Supreme Court Tariffs and the AI Governance Debate | TSG Ep. 1024

Hey, everybody. Welcome to the Textron Gang. It's Friday and went to the noon format.

I've been dying to say this, but this just in the Supreme Court is nullified the tariffs that the president has been putting out there for now, the better part of a year and a half. And a lot of that affects folks in the technology sector because well, vendors have to pay that money when they're trying to get components. And a lot of organizations that buy servers and compute and all kinds of networking gear that comes from overseas and, and sent aside money for this, and some of that money actually.

And those costs got passed along to the end customer. But John, you've been watching and following this whole thing for a while, but it seems like, uh, reason has prevailed here, and you just can't declare an emergency power randomly and cite some obscure act and then suddenly raise the cost for everybody. Yeah, exactly.

I mean, this is, as you said, Mike, this is why we moved the show back a couple of hours. This is breaking news. I mean, this is a, yeah.

Landmark separation of, of powers ruling by the Supreme Court. It was a six to three decision. And, uh, essentially our dismantling this tariff framework and, um, curtailing the use of emergency powers, as you mentioned, that bypassed the Congressional authority over trade.

Uh, I think Trump had invoked the international, I have to read this, international Emergency Economic Powers Act from 1977 that was used to levy sanctions by previous presidents. This administration used it to invoke it, to impose broad import taxes, and they cited these emergency, uh, national emergencies related to trade deficits and drug trafficking. And just for an instance of this, at the height of the policy, the terrorists range from 145%, in some cases against China to levies of 10% on, on most global imports, Canada and Mexico face 25 to 35% tariffs.

So in a sense, they've blown this up. Um, and it's gonna be interesting, as you said, the impacts not only on the tech industry, but I'm also thinking about consumers at this point. It's not very clear, um, what that would mean to consumers.

For instance, I have a friend who, who buys watches every couple of years. He bought one from Japan. He paid a $600 tariff for this Swiss watch, and now he's wondering if he's gonna get a refund.

I, I doubt it. But that, that brings up an interesting question. The, the ruling kind of leaves this federal government with this logistical nightmare.

So, um, the treasury's collected, they claim more than $133 billion of duties. And so, um, it's gonna be interesting to see how this falls out. This was a key tenet of the economic agenda of the Trump administration.

And so, um, from little of what I know so far, his reaction was he heard about it, and he, of course used the word disgraceful to de describe what the court had done. Um, we're in the, you know, process. This is breaking news, the process of reporting on this.

So I'm reaching out to a number of people to get more in depth, uh, response on what this means to not only consumers, but especially to the tech ecosystem. But, you know, we saw this coming. We were, there was a bit of uncertainty really, was I was doubtful that the court was gonna rule on this at all.

And it has, and in a sense, we can go down this road, probably maybe shouldn't, but it's gonna create this whole schism between Trump and Roberts, but for now, put to the tariffs. Yeah. But it was, I believe it was six to three, so it just wasn't Roberts that kind of carried this conversation.

Yes. Right. There were seven, I, I believe there were seven, um, uh, decisions that were written.

So there was a bit of disagreements, uh, uh, among even those who voted in the majority, they voted for different reasons, but the, the three who voted against it were Kavanaugh Thomas, and, uh, Alito. So, um, yes, six to three was a surprising, surprising edge. Hmm.

Jack, you have any thoughts on this? Because I know you kind of have a, uh, some political passions, shall we say? Well, Alan and I, Alan isn't here to get into it with me, but let's just say that, uh, everything right now I think is up in the air.

The Trump administration said before this ruling that if they don't, uh, if the ruling doesn't go their way, which obviously it didn't, that they would seek to impose tariffs under other, um, legislative authority that they have. Right. So there's a couple of others.

Now, I think the key difference between these is in the, using the I-E-E-P-A or IE, they were able to impose tariffs, um, on a nation by nation status. The, they in addition, have imposed tariffs on, excuse me, tariffs on specific industries using other legislative authority. And those, the Supreme Court said those stand, and they still have the ability to do that.

So that those things like semiconductor tariffs and steel and other industries, and they administration has announced that they're gonna continue to try to use that type of means to impose tariffs looking at an industry by industry instead of say country by country. It does, There's a, oh, I'm sorry, Jack, go ahead. I was gonna say, I was just gonna mention, I was gonna expand on something you said just because I'm, I'm glad you mentioned this, because the trade war is not necessarily over.

I'll just point out two things. So there's a, uh, uh, information Technology and Innovation Foundation reached out to me and told me that there were two avenues that the White House might look at to explore other avenues, uh, to expand tariffs. Um, section 3 0 1 1 22, the Trade Act of 1974, or section 2 32 of the Trade Expansion Act of 1962.

Just to, to, to follow up on what you said. So I apologize for interrupting you. Uh, no, that's fine.

I, and I was just gonna conclude with essentially what you said, which is the trade wars aren't over, there are still a lot of things to go. Now it's the Supreme Court punted on the decision of the refunds and what happens there. So that will probably not be decided until it gets litigated again through the Supreme Court.

So mm-hmm. You know, I wouldn't hold your breath that you're getting any money anytime soon. Right.

Um, You know, yeah. And, and from a, I look at this, you know, from a technology perspective, what does it mean? And right now, I don't think, it doesn't mean much of anything because the semiconductor tariffs still exists.

And I think that the Supreme, the, the Trump administration is really gonna work hard to maintain the tariff status quo as best as they can, and to continue to use it for negotiations and negotiation leverage. Mm-hmm. So you mentioned, I, I mentioned one more thing.

You said retailers, I mean, consumers. So you're right, it is very murky and opaque. I think Costco has already initiated, uh, litigation in lower courts to seek refunds, but the process of those refunds has been described as a mess.

So, um, that means it's probably not gonna happen for a long time if it happens. Well, John, you, you, you mentioned the mess. The, what bothered me about this whole ruling was the dissent.

It really bothered me when I read the dissent, because the dissent basically said, this is gonna be too big of a mess, so we should just let it go. It's like the Epstein case. Too many high level people are going to be impacted.

So we won't bother with enforcing the law. And that is what the Supreme Court said. I was shocked.

Mm-hmm. It's a mess. It's a mess.

That's not their problem to solve. Right. Their problem is to decide if it was legal or not.

But in the dissent, they pointed out, well, we believe it was, it's, it's legal to, to use this IPA law to, uh, initiate broad tariffs across everybody. And we think it's too big of a mess to clean up. It's gonna be too much chaos for them.

To put that in the descent really annoyed me because it's like, they're not, they're not acting as judges at interpreting the law. Mm-hmm. Not only that, I mean, how long has this been?

How ineffectual has the Republican Congress been sitting on their hands or taking very long vacations instead of trying to address these problems in the way it should have been addressed a very, very long time ago. So we have, They abdicated they don't do anything, essentially. They didn't do it.

It, it's a glare. It shows our, our, uh, holes in our system all the way from the Congress through the Executive branch to the judicial branch. When they say, we shouldn't deal, we shouldn't enforce the law, because now it'll be too messy.

Who cares? Yeah. The separation of powers is interesting because Kavanaugh, his dissent, um, refers to, uh, he says the ex executive's actions fell within the scope of, uh, historical presidential authority, which just sounds a little bit ominous to me.

Um, which means, uh, but this would, this didn't surprise me. I thought they were gonna pun on this. I didn't think they would rule on this.

They've been a rubber stamp for the administration, regardless of what's me. In a sense, Trump assumes that the people who he appointed the Supreme Court are his an extension of his personal, his law, his law firm. Um, and so for, for me, this was shocking.

And it was, uh, precedent setting. Well, three of 'em did. Three of 'em said it was okay.

That bothers me. They can't read the Constitution. And their Supreme Court justices, they're been able to read at the Constitution and interpret the law.

They didn't do that. Three of 'em, All three of them, and all three of them think that they're strict constitutionalist de boot. But, um, Congress, by the way, doesn't appear to be likely to take up this new responsibility in any meaningful way on behalf of the President, despite the current majority, at least within the Republicans, there's a large number of folks that are not pro tariff at all.

So, uh, Jack, do you think, you know, at the end of the day, we're gonna, let's see if we go with other rulings or other laws that we try to enforce the tariffs with, by the time we get all that together, you know, we could be looking at a different administration by the time we sort this thing out. What do you say? Uh, yes.

The short answer, yes, the long answer is, um, the word mess describes anything and everything revolving around this issue. Uh, the laws that Congress passed do give the administration a lot of leeway. The administer, the, you know, they give every, uh, ex they give the executive branch in general, a lot of leeway.

In this case, the executive branch took it. What I think the const, what the, the Supreme Court would say was a step too far, and they're reigning back. What the admin, the, the current administration says about it, um, the, the overall, from a congressional point of view, they're not interested in this issue whatsoever in terms of solving it.

They're interested in using it for political gain. And I think, we'll, I'll talk about that again in our next segment about some other issues. But, you know, it is very easy to have a slogan and say, you know, this is wrong, but not propose a fix to it, because it's easier to use that for.

I mean, realize, you know, in another week or two, we start the silly season of everybody running for the primaries for the 2028, uh, election season, right? So right now it's much more about campaigning and posturing than it is about problem solving. And what's Trump gonna do?

He's just gonna go, you know what, uh, that's fine. You guys can say that, but you also already told me that I can break the law off it's within my with if it's in my presidential duties. So why doesn't he just break the law and have the Treasury Department continue with the tariffs?

It's Like, What? That's not all the rights. Yeah.

That's not, that's not as, as, yeah, as crazy as it sounds. Well, at some crazy, as it sounds so normal, At some point, injunctions will be issued, and then some people will get arrested if the Treasury Department, if that happens. But it will take a while for that to play out by who The DOJ Pam Bonde come on.

They'll have to send some, they'll send, Congress will send over the Marshal to actually put some cuffs on somebody sitting in the Department of Treasury. So none that It will not happen, At least under the current laws, that's how it's supposed to play out. But, um, underneath all of this, I think a lot of it execs and business execs in general, Jack, maybe correct me if I'm wrong, but one of the things they didn't really like about this whole thing is that there's too much uncertainty.

You wake up every morning and you didn't know what the tariffs were gonna be, and we couldn't plan anything. And it, you didn't know what your bottom line was gonna look like. And frankly, this whole approach where we are using tariffs and businesses as hostages to advance foreign policy goals, it's not the right way to go for anyway.

Uh, there are arguments to be made both ways, but I think you're, you're a hundred percent right that the big issue is the uncertainty. And right now, unfortunately, Supreme Court ruling doesn't fix that one iota. Right.

We still have a whole ton of uncertainty. We don't know what Trump or the administration's gonna do. We have no idea what's gonna happen with, uh, refunds.

And, uh, it's really a wait and see. Now, I think anybody who actually makes a decision and takes action based on the Supreme Court ruling today is really asking for trouble because it could all change tomorrow. Mm-hmm.

Right, right. That's, that's what's, yeah. I'm glad you, you made that point, Mike, and, and, and Jack, it's just, it's this maddening idea.

How do you plan anything, um, given the unpredictability, and even if by cozying up to the administration, it really doesn't view any goods, um, he's gonna act the what's most prudent and expedient for himself to the exclusion of others. So, um, it bears watching. Mm-hmm.

Meanwhile, do you think that the folks who manufacture goods overseas are looking at this going, well, thank God for that. At least we can now maybe, you know, compete in that market again, and We'll, we just go back to, who knows, John, what do you say? Business as usual.

We'll just keep buying servers from where we used to buy 'em anyway. Oh, boy. Yeah.

That's, I mean, you know, for, as you think of these global trading partners, including the eu, Japan, and South Korea, maybe this is a moment of relief, though I would caution them that the, was it, the Congress Congressional Budget Office warned that the long term economic impact of the tariff era, it's gonna reach $3 trillion over the next decade. So, um, it's chaos. And I also think it's, I, I'm, I would be stunned if, if there's not another avenue pursued to, to continue the tariffs, and I, I'm almost, I'm 99% certain this, this will continue in some sh form or some manner.

So, um, no, there's no certainty whatsoever. There's no relief whatsoever. Mm-hmm.

But to, but to your point, Mike, the, from a, a technology perspective, if you are looking at it, if, if you're an exec and saying, okay, we're about to do a buy of, you know, half a million dollars worth of servers, a million dollars or $20 million worth of stuff, do you wait because you think the tariff's gonna change it? The answer is probably not. Because the percentage impact on waiting, you know, the percentage impact you might save as the tariffs roll off versus the cost of your business to waiting three months or six months for all of this to settle out is huge.

And so, you know, you really have to think about it as an ROI equation and say, what's, what's the true impact of tariffs on my supply chain and what I need? Now, of course, you know, it's different if you're manufacturing and these are raw materials that you're using, versus if this is, um, just business equipment, right? So if you're an IT shop and you're saying, okay, it's time to refresh our laptops, or we need to do a big buy servers or storage, you might as well just go do that because you may save 5%, or it may cost you an additional 5% if you wait.

And the cost, what is the cost of waiting six months to do your refresh, refresher your buys, right? So it's alright, I, I, I might go the other way and I might reduce my current order, and on the assumption that, well, what's the minimum number I can get by with today while I wait for the prices to come down later? Right?

And then, and, and then of course, you always have the risk that, well, DRAM is now gonna be even more expensive simply because the AI vendors are gonna suck it all up anyway. Well, there's that too. All right.

And speaking of ai, let's leave this where it is because, well, like most breaking news, it's still developing story, and I'm sure we'll be talking about it again in there days and weeks ahead. But let's jump over to some kind of things that maybe, um, less argumentative. We shall see.

Jack, you have an article on Security Boulevard commenting about a post that was on salon where somebody was arguing that AI and AI agents will soon leave to, to the destruction of democracy as we know it. And of course, you've said basically that was a whole lot of chicken, little arguments. But walk us through what's going on here?

What is the concern, what's real and what should we be thinking about here? So, so Salon reprinted an article from, uh, a researcher, a set of researchers who were bemoaning the fact that, uh, they, they did a study of social networks, and particularly Twitter X slash x, where they found, and they were able to create, they found some, and they were able to create their own AI driven bots to driving increased engagement. And from that, and the way they push created their study and their, their little experiment they declared, they, they raised their hands in the air and said, let's throw a banner headline out there and say, AI is a threat to democracy.

And this comes hot on the heels of all of these articles saying, AI is taking our jobs and AI is doing this. And it's, you know, it's the, it's the, the end of the world as we know it in one form or another. And you know, Alan's been talking about this, and I'm sorry he's not here because I saw he wrote a column the other day and he, you know, sorry, he and I were right along the same lines talking about the fact, the exact same things.

Well, you know, we've had many technology revolutions. We had this, you know, the printing press, the steam engine, the industrial revolution, transistors, all of these things were predicted to end the world as we know it, and they were all proven completely wrong. And part of that is because humans are incredibly adopt, uh, adaptable.

And part of it is because yes, the jobs change and the nature of the jobs change, but when you increase worker productivity, you unlock things that we never thought were possible before. Right. And if you think about, you know, we, when the printing press came along, all of the scribes who were hand scribing and copying and creating extra copies of books thought that, um, that was the end of their job and the end of, you know, the world as they knew it.

And now, you know, a hundred years later, everybody was writing books and reading books, and it's sparked an entire knowledge revolution that we haven't seen up until the age of the computer. And, you know, when computers came out and exec, you know, all of the secretaries now, no. You know, what happened to the typing pools that we had in the 1960s and seventies, right?

They're all gone. And in fact, uh, I was actually talking about this with my family yesterday, that the, the age of the administrative assistant is even gone because we don't need, uh, a, an executive assistant to manage an executive's calendar anymore. It's too easy for the executive to do it with the computer tools he has, or she has.

Does that mean that the, the, you know, the world comes to an end? No, it's increased productivity allows people to do other things and create things that we never thought possible. And I'm, I'm just really sick and tired of the world is coming to an end because we've got some new cool technology here.

I mean, let's just be real about these things. Well, let, Let's, let's take this apart though. All right.

So what they're arguing, I think, is that they're saying that, or if we go back to Jefferson, Jefferson would say that, you know, the foundation of democracy is that if you put a reasonable amount of information in front of their reasonable average souls, they will most often come to the right decision. Um, but John, I'll throw this in here for you, it's like, well, how do I get that reasonable amount of information into their hands of those reasonable people when we have these AI agents creating a massive amount of noise that is far from factual and closer to propaganda, and nobody can tell what's real and what's not real anymore, Right? Right.

So, so, you know, on principle, I I, I agree with, with Jack, I mean, people are very good about adapting to technology. They're always resistant to it because they feel it's a threat to them, but then they learn how to work with it or work around it and enhance themselves. And those who don't, I'm not gonna be overly dramatic perish is kinda the way I was thinking in terms of professionally.

So, so as you mentioned, uh, Jack mentioned this, uh, this column that, that Alan wrote the doomsday column, and I mean, it was a little bit tongue in cheek, but it does address this idea that we're hollowing out these kind of, these jobs in the middle. And I do think, given what we're hearing from the tech executives, and they're not helping matters at all, by the way, when they make these predictions from Microsoft or, or Salesforce or Nvidia, how the, the workforce is gonna be completely reshaped. What they, what they first told us was that we were going to, and Amazon, where we were gonna replace some human jobs with robotics, or we were gonna create these, these armies of AI agents to do these tasks that were repetitive, which in a, in a sense, is gonna affect a lot of people.

Um, they, they were, they were the messengers that created a bit of a panic. Um, I think we're not at that stage. I don't know, nowhere close to that stage.

I think there, there is gonna be a seismic change in terms of the workforce over the next couple of years, next three years. I've seen some estimates of up to 2 million jobs being kind of displaced or changed in some manner. But it's also, we're gonna create a number of jobs we don't know exist now.

We're gonna, we're gonna see this whole new era of things done differently by different people and different expertise and, and skillsets. I think one of the, one of the problems too, in addition to the executives hyping this and scaring us, is they've done a really damn poor job of, uh, upskilling us people on the trenches. You know, people most, for the most part are, are being told, oh, ai, you need to use this.

But they're not being trained. They're not being trained how, how this is gonna happen. They're, they're encountering disinformation when they use, uh, chat GPT or Claude in some instances.

Um, they're, they're, they're getting slop. They're getting, they're getting misguided information or advice outright bad advice. So they, we have all these factors moving kind of against one another.

And I, I just think it, it creates all sorts of anxiety that is being fed into by this AI race where everything is accelerating. 6 just came out with 12 days after, you know, a previous iteration. I mean, we're just being overburdened overwhelmed with this.

And I think we have to just, we want, we feel like we have to play catch up or at least be in some sort of position to, to absorb this. And I, and I, I, I understand, I understand the fear and anxiety, but I think it's overcooked at this point. For now, at least I would look at it this way.

It is possible that democracy might save us from ai. Think about it in another way, go about this, is like, if I lay off, you know, half the people in a particular state or can't find a job, well, they're gonna vote in a certain way and enacting certain rules and regs to protect themselves because the ballot box is their last line of defense. Tracy, am I crazy?

Uh, maybe a little, Maybe was a little, uh, I don't have that much faith. Um, but you know, the article is, is re I, I mean, I think that, let's use the word democracy to really unders the way I read the article to, to understand what they're, what they're trying to say. And last week, I think we had a discussion on, is AI the new, um, immigrant, right?

Is that the new enemy to our jobs? But in many ways, a and people don't understand this, but in many ways, AI actually cre democratizes so many elements of our economy medicine, right? Mm-hmm.

Um, it democratizes even software development. More people are gonna be able to do software development that aren't as smart as the people who did it. Would, it would, they had to code every single line, right?

So it may open up opportunities for people because it's democratizing, it's lowering the entry point for some of these harder jobs. I think what happens is, when we call it, because it's called artificial intelligence, it sounds hard and scary, but to be honest, this is probably the easy, the, the reason why adoption has accelerated in the way it has is because it's got a pretty low entry point to start using it, a very low entry point to start using it. Unlike previous technologies that were really hard to, to to use.

So maybe AI is democratizing some jobs and allowing more people to get into those industries. Now, everybody can't be a software programmer, but there are other areas that AI is helping writing, uh, creating art. Uh, there are, there's it coming up with recipes.

I use it to create my own crazy recipes 'cause I like to cook. So maybe we have to start looking at it in a different way. I don't think, I think the doomsayers are going overboard.

There are plenty of things that's impacting our democracy right now, uh, such as, you know, getting all of the billionaire jerks who are taking out things like the Washington Post. Come on Jeff Bezos. Really, this is what, this is how we save our democracy, is identifying what's really impacting it.

AI is not this point. Mike, two things on that. And, and Tracy, I think you're right.

In one sense, and I alluded to this in the last conversation, which is there is a very strong political component to this right now because the we're in, and, uh, over the last year and a half, the Trump administration has done a very nice big bear hug to the leaders of Silicon Valley and the, particularly the AI leadership, right? And saying, we give full throated support and we, uh, want to continue to not regulate or not heavily regulate AI industries. So if you are a, not a Trump supporter, as the Democrats tend to be, and you want to find a hook to say, I'm against Trump, one thing people are doing is saying, AI is bad.

AI is coming for your jobs. And if you look at the statements and what is happening from the Democrats running for office this cycle, there is a huge negative AI component. It's making energy more expensive.

It's taking your jobs. We have to regulate it. We have to prevent kids from it.

There is a proposal being floated right now to have put a moratorium on AI dentist centers that you can't open a new AI dentist data center for a year, right? I mean, that, and, and to me that's like saying to, uh, Henry Ford stopped production of the, you know, the, the, the, the, the automated, uh, assembly line was taking away too many jobs. Therefore, we can't, you can't build another automated assembly line.

You have to stop selling cars that people want and stop commerce because we don't, we, because we wanna use it as a, a cocktail to beat up our opponents, right? I mean, it's, the political component is, is a huge part of what is driving the fear. And it was my Secret could get away.

Yeah, we that, yeah, we talked about that last week because, you know, since we talked, meta has made a, uh, $65 million, uh, commitment to, to, and I think it's, it's, it's around the midterms and beyond. They, they are working. They have two pacs, they're working with one that's a Republican PAC and one that's a Democratic PAC on, on influencing AI and the influence of ai.

They don't want regulation. Obviously. Anthropic on the flip side, has got their own, their own agenda, which is a little bit more AI safety or ai, uh, responsible AI use.

Um, so, so yes, it's AI is gonna become, and we, I I know we talked about this last week again, but I know AI is going to be more of a topic during campaigns, um, especially during the economy and concerns about the economy and about affordability and job certainty, and as you said, energy consumption, uh, land use, uh, energy bills, uh, et cetera. So you're right, it's, this is, this is, uh, gonna only be rushing it up. So here's David hopeful about, imagine a world that goes this way, right?

Because I would argue that even before ai, the average voter and the average person was being spun out of their shoes by the left and the right, and all kinds of people with all kinds of agendas. Imagine a world where you each had your own AI that helped sort out all that crap and kind of gave you the actual, you know, facts on the ground that matter versus all the crap that comes down of all these spin factories that winds up just trying to confuse people and, you know, convince them that something that they're for is actually something they're gonna vote against because of the way I worded the ballot. Tracy, what do you say?

That's what I'm saying. AI can democratize things, right? That it can bring things down to a level that everybody can understand.

Um, so maybe the Democrats, because many of you know, let, let's, let's look at attorneys for example. Do you think I call my attorney when I have a new contract? No, I cut it, I just push it through chat, GBT and then Claude, and then Jim and I, and I get a consensus between all three because I don't trust a single one of 'em that the, the particular paragraph I'm worried about is, okay, right?

So now I don't have to spend money on an attorney. That to me is freedom. It's liberation, right?

I'm a small company, I don't need to call an attorney, and then they come up with 10 or 15 new issues that I have to be worried about just so they can justify their job. So yes, I think that it, it played the right way with companies who believe in responsible ai, getting rid of some of the hallucinations, addressing some of the issues that we have. AI can provide a, a better answer to people as long as they turn off their history and they're getting the answers correctly.

And it's not just reflecting back what they wanna hear, which is what it does. Now, Tracy, I'm, I'm, I'm so so glad you brought that up, because my father is, uh, one of many who says, you know, I hate talking to my doctor. I'm gonna go and search GPA search theis for what, you know, what ails me and what meds I, what, what meds am I taking, what should I be taking, yada, yada, yada.

And, uh, being concerned about that, my wife said, let me look and see what you're doing. And she found a site specifically for that issue. I believe it's called consensus.

That the idea is that you ask it a question and it goes, and it fields the question to a whole bunch of ais and synthesizes the answer and says, what's the consensus among the five or 10 or whatever things that I've done, right? And, and it's an AI can see, you know, building that consensus out of that. And, you know, it brings me back, I, so I I love that you do that manually, right?

And now there's tools that automate this, right? Uh, and, and you know, I, I think this part goes back sort of to Mike, you had the question earlier about, you know, where this all started with Salon and the actual article and talking about disinformation and, and, uh, you know, AI manipulating the, the forums, right? The Twitters and Public Opinion, and Oliver Wendell Holmes, I believe it was our Supreme Court Justice, uh, coined the first phrase, the marketplace of ideas.

And the question really is, can the marketplace of ideas survive the flood of information from ai? And, you know, I believe, you know, I'm, I'm very much a free speech ols and I believe in the marketplace of ideas, but I am concerned that AI may make it harder to understand the consensus and what is really happening and what is real or not real. And then, well, I wouldn't find that tool because I, my prompts would be a lot easier because I won't have to ask chat GP to go ask Gemini and Claude, Oh Wow.

Let's, let's be extremely honest. 99% people probably already made up their mind about an issue long before anybody ever sent them an email or anything about it. 'cause they're just kinda like looking for things that confirm their bias anyway, so, mm-hmm.

Yes. Hopefully, hopefully, you know, AI will, you know, straighten a lot of that out for folks and maybe we'll see some more rational decision making, but I'm not sure ai, no matter what changes the human condition. All right, let's jump onto to our next topic, which is a little bit more ai, been a little more practical, but GitHub is now talking about how they're making AI agents available to help manage your software repositories.

Tracy, is this a good thing or a bad thing? So, you know, we'll see that's what, that's the answer to that. Eventually it will be.

So for those of you who don't have GitHub re repost to manage, um, if there's a lot of busy work, if you wanna keep your repo clean, um, there's documentation you have to update, there are issues you have to track. There may be old issues that nobody's responded to. You try to get somebody to look at.

So there's busy work when it comes to, um, managing a get repo. It's just, and that's the best way to describe it, is busy work. But there are a lot of people in this industry that likes that busy work and sees it as a level of power and control.

It's true, I know is true. We, this is, this is one of the, one of the things that we do, one of the things we do, we like that kind of power and control. So for, for a company our size, or even for our open source project, orus, I would so use that because I don't hire somebody to manage repos.

I would prefer somebody that to be automated. And the way they're doing it by just opening issues or pull requests or comments is a really low bar. It's, it, it actually does work that, um, nobody really should have to worry about controlling, but developers right now don't necessarily trust AI yet, and they don't trust it because some of the tasks that it's taking over have to do with our culture and who's in control and where the power is.

So there's going to be a push and pull between, and this is probably across every industry, uh, for some of these remedial tasks that are remedial, but they're, there's really solid decision making going on in the process that they don't want to hand off to ag agentic workflows. Mm-hmm. I would use it, but I don't think everybody will.

Do you, are you concerned, and we've heard stories about this, where somebody, you know, woke up one morning and the AI agent deleted their entire production database. So, you know, is it gonna be possible that an AI agent's gonna accidentally delete something Jack and we don't have a way to recover it, or maybe we should figure out how to recover something before we turn it over to the AI agent. What do you say?

Well, that, that specific use case always bothers me because humans are much more likely to fat finger your database and destroy it than I, than an AI is. And if you haven't protected yourself from your human users, right, you know, then, then you, you deserve whatever you get, right? If you don't have your backups and your data protection in place.

That being said, any tool that automates, whether it is, uh, AI in GitHub or it's infrastructure's code or any of these other things, you have to have the appropriate guardrails on it and say, what can it do? What can it do? How does it do it?

Where does a human need to be in the loop or not need to be in the loop? How do you audit it and how do you review the changes, but the, the, the power, literally the power that you get from deploying automation. We've long recognized that that's why nobody in the DevOps world has an objection to infrastructure's code anymore.

Because we would much more be happy saying, Hey, look, I need eight instances of this, or I need the network to run it. You know, this, you know, I need to be able to talk to this guy over a hundred gig ethernet link rather than sit there and remember what Cisco command it is to get to that particular router to do that particular task on that obscure configuration setting, right? Nobody wants to do that anymore.

And the same thing as Tracy said, nobody wants to sit there and manually do all of these boring, repetitive tasks. And so this automation is good, and if you put the right girls guardrails on it and you understand how it works, I mean, nobody, nobody hand compiles assembly code or, or bin to binary code anymore. And nobody hand assembles, you know, any of this stuff.

That's what the computers do. And we've automated all of this stuff. So we just have to continue doing it and put the right tools in place around it And GitHub's making sure that has those guardrails.

It's all based on a pull request or a comment. And right now, that's where it should be. I'm hopeful that something good could occur that looks like this.

And again, feel free to call me crazy, but if I am, but we've been kinda unhappy with the fact that all these repositories have a lot of code in them that shall we say is not of the highest quality, contains a lot of vulnerabilities, and then we create all these DevSecOps issues that ultimately wind up with vulnerabilities in production environments. Is it possible, Tracy, that these AI agents will kind of improve the quality of our code because they'll actually go into that code and find those vulnerabilities and maybe even fix 'em for us? What do you say?

Garbage in, garbage out. I mean, they're, these, they're pulling from code that's out there that's bad. So unless we all go and scrub all of the, basically the entire GitHub, GitHub global universe so that these LLMs have super clean co code to build from, it's, we're gonna still see the same thing.

I don't, I don't see that getting fixed and they're, I mean, I look at our own repo, we need to sit down and delete. So we need to archive some of these repos off because we get vulnerability scans against this stuff and it's so noisy and we don't even use it, right? So this, when I get something that comes through that is truly a vulnerability, sometimes I miss it because I've gotten in the habit of deleting my, my sneak notices off the open source repo, right?

Because it's too much, it's too many. I I I can't look at every single one. And, and some of 'em are in the repos we don't use, that's why, that's why deploy hub's all about post-deployment vulnerabilities.

Let's catch the 20% that are really, uh, in the attack surface and not worry so much about the ones in the repos. I don't even use All. So Jack, what do you say we need, like delete your repo day every 30 days.

We should all just sit down and say, we're gonna delete repos for fun. Uh, heck yes, brother. Absolutely doesn't matter.

We do. There is, unfortunately, it's, it's the nature of the beast. As as, as an engineer, you write stuff, you use it for a little bit and then you say, oh, that was complete garbage.

It doesn't do what I wanna do, throw it away and do something else. But then you, you never really throw away or delete anything. It just lives on forever.

And 'cause the, you know, and this is, you know, as a side note, this is just a challenge in general for computers and storage and why storage never shrinks. It always expands. The cost of making a decision to throw away or keep data is much more expensive than just buying more storage and keeping it, particularly with GitHub.

'cause you're not really paying for the storage. And so we just keep this stuff forever and we just never, you know, that module lives in the middle of the code stack that we're using, but it just isn't included in the compilation, right? So there's, you know, there's 20 files that, that, you know, you, you make a request, you download the code and you look at it and you're like, I have to go through 20 files, but only one file is used because all the rest are, were thrown away at some point or another, but they just live on forever.

It would be great if we, we could afford to do that. It's our human nature. George Carlin used to do this whole thing on storage, storage container units.

You know, we buy, we buy storage container U units to store stuff. GitHub is just a giant stuff, container storage container unit, right? Right.

We just put stuff in there because we, it's too hard to make a decision if we should delete it or not. Should it be thrown away or not? It's easier to put it in a storage container.

So I just don't wanna make the decision, is what you're saying. So I just keep buying more storage containers and then I shove it in the attic or their basement or wherever until somebody inadvertently trips over them one day and winds up, you know, using some code that probably, you know, is outdated, shouldn't be used and fundamentally insecure. So what can go wrong?

Tracy, I gotta ask you something though. It's a wonder that anything we've built so far actually works. So how have we been holding this together?

A lot of good programmers looking at their code, code reviews, right? Testing, uh, new tools in our DevOps pipeline, signing attestation, you know, refactoring code. We, we do a lot of work to clean this stuff up.

There is no doubt. And does anybody actually enjoy that part of the job? Or is that the whole toil that they just assume have an AI agent do?

Well, I think the companies that are in the business enjoy doing it, and it solves some problems for the companies who are not in the business and they don't wanna delete all that crap code and they need something to some, some tool to automate checking it, right? So it creates an industry certainly does, but it doesn't help the LLMs, it's not gonna help Claude write better code because we still, they're still looking at, you know, repos that have bad code in them. So how do you fix, how do you fix that?

Claude's gotta get better at writing code without having something to look at. Alright, I think we've come to a conclusion here, folks. If we want better software written by ai, let's kill a few repos and see what happens.

I I want to thank everybody as always for sharing their knowledge and insights. I wanna thank you all for spending some time with us. Once again, please stay tuned.

There's a whole text on TV lineup right behind us. It's the replay from this morning stuff. But hey, just like this content, it's awesome stuff and, and that Happy Friday.

Have a great weekend and we'll see you all Monday. Take care.