Barron's Streetwise - A Compost Man Talks A.I. Vibe-Coding. Plus, Agco’s CEO.

Episode Date: February 20, 2026

Jack and Jackson discuss what’s ailing software stocks, and list Wall Street picks. And a tractor executive has the latest on farm incomes and planting tech.  Learn more about your ad choices. Visit megaphone.fm/adchoices

I should be sitting up straight to help with my voice, but I'm too lazy. Is this, does it, does it come through in my tone that I'm slouching? You can give that table a couple cranks. You just crank it up a little bit, a little higher so you can sit up straight. Higher, all right. I'm going higher. That's what I do. Up periscope.

Hold on. Barons can get you, and after all these years, an electronic table. No, no, no. I see you cranking it up. I'm not looking to break the budget over here. This is not an ergonomics podcast. This is the Barron Streetwise podcast. I'm Jack Howe.

Longtime listeners will know you as Jackson Cantrell, and you left me for garbage. We've talked about it. I'm not bitter. I totally understand. You were an audio producer in this podcast. You've popped back over here to help me for a few weeks. But, you know, we're going to hear about, tell us what you've been up to. Well, it's not garbage. Come on. It's industrial composting.

Food waste to compost. We collect compost from hotels and stadiums and grocery stores about a million pounds a month. And we have these like 53 foot long machines at our warehouse just south of downtown L.A. where we turn it into compost. And what and the compost is used for? Right now it will go to sort of small family farms. It was interesting to talk about the beef episode because it's very similar where you have like medium-sized ranchers who are out in what we call the high desert out beyond the mountains in L.A. They'll use the compost as organic fertilizer. We also send a lot of like expired vegetables to them because they'll use it as substitutes for some of their animal feeds. So they'll have like hogs and cows. I know I didn't mean to make light when I said you left me for garbage.

You also left me for rotting vegetables. Okay, I've got it. I've got it. Now, look, I want to hear you have direct experience with something that I'm very interested in right now. It was something that many investors are interested in. And that is software and coding potentially being replaced by AI. And that has led to a severe sell-off in software stocks, what Wall Street is calling. There's been a lot of wordplay.

The Sasaker. Right? The Valentine's Day Sassaker, although it started well before Valentine's Day. Folks have been saying the Sassapocalypse, but I think I like Sassiker comes from you. I think I stole it from a research report from BTIG. I don't know if it was original to them or I don't know if it's been going around. Let me tell you what I touched on in the column this week, and then I'm going to talk to you about your experience. And this comes from a piece of research from Deutsche Bank. They wrote recently the software stocks are down 19% since the beginning of the year, and multiples have compressed. That's a fancy Wall Street way of saying the price to earnings ratios or price to free cash flow or whatever it is has come down.

And it basically says they're searching the group looking for good deals, names that appear oversold. Here's what they write. They say it's difficult to disprove the bare narrative in software because the fears about Gen A.I. have more to do with the out years. They say any meaningful disruption will likely play out over a much longer timeline than investors anticipate. Quote, we remain steadfast in our belief that vibe-coded replacements for package software are unlikely to gain traction anytime soon. I keep saying jive-coded, it's vibe-coded. We'll talk to you about what that means in a moment. Deutsche Bank goes on to say, we are more concerned about switching costs and software being eroded by cheaper code generation, reduced integration costs, and easier change management, which could impact software economics going forward.

Sounds like a little bit of hedging. Sounds like they're saying, this isn't something we think will be a big deal or it might not happen for a long time, but it could sort of happen. And here's ways it could influence the economics. I feel like that's common in these sorts of notes. I like to be right either way no matter what happens. But I mention this for the stock picks, because what Deutsche Bank did is they did kind of a screen of their whole universe of software stocks. And first they focused on what they call barriers to exit and the ability to adapt to the Gen A.I. In other words, these are the companies that they think are safest from this.

And then they took that group and looked at their buy rated stocks and they overlaid a valuation screen. And they came up with four names. And I will give listeners these four names if all of you promised to do some further research on them. This is not me saying let Little Timmy's entire college fund ride on these names. This is Deutsche Bank saying these stocks were attractive and me saying this is a list of companies for further research. The first one is called Celebrite. The ticker is C-L-B-T. I'm looking at the description.

It's got a lot of these multi-syllable words that you string together and they don't mean anything. Empowering organizations in mastering the complexities of legally sanctioned digital investigations by streamlining intelligence processes. What does that mean? Is that, that's law enforcement stuff, Jackson? I think so. I think it's like government data. That's an Israel-based company. The second one is CRM, of course, that's Salesforce.

I-N-T-U is into it. That will be intimately familiar for anyone who's been turbo-taxing recently. And service now, that is for, oh boy, end-to-end workflow automation platform for digital businesses. So it could be a hot dog stand, anything. I don't know. What does that mean, Jackson? Jack, they're empowering both customers and employees in every corner of your business. Yeah, I recognize it as just a general software name that's widely used by businesses.

I see the name around here and there. They use the acronym CRM. That's also the ticker for Salesforce. It stands for customer relationship manager, which again, could mean anything. But usually it's like software that helps you stay in touch with the people that you're trying to sell to or the people you're trying to service or what have you. Basically database software that helps you make more money. Okay, so those are four stocks that Deutsche Bank has put on its second.

Massacre shopping list. Those are my words, not theirs. And by my words, I think I mean BTIG's word. And that's where I want to turn it over to you, Jackson, because I have never coded and I've never vibe coded and I've never jive coded. And so you have coded. You've done some coding in your industrial composting startup. Start by telling me what kind of coding you've done over the past year. What have you made? What kind of software are we talking about? Well, the first thing we did was make dashboards for our customers so they could see how much food waste they were composting. And then as AI tools got better, we ended up building Android apps for our drivers and AI pipelines that analyze waste bills. And now we have a computer vision program that identifies food waste from pictures.

I had many, many years ago someone paid me a backhanded compliment. they said, you know, you're not a bad writer for a stockbroker. Have you had someone say to you, you're a pretty good software maker for a compost guy. Has that happened? Yeah, I mean, it was just kind of based on what our customers were asking for. So after we built this, you know, way for them to track their food waste, they wanted us to also track their recycling and trash and e-waste. And so it was...

So they're coming to you now for the software. not just the compost services. Yeah, and we actually have some software-only customers, too, who we aren't able to service, but we can help them track their composting with other providers. What did you use or how did you make this software and how has that changed? Yeah, so I could say pretty safely that about a year ago

I was absolutely hopeless when it came to coding. I'd taken one class in college, which culminated in me spending, about 120 hours building a buggy version of flappy bird. Is that good? Is that a good thing? But that's, no,

it's like a little game. It was coding in, oh, buggy. I was picturing a buggy like a cart, pulled by a buggy, meaning a lot of bugs. I barely passed that class.

So about a year ago when these AI tools were taking off, I think they'd been around for a little bit. They were mostly like auto complete. I think of, out this in like the levels of self-driving where level one is hey I can see you're writing this function here's how you can finish it up so it won't break or here's where your code's breaking here's what you can use instead and that sort of tool was useless for me because I didn't know what it was you didn't have enough to to auto complete no not not at all and so we're really started for me

is when the tools got good enough to build front end code so we were building this dashboard and how it was working is I was going back and forth with a contract developer. I'd say, hey, can you build a dashboard? And then he'd send me his version and I'd say, oh, I want this filter or this button over here. And then the note would go back and he'd change it up a little bit. And that process would take weeks just because, you know, this guy's a busy guy. And, you know, I was kind of a little nitpicky. Did you try saying, hey, you don't know who you're dealing with here, fellow. I'm the creator of buggy flappy bird. Did you try that? No, that would have been good. But instead, I just started building the dashboard myself.

And it wasn't connected to our data. It was just the front end. But I was able to just sort of whisper into this machine and it would come out with a result that was, you know, much closer to what I wanted. And if I needed to make a tweak, it would happen instantly. And then I sent it to the developer. and he was able to plug that into our data in about an hour. What was the tool? It was called V-Zero. So Vercel is like a company that helps you host your website and they do a bunch of other things.

But they had this AI front-end coding tool. And is that where you are now or you've progressed since then? That was like six, seven months ago. Oh, a lifetime six or seven months ago. Yeah. I was still listening to A-track tape. back then. Yeah, that was level two. Now, after that came what I call level three, have you heard of this tool called cursor? I'm guessing it's not a person saying bad words, right?

Because I could make you a tool like that. So it's, you could just push a button and a 53-year-old man says, what are they talking about with all this AI? Yeah, yeah, yeah. Basically just a chat bot inside an IDE. So that's what developers use to look at all their code in one convenient place. What does IDE stand for? An integrated development environment. How this works is you chat with the chat bot and then it writes all the code for you. Wow.

And it could write like front end functions, functions that talk to your database. And here you're able to actually create working apps. How dumb can you be? I mean, let's not say dumb. But I'm thinking, can I get on this thing, never have encoded a line in my life, and tell this thing that I want a piece of software? Could that happen? Or you have to have some basic knowledge? I think for this, that tool specifically, about like four or five months ago, you had to have some knowledge.

Because this tool would never tell you that you're being an idiot. You know, if you asked it to do anything, it just wouldn't push back, even if your idea was, like, inefficient or wouldn't quite work. And another thing that it tended to do was it would write brand new code anytime you asked it to do a task. And in development, that's not ideal because it could be writing the same function to accomplish the same thing in five different places in five different ways. So you said four or five months ago. Where are we today? What's the best and brightest out there now? So you're familiar with Claude Code and how that was like all in the news around Chris. That's the company Anthropic. They came out with a new model called Opus 4.5, and it really surprised people because it was

the first time, as you said, you could know absolutely nothing about coding, just type something into their interface, and you'd conjure a software product into existence. So this is something, Jack, 100% you could build pretty much whatever you wanted, as long as the scope was fairly small. I could make a program that would do anything I wanted to do right now using this thing. You could do it and you would never have to look at the code. Good Lord. I mean, what kind of software improvements do I need around the house?

Now I'm trying to think about how can I improve my everyday life with this thing. What do you have coming up? What are some big things that you have to, like big chores you have to do? I'm almost out of firewood. I had such a pile of firewood, but it's been a cold winter and we've been going through it. I've been hoping that a storm would come along and not knock out my power, but also take down like three fat red oak so that I would have something to chop up for next winter. I think you need an axe swinging robot for that. And I'm not sure I want it. Tesla, Tesla is coming out with one of those soon. I don't think, I'm hoping they're not going to give it the axe on day one as a as an attachment.

But, you know, let's let's start with maybe like a vacuum cleaner or something like that. Anyhow, so that was what, level four? You're talking about it in terms of like robocars. That's level four. I guess what do we go up to? Five, like where would you say we stand now? And what do you think of Deutsche Banks take? How much of a threat do you think this is to the software industry?

Yeah, I'd say we're about like a 4.1. And I'd say mostly agree with their take that no one's replacing their enterprise software with this kind of stuff. But I think it has made development just so much faster. I imagine we could get to a future pretty soon where a team of five, maybe 10 people could build very similar software to a team of 500 people.

And that's where I think maybe not some of these whole companies, but some of their products might be at risk. And this is why some of many people are talking about what AI means for jobs. And are we ready as a society for what will happen. One last thing. You said you experimented with level five of what you're talking about,

and it involves Wix. What are we talking about here? Basically, what I call level five is when you scale from one agent, you know, doing some coding tasks for you to a whole team of agents. Or in this case, I was using a whole town of agents. So there's something called a gas town that this guy, Steve Yeggie made. It's based on Mad Max, if you know that story at all. I'm hanging in there. So I can't recommend this because it's quite expensive for the average, you know, layperson

experimenting with coding. But basically, it spins up different agents and gives them instructions where each of them has a different job. So in this case, you have the mayor of Gastown. you have a bunch of pole cats who each make a feature. You have a dog that goes around and wakes the pole cats up if they get off track. And you have a witness writing everything down. These are all AI agents that are involved in the job of making software for you. Because we could be talking about a Lego movie right now for all. Yeah, exactly.

So each of the agents are telling each other what to build. So you're out of the loop at this point. You just give it like a plan and this thing will just, this town will just work all night. What got me using this is I realized that for the last 10 years, I had been paying Wix, I think $140 a year to host my personal website. Wix is a company that hosts websites? Yeah, so Wix makes very easy to build low code. It's called a low code platform where like anyone can build a portfolio. website. And so I had my portfolio

website and I was like, well, one, I haven't updated this in a couple years and two, like, what the heck am I doing? I code with AI every day. This is going to be really easy. And I wanted to try out GasTown because I had heard about it. So I thought, what could make this like portfolio

kind of fun or unique? And I wanted to go back to my college radio days. So I came up with like a fictional 2002 radio station based in Southern California where I... Let me make sure that I understand that I haven't been thrown from the vehicle yet on this ride.

You wanted to remake your website, which was being hosted by a company called Wix. And you said, I want to try out this new tool called Gastown, where AI agents will talk to each other and the work all night on some kind of job you give them. So you gave them the job of remaking your website and you said that you'd like it in the style of of your old college radio days. How do I get to this website? It loads up like a retro radio station, and then it looks like a retro desktop, and it says Jackson Cantrell.

You've done it. And this is cheaper than what you were hosting it on before. This is free. You can host this on Cloudflare for free. It's a little busy. Where would you like me to click? Click on a music snob. It's in the left.

An icon. On the left side. And this is a working AI chat bot or you can turn up the snobbiness and it will judge your music taste. I'm turning the snob. Oh, it goes all the way up to 11. That's a shout out to, what do you want to?

Spinal tab. Spinal tab. I've turned the dial up to 11 on snobbiness and I'm going to say to the chatbot, what? It says type of band genre. or music topic. I will type the greatest American band of all time.

Credence, Clearwater, revival. And hit enter. Audible shutter. Oh, CCR. John Fogarty's lazy. Oh, John, if you're listening, turn off the podcast now. You know, before he became synonymous with looking out my back door and actually listening to actual folk music, people like, say, David Crosby, Yeah, Crosby's 1967 album, if I could only remember my name, was doing that vagabond vibe before Fogarty and even knew what a banjo was.

This is hurting me to read. That's brutal. This is a remarkable accomplishment, Jackson. Anything else you want to tell us before we go to break? Yeah, I'll just note that the AI hired other AI to create songs or AI music. and radio station news for a fictional radio station. That was probably the craziest thing. If the songs are not real,

does that mean we can play one on the way out to break and the lawyers won't get mad at us? We can play any of these songs on the way out to break. All right, since we can't afford fortunate son, hit me with some AI music, Jackson. This isn't bad. I think I heard this one at Tencent Draft Night. down at McGillacutty's back in college.

On this coffee shop, it's super underground. Behind the dry cleaners, you won't hear a sound. We'll be right back with the CEO of Agco after this quick break. It's pretty niche. You probably never heard of it. The real Glendell isn't the American... Welcome back. I had an editor at Barron's tell me recently.

This isn't Farmers Weekly. What he meant is that I had already done a five or so page cover story on beef and cattle raising. And I wasn't about to get another full page in the column I write to talk about Agco, which is a big tractor company. But we got to talk about Agco because the stock is up, I saw it up 36% year to date. That's a lot. So I snuck a mention into the column. You may or may not recognize the company's brands, Fent, Massey Ferguson, Valtra. They're familiar to farmers. The company is not just a tractor maker. It's a precision agriculture company.

All the farm equipment companies say that nowadays. They have sophisticated software and data tools and machines that help farmers get better results. But in Agco's case, as you'll hear in a moment from CEO, Eric Hansosia, it makes precision farming systems that farmers can use on their older machines, even machines with different brands. That's important now because farm incomes aren't great. at the moment. Some farmers are hard-pressed to afford new machines right now. We'll hear about that too. One of the reasons Agco stock is up is that it was cheap. There's been a rotation or broadening out of the stock market rally in the U.S. Agco also had well-received fourth quarter financial results. Oppenheimer writes that they were pleasantly surprised by a number of the underlying fundamentals.

They mentioned record global market share, strong price, solid progress on North America dealer inventories. In other words, if they're clearing out those inventories, that'll be more room for selling more machines down the road. In its investment thesis, Oppenheimer writes, we believe Agco is well positioned to benefit from secular trends of digitization, automation, and autonomy unfolding in agriculture. Given a focus on creating value for the producer, we believe Agco is effectively leveraging its technology investments to, you get the idea. Why am I going to read it from Oppenheimer when we can hear it from Eric, who was good to do it. We're going to do it. We're good enough to jump on the phone with us recently. Let's get to part of that call now.

What's the state of the farmer? What's going on with farmer incomes? How are farmers doing these days? And depending on which order we go in, the farmer is extremely challenged, especially in North America. There was less tractors bought in 2025 than any other year since before World War II. And 26 is going to go down. The industry is going to go down another 15%. So the farmers are under an enormous amount of pressure. So we're working really hard to try and help them through with solutions that make them more productive in this time. Use less inputs, trying to bring some profitability into their operation. How do you do that? What are you focused on? What kind of adjustments have you had to make to continue to do well during this tough period for farmers?

Yeah, we went through essentially five different transformations all over the last couple of years. We invested $2.2 billion in creating this business called PtX. It's precision technologies multiplied. We got the assets, the ag assets from Trimble and brought them over to Agco and joined them with our other tech business. And so we've got this big tech company. I'll talk more about that, but that's a PtX, where we sell to all harmers regardless of brand. Number two, we exited green and protein business, which was kind of a distraction for us. It wasn't the same channel, wasn't the same sales process.

So that allowed us to focus a lot more on precision A. Number three, resolved our issues with one of our investors called TAFE, allowed us to do share buybacks. That was another distraction that was bothering us. Rewiring how the company operates, taking a lot of the waste and cost out of the business, figuring out where we do the same thing, eight different ways across the company. We've taken $200 million out of our overhead off of a $1 billion base. So it's like a 20 or 20 plus percent efficiency improvement.

And finally invented this thing called Farmer Corps. It's a shift of the distribution model where instead of the farmer having to come to the business, mean, come to a brick and mortar dealership store, we come to the farmer, do everything on site. Farmers can spend a lot of money on some of these sophisticated machines. So at a time when they're looking to save, how do you allow them to upgrade their capabilities to get some of these modern capabilities without having to break the bank? So we're serving every farmer and we allow them, instead of having to buy a whole new piece of equipment, they just buy a technology upgrade kit from us from this PTX business.

It's called retrofit. So there's retrofit to the mixed fleet, retrofit to any brand of farmer is how we answer that question that you asked. It allows the farmer to get the latest and greatest technology without having to buy the whole new machine at a much lower cost point, yet get all the productivity enhancements. What's that to do for you financially? I'm guessing there's a subscription component to this. What does that do in terms of the trajectory of your business? The market has been coming down because of the pressure that the farmers are under. Retrofit only comes down about one-third as much as the overall market.

So revenues are less volatile. This is a high-margin business where we aim for a one-to-two-year payback for the farmer, so they're getting a good deal, and we get a high-margin business that's growing fast. We try not to use subscriptions very often because farmers tell us they don't like them. They want to buy the thing. They're overloaded. They've got nine sign-ons now just to watch TV. Everybody's overloaded, right?

Exactly. And so they are very frustrated with subscriptions. From time to time, they'll find something where the subscription model fits. Like we have an autonomy kit where they can buy a kit to put on their tractor. Could be our brand or someone else's brand. And it makes the tractor autonomous when you're using combine and green card together or when they're doing tillage. So some of them want it to subscription model for that because then they get the

latest and greatest things updated all the time or on our targeted spring. Some of them like that, but some of them don't. So we're flexible on the financial model. Give me some insight into what the farmer is seeing right now. When you talk with these folks, what are there, you know, one or two biggest concerns, what's hurting the most? and what are the prospects for things getting better down the road? When might that happen?

I think the first thing they'll talk about, one, two, come fast. Number one is market access. They want to be able to have market access for their green. I'm talking to a grain farmer, but livestock is the same. Livestock profitability has been pretty good. So the main concerns are from the grain farmer, and they want access to the grain. And during the trade negotiations and talks over the last year, some markets have come in and out, you know, China pulled back from buying from the U.S.

They shifted to Brazil. Now they come back in the market, but stable markets is number one. And then input costs like fertilizer seed and other chemicals are all up. And so that's a secondary concern. So they're kind of getting squeezed from both on the top market, top line on terms of revenue and who they can sell to. And then their bottom line in terms of their cost structure. And if there's a third one, it would be about labor availability. They can't get labor to be able to manage these complicated machines.

So that's why we're selling them autonomy, take a labor out or farm-engage to be able to manage their fleet. You can send the task to the machine. So when the operator gets in the field, he just says, accept task, go do it. And the machine is already set up. They don't have to know how to make all the adjustments real time. What are the prospects for a farmer seeing better days in the years ahead? What do you think has to happen? and what do you think of the chances of it?

It comes down to green prices and profitability. There's a couple things. I think China and the U.S. are, feels like I've settled in now on a market that we can count on. This is about tariffs and reciprocal tariffs and kind of trade war stuff. My understanding is our farmers invested years kind of showing their farmers how to take these soybeans and feed them to pigs. And that was a very good business for a long time.

And now that's hurting because of some, you know, tariff backlash. Have I got that right? That's exactly right. Yeah, China stopped buying U.S. large soybeans, but soybeans and corn from the U.S. farmer to put pressure on the U.S. government to end up with a good trade deal. Before I let you go, I want to hear about one or two G-Wiz technology things. Be as granular as you can. Something that's happening that's new, that's taken a specific task that was a pain in the neck for the farmer in the past and that's making it easier or faster. for them in the future. What are you, what are these machines able to do now or what's on the horizon? Arrow tube is one that we just launched at Winter Conference of Precision Planting. It's the very

first time anybody on the planet has been able to deliver a solution that creates the seed orientation that the farmer wants. So it gets the setup where the seed is always planted, tip down. And the flat is, is perpendicular to the row. Now, what does that matter? That means that what matters, especially in corn, but other crops too, is you have equal emergence. If every plant comes up exactly at the same time, you get the full yield, you have the chance to get the full yield potential. If a seed next to another seed is delayed by even as much as 48 hours, that seed is

essentially a weed. It will generate almost no yield. So 48 hours makes a huge difference. 24 hours makes a difference. 48 hours makes a lot of difference. And so the notion is, how do you get that seed perfectly replaced? We've been working on depth and singulation and spacing, but now we've got orientation. We're the only ones on the planet that can do that.

So this combination of placement in both orientations, what we expect is going to allow another yield jump, much like we had depth control a few years ago. So that's one of them. Second one would be we're the only ones that are doing a dual boom targeted spraying. So we use cameras on the boom of the sprayer. Artificial intelligence looks at the field as the sprayers going through the field, identifies only the weed versus the plant, sprays only the weed. That's one of the booms. But then you can also at the same time put down a residual. What the problem was before is that you

had to choose either one or you had to split your tank. And as a farmer knows, the worst logistics machine on the farm is the spreeer. Only about 25% of the time is a sprayer of a spreeers hours, boom on spraying. So it's getting tendered. All the logistics of keeping that sprayer are. So the solution of splitting the tank, we made the logistics even worse. One tank is always running out before the other. So you doubled the problem on a sprayer. Well, we've made one tank with injection systems. So you're always spraying residual.

And then when you see a weed, you spray contact riverside. So that's another great breakthrough for farmers that they can not hurt their logistics, but get both types of weed treatment done in a single pass with artificial intelligence. that saves them like 60% of the chemical. Eric, the seed orientation thing, is that something that could help with my wife's tomatoes because we've been getting some pretty lumpy results. I hope she's not going to hear this.

But the past few years, it's been kind of lumpy out there. Does that matter on tomatoes? I'm not sure about tomatoes, but I can check into that. Get your salsa yield up this fall. Thank you, Eric. One other thing that Eric told me, I left it out in case listeners were all beefed out from last week's episode.

But it was about cattle raisers. Remember, beef prices are very high. We were looking into why ranchers aren't raising more cattle. But Eric said that dairy cattle raisers have been crossbreeding with Angus. That gives them the option when they have bull calves of keeping them in the dairy pipeline for creating more dairy cows or selling them for beef. He said, when the market shows pricing power, creativity starts flowing in, and I think we'll see a rebuilding of the beef herd.

I want to thank Jackson Cantrell, not just for filling in as our audio producer, but for taking me on a tour of, what was it called again? Flavor Town. What was the name of that town? Where you, where you, gas town, but maybe I'll make a Flavor Town version. For walking us through Gastown and what all his AI buddies are doing to make software, even for people like me who are too dumb to code. That's going to be the title of my memoir someday. You're just jive coding now. I want to thank all of you for listening. If you have a question you'd like played and answer it on the podcast, send it in. It could be in a future episode. Just use the voice memo app on your phone and send it to jack.

Dot How, H-O-U-G-H-O-G-H at Barron's.com. Subscribe to the podcast and Apple Podcast, Spotify, wherever you listen. If you listen on Apple, please write us a review, unless your name is John Fogarty. You might want to hold off on the review. It wasn't me, John. It was the AI bot that Jackson's AI bot. See you next week.