectations and how quickly we can derive value from these systems. And a lot of that was sold to us by these same people who are telling us we need all these damn data centers.

" And then the narrative changed. The narrative changed to, "Oh, you know what? " And phase three is exactly this: "Oh, we sent those humans out, and what they found was a nasty ball of chewed-up gum, hair, and all sorts of nastiness, AKA technical debt" And it's a lot harder to tease ROI out of these environments than we were led to believe.

I think a lot of us here knew that that was going to be the case. But the final interesting piece here, even companies that sell AI are challenged with proving they're getting a positive return on their investment. " But in that middle zone, they really don't know how it's being used effectively.

So that middle part that involves, guess what, bunch of other people, going in and having conversations with people about what are they doing, what do their workflows look like, that's a long slog. So this is the CTO dilemma. That hasn't changed.

Technical debt's always been there. I think the difference is we've all been kind of conditioned to think that the magic was going to happen sooner than it possibly could. To DC, you live this, man, right?

In the CDF and everything else. What do you say? It's the same point I was about to make before about the data itself is probably the biggest issue.

That's the issue that we're seeing. It boils down to AI's not creating anything new. It's not going to think up anything new.

It only has what's in the model. And no matter how you kick it, no matter how you push on it, you can only get back out what's in the model. And so those things are the things that we already know, but shaping those and matching those beyond the normal expectation is where new things come from.

And so if the model is trained to give you the thing that it's most common, that's not necessarily going to be the new thing. So that's the real challenge, and that's why you need a body in the seat. But that complicates things, right?

So we don't have the mechanisms yet to sort through the data. We don't have the processes that make the agents work more agentic. But all of that, if it showed up on the market, would cause a lot of these things to collapse.

Like the need for data centers if the agents are more deterministic and need less thought processing to move forward, is going to collapse the market. So this is eventually, like so many other things, it's going to start to take care of itself in its own space. But a lot of the assumptions that the people who are selling it to us, who want us to say, "Hey, spend your money.

Invest in this. It's going to have a great future. We're going to be huge.

" As the process improves and the technology improves, it's going to get smaller and smaller and smaller, and you will have given away all of that money for data centers that you don't need in the long run. I can't help but wonder if the entire argument around ROI is kind of nonsensical, and I'm going to throw this at Dave. But if I go out and I use AI, and I'm a hospital, and I automate some process, and it's faster, well, that's great.

But two things. One is there aren't going to be more patients, and two, the hospital down the road is also going to do the exact same thing. " So where the hell is the ROI going to be in the first place?

Yeah. We're good at calculating the I because investment is cost, right? The R, the return, is a lot harder to quantify in some cases.

And look, I can tell you that the folks in the programs that I teach are battling this, especially in healthcare. But not everyone is looking for return on investment in the form of profit. It can be services delivery.

But there's another big question in there, Mike, and that is to whom do the benefits of AI accrue? If I, as an individual contributor, suddenly become 300% more efficient, does that mean that Alan, the guy with the stick beating me over the head for that productivity, does that mean he just gets more boats? Or do I get an extra Friday off every month?

Or do I share in those benefits? Because if I don't, we talked about the quiet quitting and all of that. Boy, I'm going to be on YouTube a lot more secretly with that time that I've reclaimed with AI.

So this is a big societal question of to whom do the benefits accrue? And return is a tough thing that we're struggling to figure out the return side, and DC's ... comment was very well-grounded in reality of how these systems develop over time.

I'm just waiting for 20 AI agents to let me go to that afternoon game for the Yankees on a Monday or Friday, and if they have to go more, they can go more. Not happening, Mike. We want more of articles.

More. We need more. But Stacy, let's look at the human side of this.

So- Part of me, Stacy, says that Dave should relish that his 3X more productive than he was before because he wants to be the best Dave he could be. He's not just looking to cut a corner and grab a Yankee game now and then. Is there an HR human element to this to get people to understand how do the benefits accrue to them?

So I think this is one of the things, you spoke about millennials, for example. " Right? And so the question then becomes, well, does that mean that they can say, "Hey.

Okay, great. Here's AI. " Or are we going to use that and say, "Oh, that's great.

You just found 20 things that AI can do. " And I think in this country, we're always going to keep building upon that and finding a way. And I think we are at a choice point right now where we can take a look and say, where do we want to be?

How do we make these decisions? Are we going towards this AI, this impersonal, data centers and tech, and all of that, which can be good, but I don't think people know how to use it yet. I don't think we really understand how AI and humanity are going to work together.

And so people are told, "Go use AI. " But then they're wondering, am I replacing my job with AI? Am I enhancing my job with AI?

If I find a way, then they can get replaced. So they're not necessarily thinking about, can I set this to go and then go take my four-day work week or go watch YouTube? So I just feel like we don't know what we're doing half the time, even at the higher levels, even I think the people who are working at these organizations, who are working at these different AI companies.

" Well, how? "I use AI. I use it," and that's it.

But Didisi, I'll come to you. Mike, I'm sorry to jump in here for you, but- I was just saying John's on mute, and he doesn't realize. Oh, John, you're on mute.

Well, he was on mute. I'm going to Didisi. Didisi- What the hell?

I'm sorry. Sorry about that. Go ahead.

I might have lost a word. Didisi- I lost my place, Dave ... 5 trillion in technical debt.

Some would call that a ticking time bomb. Oh, go ahead. Go ahead.

No, go ahead, John. Sean, go ahead. No, sorry.

What I was trying to say earlier when I was on mute for some reason, is this idea we're kind of like in this era of purgatory or kind of this disconnect where we really don't know what's going to happen next in terms of the impact on our jobs or what the new jobs that are going to be created are in terms of AI and the people who are behind all this who wanted us to think it was going to be easy and we were going to have this 30% productivity boost. This is one of many speed bumps we have hit that is part of this elongated story about AI agents in particular, where we all thought it was going to be easy and simple and fast, and it's not. This is a major disconnect.

But it makes us want to adopt it faster if we think we can be faster and better. You're right. We bought that story.

We've been sold that this is going to make life better, easier for you, and companies are eating it up. Yeah. 5 trillion in debt?

Here we go. All right. So- Dave ...

5 trillion in debt, guys. So let's take a step back with AI right now because we're falling into the same trap that's the sales side of AI, right? We're talking about the hype of it, not the actual technical of what is it doing, how do we take advantage of it.

If you look at the industry now and all of the tools that are available, especially in the SDLC space, there's a lot of repetition. We're building the same things over and over again. And had we done something really, really radical like that very first CI/CD tool, everybody built on it and kept expanding it and worked together to have one way to do a thing, we'd have less toil.

There's so much toil out there. There's so much debt out there. Someone needs to be rethinking how we bring information together and use it with a system that wants to predict.

That means better sectioning of our data, pulling it together in a way that makes it collaborative so that we can analyze it better. All of these steps are going to improve the system overall. But then we can use that same AI to kind of make some of those corrections in our own process.

That's literally the thinking that we have moving into the future in the space that I'm in now. We need to change our approach to how we work in this space because we know what the tool does, and if we understand how it works and how it's making those selections, we can actually drive our work in a way that takes advantage of it. Fair enough.

Mike, it's that time. Perfect. We got to hop to three.

Well, we are going to hop to three because, well, it's part of a continuing theme, but there's been reports that say that people are starting to file legal documents with prompts that are embedded in those that are trying to direct the judicial system in a particular direction, and they're getting called out for this, and other folks are starting to use AI to kind of Maybe game the system a little bit, and there's an example involving 3M that said, come up with a legal case here that shows that 3M has no responsibility for X, Y, and Z. But Alan, you have come from this background. What are the dangers here of using AI in this context?

You can see how AI would benefit legal stuff, but it seems like people are also being people, and they're doing bad stuff. First of all, Mike, I got to go disperse the line of people out my door who are waiting there because they feel sorry for lawyers. Other than the AI oligarchs, is there a more hated group in America or the world?

Shakespeare said it. But as a reformed lawyer, here is the issue. One of the things they actually teach in law school, and not a lot of lawyers are going to tell you this, is it's okay to give the client the illusion that there's magic going on here.

Never let them behind the black curtain to see how it really works. Right? It's almost like when the Roman Catholic Church had the masses only in Latin.

Most people didn't speak Latin, they didn't know what the priests were saying, but it made it holier somehow. Right? And then, of course, the church went to making mass in, funny thing for a Jewish guy to be talking about, but went to having mass in regular languages so that people could understand what the Bible said and what the priests were talking about.

Very similar in law. Lawyers like to talk legalese, so that the regular people can't be involved, they can't do it themselves. Now, all of a sudden, with AI, they can do it themselves, right?

And that opens, like everything else in this world, a good side and a bad side. The good side is you have people who can't afford lawyers that can actually defend themselves. My son built a chatbot in his senior year, in his last year of law school, to help tenants go into landlord-tenant court and defend themselves, because if you don't have money to pay your rent, you probably don't have money to pay a lawyer either.

And it was a huge benefit for tenants. However, the abuse side of it is people are going to use it to file frivolous lawsuits, to clog the system, to do things that are wrong. And then the lawyers themselves, Mike, the lawyers themselves are abusing it because no one's lazier than lawyers, my friends, and I say that with all due respect.

They're using AI to write their briefs, and the AIs are hallucinating up precedents and case law and so forth. And now the judge, not the judge, because judges do very little real work, but the judge's law clerks have to sit and go through these briefs and find out what's nonsense and what's real. Right?

Because that's just the nature of how law works. And so we've seen a whole spate of those things, too, right? Where judges are now holding lawyers in contempt if you give them- Alan ...

nonsense. Alan, can I ask you a wonky legal question? Sure.

Okay. So does the improper use of generative tools by expert witnesses, does that undermine evidentiary standards required under court rules? No, not at all.

So, you said a word there, expert witness. An expert witness is different than a non-expert witness- Okay ... if I just called someone, right?

But if I call Dr Stacy Thayer to talk about psychology or something like that, I have to qualify her as an expert. "Stacy, what expert qualifications do you have? " Okay.

Now she has the impromptu of being an expert, but it's still up to a jury, or the judge if it's not a jury trial. The decider of facts to weigh what Stacy's saying. " And she says, "Yes," that's up to the jury to decide.

Ah, it's AI, it's b******t, or it's not. Right? But ultimately, it's not automatically excluded.

It goes to how much weight it should hold. Okay. Best I could answer that for you.

So, I think there's general agreement that the legal system is slow, expensive, and a lot of the people in it have incentives to- Keep it that way ... to keep it that way because they're paid based on time- Mm-hmm ... rather than anything that feels like an outcome.

So, I get that there are negatives here, but maybe the preponderance of the good of AI is that can we shake up this legal system so that we're not all feeling like we're getting robbed every time we talk to a lawyer? Dave, who are the people stopping the data centers? The lawyers.

The lawyers. The lawyers, Alan. You're going to have to do legal genocide or something.

You're not going to stop them. Look, at best, AI just becomes a form of an arms race in the legal profession. One of my students, sort of off to the side of legal in the healthcare space, revenue recognition.

So major hospital system, $3 billion a year in revenue. The way they get their revenue is by billing insurance companies for the work that they do. Well, guess what?

Insurance companies are using AI to decline, refuse, delay, do all that stuff. And so she has to use AI on the other side to address this, and it's this completely crazy environment. Back when Alan and I were teenagers, I was working for a company that did automated litigation services, and I watched these teams use the process of discovery to try to bury the other side.

And in every single one of these business litigation cases, the only people who won were the lawyers in the middle. Everyone else lost. And so, yeah, it's an arms race.

Frankly, the read here, the injecting of prompts, I thought it was hilarious. I laughed my rear end off at the whole idea. " So I don't know where it's going to lead.

But the courts do have the power. This goes back to something we spoke about yesterday, right? With Claude embedding watermarks in AI-generated stuff.

Are we going to start looking for AI watermarks in legal filing? When everyone's using it? I don't know.

But let me return to the very first thing I said. For those underserved elements of society who cannot afford legal representation in important litigation, AI is their last best hope. 100%.

Because if it's a criminal case, you're going to be assigned an attorney, but if it's a civil case, you don't get an attorney. And the guy who has the better lawyer wins then, because he's the only lawyer usually. Mm-hmm.

Right? And so for those people, this is a real lifesaver. This is like throwing someone a rope.

See, Alan, I think you're bringing up, these are the parts of AI that give us hope, like when we're talking about curing cancer. And there are positive use cases for AI, and I love hearing about them because that, like I said, gives me hope. And then there are times when I think, oh my gosh, the internet gave everybody and their uncle a platform to talk, and they did, and now look where we are with media.

And now, not only do they have a platform to talk, they don't even have to think critically about it because they can just enter in their thoughts into AI. If we're using AI now in legal jargon, where do the lawyers even come in? At what point are we handing over the keys to our kingdom to think critically and to serve our purpose in humanity and our tools and techniques, and then we don't want to throw the baby out with the bathwater where it is doing such good for people.

So when it's used, it's like anything else. When it's used for good, there's so many wonderful things that it can be used for. It's just we, as people, to Mike's point, we're going to human and people are going to people, and we're going to do things that are bad with it.

And I think as we become more mature and we evolve as understanding, much like we're doing with kids and the internet, where we're saying first it was no screen time for kids, and now it's, well, maybe a little bit. We'll learn to use it, I hope, in a way that is responsible and makes it for people. So OpenAI did that, right?

With ChatGPT for Teams. They just announced that. And you're right, Stacy.

Right now, I think the narrative wholly, though, is the pitchfork crowd has got the megaphone right now. And as these other use cases come out, whether it's medical, legal, et cetera, we will see the benefits. But right now, the narrative is very black and white, like these things tend to be in society.

Well, I think- Guys, aren't we talking about the... I'm listening to you guys. We seem to be talking about the bad actors in one sector and the things that they might do and the benefits of the technology.

It's never the technology's fault, right? It's always the people. And so this is just more of the same, right?

We're trying to say, hey, there's a group of people who are going to do some really horrible things with these things. There's nothing we can do about it, but maybe there is, and we have to start to look at that so that we can build towards Stacy's hope. I think we have this opportunity in these kind of collectives where we are talking about it and highlighting it to say, "Hey, world.

" It's not necessarily that the technology is doing anything bad. This is the people who are picking up the technology are still doing bad things with it. We have the same problems with people who drive and have road rage.

" The issue is the humans in the loop, not necessarily the technology. I think that there are all those bad examples out there that you cited, but at the same time, confidence in the justice system is low, and AI is not going to help that at this point. And then things start to break down even worse.

So I got to say that maybe we need to step up here as a society a little bit more in saying that this is not just a technology issue. This is going to be, how are people using this? And whether police officers are abusing AI to go track girlfriends and boyfriends or whatever else to crappy cases are going to be built on AI that we're going to wind up spending millions of dollars because somebody went to jail for the wrong issue.

This is crazy. Mike, if you're not thinking the wrong things, you don't have anything to worry about. Sure.

That's always the story, isn't it? As the taxpayer who winds up paying that guy millions of dollars who went to jail for the wrong reason, that money doesn't come out of thin air. It comes out of our pocket.

Yeah. Only the guilty worry about any of this, Mike. I'm being sarcastic for the folks that don't know Mike.

You're good. You? No.

We need a flock camera in front of your house. I want a flock camera in front of his house. But look, we live in a world where we are under video surveillance when we're outside a lot more than you probably realize.

And the only way to really harness and get a handle on that really is to have some form of AI. I had this recently. My car was involved in a hit-and-run accident by someone who was supposed to be cleaning my car, took the car for a ride.

And I was amazed when the police showed me what What roads are under video surveillance 24/7, and you just type in a license plate number, and it shows you every time your car's been on that road recently and what it looks like. I mean, the actual video of the car moving. And that's not unusual from what the detective told me.

And then they were able to trace it from that road to the Publix parking lot where there was a hit-and-run accident reported, and had video of that as well. And so yes, Mike, there might be some people, there's always some people who are wrongfully accused, wrongfully convicted, but there is a legitimate use of catching people who committed crime. Back to this whole lawyer thing, our system is built on the premise that it's better that nine guilty people should go free than one innocent person should go to jail.

And that's a tenet of the American justice system, right? And why we have such beyond a reasonable doubt, such a high level before you can convict someone of criminal. Other non-criminal cases, it's different.

Preponderance of the evidence and all of that stuff. But for criminal, it's a pretty high bar. Make sense?

Should be. It does to a degree, but I don't think that there's a lot of innocent people who are not going to jail. Seems like there's a lot of innocent people, or too many.

I mean, that just- Well, no. When the premise is better nine guilty go free than one innocent goes, even one is too much, right? But you kind of do the best you can.

AI is, again, don't blame the tool, as Stacy, I think, said. Blame the human. Speaking of humans, we're about out of time, humans.

It's right on the 45-minute mark. Guys, what an interesting, spirited discussion. Dave, you bring something to it, as always.

Thank you. Didisi, thank you. I hope to see you regularly on here.

Stacy, you're a breath of fresh air to have back on. Please keep coming back. John and Mike, as always, you do a heck of a job.

Thank you so much. Thank you for watching Techstrong gang. In case you don't know, we're live every Monday to Friday from noon to 12:45 Eastern Time.

tv, TechStrongTV YouTube channel, or on TechStrongTV OTT app, which is iOS, Android, Amazon, Apple, Roku. Whatever screen you watch videos on, you can catch Techstrong gang. We'll be back tomorrow with more.

We have TechStrongTV coming on, on this network if you're watching it. But until then, on behalf of the gang, I'm Alan Shimel. We're out.