Skip to content

Podcast

AI & Cloud News of the Month - Feb 2026

The Reasoning Show

Source ↗ ← All highlights
  • Podcast Rebrands To Focus On AI
    • The Cloudcast is rebranding to The Reasoning Show to focus more heavily on AI while keeping cloud topics.
    • Aaron Delp and Brandon Whichard framed the transition as a shift in emphasis, not an audience or feed change. Transcript: Aaron Delp Brandon, how are you, man? I’m very good, and I’m excited because I believe this is the transition show, right? Brandon Whichard This is the last cloudcast, and we’re going to roll right into the reasoning show. So I feel honored to be part of the handoff. Aaron Delp Yeah, no, it is. This is, yeah, officially the last official The Cloudcast show. For those of you that didn’t listen to last week’s show, as many of you sometimes will get backed up, we are continuing the feed. You will not have to change anything, but we are changing the name of the show to The Reasoning Show or, you know, shorthand for reasoning. We’ll be a little more AI focused, probably a lot more AI focused, but still doing some cloud stuff. But, yeah, this will officially be the last of this era. So 15 or so years. (Time 0:00:24)
  • How Listener Feedback Saved The Show Name
    • Aaron recalls the show’s origin story shifting from Two Guys One Wire to The Cloudcast after a listener warned about an unfortunate meme.
    • The anecdote illustrates how small cultural signals and listener feedback shaped the show’s name for 15 years. Transcript: Aaron Delp So, long, long, long, long, long time listeners will know this. You guys on Software Defined Talk just did a show called, was it Two Guys and Tokens or something like that? Brandon Whichard Yeah, Two Guys and Their Tokens. Aaron Delp Two Guys and Their Tokens. So, anybody who has listened to the show for a long, long time, I don’t think we’ve discussed it in probably a long, the original name for this show, so this is not the first time we’re switching The name of the Cloudcast. It’s actually the second time. The original name for the show was called Two Guys, One Wire. Oh, okay. Wow. Well, because it was during the days when there was sort of like unified computing was coming out, and it was this whole idea of like, you know, everything’s going to run over the network, Storage, and compute, and all this sort of stuff. And then a longtime friend of the show, listener Stu Miniman said, hey, you guys may want to look up this thing called Two Girls, One Cup because, you know, you guys are dangerously close To that meme. So, yeah, the original show was two guys, one wire eventually became the cloud cast. But here’s my theory. I like your idea of, you know, the way to solve migration problems is to take two people who have, are sort of experts in adjacent technologies, right? Your whole idea, the whole context of this was you guys were talking about like COBOL evaluation or COBOL modernization. We’ll get into this. And your idea was, you know, take a really smart COBOL engineer, if you can find them, and then a really smart whatever, you know, Java engineer, put those two people together in a room And then like unlimited tokens and they’ll go solve, you know, whatever the problem is, right? That was sort of your theory, which- Absolutely, yep. I think it’s fine. It’s really hard to scale that, right? You got to find a lot of two-person pairs. But I want to come back to this thing I was thinking about. We’ve sort of gone away from the idea of being fascinated with the two-person sort of miracle, right? Like that was the origin story of like Silicon Valley, right? It was Hewlett and Packard. It was Jobs and Wozniak. Cisco was started by like a husband and wife team out of Cisco, right? There was a lot of like sort of two-person, multi-person teams. (Time 0:03:28)
  • Most Big Tech Wins Are Small Team Stories
    • Brandon argues the popular ‘singular genius’ founder story glosses over that most successful products grow from small teams or pairs.
    • He suggests sharing credit better because many successes actually come from collaborative groups, e.g., AWS features developed by multiple teams. Transcript: Aaron Delp Like that was the origin story of like Silicon Valley, right? It was Hewlett and Packard. It was Jobs and Wozniak. Cisco was started by like a husband and wife team out of Cisco, right? There was a lot of like sort of two-person, multi-person teams. And we were always fascinated, I think, with this idea of like two people who probably have very different personalities, right? Like you probably never put them together by themselves because one is very introverted, one’s extroverted, one’s got really good skills in this space, and one’s got – and we don’t Really have that anymore. Like we don’t even, I think the last one that we sort of were fascinated by was, was AWS being sort of Andy Jassy and Werner Vogels, even though I don’t know that they were truly the founders Of AWS, depends who you ask. But we’ve, we’ve sort of gotten into this idea of like, you know, the, the singular, uh, you know, genius, right. The, the, you know, the Jeff Bezos at Amazon, Mark Zuckerberg, Mark Zuckerbergberg do you think we need to get back to the two guys i feel like the two guy theory the two person theory tends To not not a it creates very sustainable companies because they figure out how to do their things but also just less assholes in general less you know just man but i don’t know do you feel Like we need the two people or do you feel like our our society now can’t can’t handle that? They need the one person, the singular sort of person, you know, being being the smart person. Brandon Whichard I’d say two part answer here is like I actually think most of these companies, when you dig behind the scenes, it is usually a couple people that are working on it. So it is like, let’s call it like there really are mostly two people or like small teams, like even like OpenAI. If we go into OpenAI, right, and we go back in time, we’d find a couple really key people there. But what happens is kind of back to like a little bit talking about almost like your wife’s thing. It’s like people like stories, right? And the stories that people like to tell is the hero’s journey. Like, you know, this one person rose above it all and did everything. And that is probably what gets popularized. So probably what we need to get back to is maybe just kind of what you’re alluding to. Like, maybe we just need to like share the credit a little bit more willingly and be like, it’s rarely one person can do every conceivable thing, right? It’s really usually, and it usually is somebody that’s more front-facing that often takes a lot of the credit, but the behind the scenes. And I think, you know, that AWS story is a great example. It’s like, I don’t know, when you read, when you really dig into it, there’s like a couple of different groups, like one working on EC2, one working on S3, and one working on, you know, Something like, I can’t remember, maybe something else. And it was like, those kind of teams kind of all together did it. But, you know, that part is never really told. It’s just like, you know, I think it’s always Andy Jassy wrote the memo and, you know, the rest is history. But, you know. Right. So I like it. So, yeah, I mean, to your point, yeah, it would be good. Let’s share the credit a little bit more. There’s plenty of success to go around. (Time 0:05:12)
  • Linux Release As A Test For AI-Generated Code
    • Linux 7.0’s release provokes questions about AI-written code in future kernel releases and Linus Torvalds’ ongoing role.
    • Brandon predicts a high percentage of future kernel code may be AI-generated, making kernel adoption a milestone for AI coding trust. Transcript: Aaron Delp On the cloud side, Linux 7.0 came out. Are you excited about Linux 7.0? Are you upgrading all of your servers around your house or in your data center to 7.0? Brandon Whichard I’m really, we’ve talked a little bit about on our show over at Software Defined Talk, but like, I’m only really interested in it because I think I’m really looking for it. And I sort of, I’m trying to get everyone’s prediction. It’s like, whatever the next major release of Linux is, 8.0, whatever you want to call it. Like, my question is, like, how much of the code in that will be generated by AI? And just to put my cards on the table, like, I’m at, like, 85 to 90 percent. But of course, the Linux kernel group, which you probably know more about than I do, is like famously like, you know, kind of slow to adopt new things, which is probably good. We don’t want the Linux kernel like, you know, kind of doing weird things. But I think it’s going to I just think this is sort of like it’s like a demarcation point. It’s like if Linux starts to accept or starts to use a lot of AI coding, which I can’t imagine why they wouldn’t, then that’s going to be like a real milestone. So that’s, that’s really the only thing I personally am interested in. But like, other than that though, I heard everyone says Linux now, so the Apple hardware is like better supported. (Time 0:08:28)
  • Turn Consulting Playbooks Into SaaS Products
    • Corey Quinn’s move from consulting to building cloud cost management software shows a playbook: learn problems through services, then productize the repeatable playbook.
    • Brandon and Aaron note consultants have deep problem knowledge that maps well to SaaS but face crowded FinOps competition. Transcript: Aaron Delp Provocateur, I guess. Yeah, Corey Quinn. Pondent, provocateur, I like those, yeah. Podcaster. Corey Quinn of the Duckbill Group is moving out of the consulting business, moving into the software business, you know, made an announcement, took some money from venture capital, Which is always going be interesting. He’s no longer wearing a suit. He’s now, he’s become sort of the long haired, I don’t know, dad, you know, kind of middle-aged sort of thing, but he’s, he’s making software. He’s in the cloud cost management business. And I would think, you know, his team has, he has lots of experience in this. The question becomes, can you turn all of your, they’re probably a perfect example of like, can you turn all of your internal instincts and experience into software? Because that seems like what they’re doing with this new round of funding, I don’t know, it’s around $10 million or something like that. So unfortunately, I don’t think he’s going to get to build his own data center, but congratulations to Corey on evolving that. Brandon Whichard Well, you know, traditionally, this is one of the best ways to start a company is you do consulting for a long time. Yeah. Because that forces you to learn all the problems. And then you sort of have this playbook. You’re like, I go in and I do this thing every place. And then that is the foundation for, like, really understanding the problem set and building software. But I do think it’s kind of fun for Corey because he has kind of made his whole persona as sort of taking shots at, in a fun way, in a jovial way, at AWS being like, look how stupid their naming Is, or like, why did they do this? This AWS service doesn’t do something that’s obvious, right? And it’s like, so it’ll be interesting when the shoe is on the other foot, right? Like, well, it’s like, okay, when you start building software, you start to learn why it’s like, you don’t have all the features, or you have to make naming choices. So that’s going to be kind of fun for, just for me watching. It’s like, OK, well, how does he handle it when people are like, I can’t believe the Duckville group software like doesn’t support, you know, Azure doesn’t, you know, do this thing that’s Obvious or has this obvious error. So that’s going to be maybe the more fun side of it. And then, of course, too, like he’s in an interesting but crowded. I’m kind of putting him in the FinOps group. I don’t know if that’s fair, but it’s like it’s an interesting area but like there’s tons of people inside of FinOps. So it’s also going to be fun for him. It’s like, OK, well, like, you know, it’s almost like you’re entering the arena, you know, which is fun. It’s like you’ve been on the sidelines. You’ve been kind of a pundit. But now you’re going into the arena with everybody else. (Time 0:10:32)
  • Why Transitional AI Tools Struggle
    • Cursor exemplifies transition-stage tools that enhance existing workflows but lose relevance when fully native AI experiences arrive.
    • Brandon says tools that merely augment old tools risk obsolescence versus tools that ‘do everything’ for users. Transcript: Aaron Delp First question I got to ask though is, and again, this is emblematic of like, what’s going on in the AI space. You’ve been on, we’ve been doing this for a while now. Last year, it was all about Cursor. What happened to Cursor? I have not heard a thing from Cursor in I feel like at least three or four months. Like, are they, I don’t want to say are they gone, but like, what’s going on? Brandon Whichard Well, I think what happened was a cloud code happened to Cursor, as well as, you know, there’s all the other ones, basically command line development. And I think the key learning, and I went through the same process I would ask anyone to do is like the learning here, the jump is like when you go away from like, I don’t need the IDE. I just tell this thing what I want it to do. And it does all of that work. And I’m not like, I’m not really, if you will, pair programming. I’m product managing Cloud Code or Google’s version of it. And it’s doing everything. And it’s like, you don’t really need, you know, I think, you know, for a lot of us, it’s like, you don’t really need cursor. Don’t be in the IDE. Just ask the tool to do what you want it to do. And then if it makes a mistake so i think cursor is like it really is in a a precarious situation i think it was sort of like they were like the the ultimate transition i was about to say the Ultimate transition girlfriend or something transition partner it’s like oh this is good and then you’re like no that’s actually what not what i wanted i actually found true love somewhere Else so i think there it’s good and i i say this as someone that loved Cursor as much as anyone. I know. When I first saw it, I loved it. I was huge in, but I’m like, I don’t even fire it up anymore. I’m like, why would I, why would I waste time inside that? I’m just, I’m just on the command line, just speaking regular English. Like, that’s not right. Do this. That’s all I say. So I don’t know. Tough situation for him. Aaron Delp Yeah, no, I, and again, I’m not knocking them. I think they might end up being a first ballot Docker Hall of Famer, though. It should have taken the money, you know, did a lot of interviews. Everybody couldn’t believe how fast they were growing. But yeah, you’re right. I think they’re going to be emblematic of the you know, we saw this in the cloud as well. And eventually it kind of not flipped. But like when you when you’re doing a transition, like you need to not be tied to the old thing. Like you need to be, be the new experience. And, and they were a transitional experience. (Time 0:13:26)
  • Anthropic’s Month Of Enterprise Moves
    • February felt like ‘the year of Anthropic’ as Anthropic repeatedly launched enterprise-focused products that challenged SaaS incumbents.
    • Aaron lists announcements like Claude CoWork, COBOL modernization, and Enterprise Agent as moves to win enterprise adoption and displace niche SaaS. Transcript: Aaron Delp It was like, we are right now living in the year of Anthropic, the year of Claude, the year of Anthropic. You know, Anthropic, it felt like every week this year, and I don’t know if they, I don’t know what happened. Oh, there we go. I’m back. I’m like, I can’t make it. What a matter? This show’s not video. It felt like every week this month, Anthropic was announcing something that not only spooked a whole bunch of people, but also got the stock market to freak out. Everything from, they announced Claude Cowork, which was like, you know, Claude for, you know, regular normies, you know, everyday white workers, you know, plugs into all of your Tools. They, you know, shocked the Cobol market by saying, hey, we now support Cobol, which, you know, I guess is kind of to your old point, like people have been saying you can modernize Cobol For a long time, even IBM, and it shocked the market, shocked IBM stock price. They announced they had a whole event called Enterprise Agent. So we had a whole month of software. You know, you and I talked about this, like the whole software market, the SaaS market kind of got smacked around because people were like, oh, wait a second. So you don’t need SaaS anymore, which, you know, again, debatable. They sort of came over the top of that and started announcing, hey, we’re just going to plug into all of your tools. So to your point of like, you know, they’re doing the slow sort of creep up on all of the SaaS people, but like not in their back. They’ve now gone face to face with them and been like, no, no, no, just plug into us. Don’t worry. It’ll be good because obviously people love your stuff, right? Wink, wink, you know, like, and the SaaS people are like, people love our stuff. You know, you see what’s, uh, Mark Benioff out there, like, oh, people love Salesforce. They just, you know, which is, I think, I think it’s on the, it’s on the brand. It’s comical. It’s, it’s on the Brandon manifesto of like, nobody loves Salesforce. Nobody loves that. But yeah, I mean, Anthropic is, you know, I don’t know if they’re planning to go IPO this month, but boy, they have put out so much raw meat for people to be like, oh yeah, the software industry. Yes, just, you will just give us all of your money. Like all of your money will eventually transition to us. It’s been, it’s been an interesting month. (Time 0:16:34)
  • Anthropic Wins With Task-Focused Enterprise Tools
    • Brandon sees Anthropic’s strength in targeted enterprise tooling and a ‘do-it-for-you’ flow that produces immediate productivity gains.
    • He contrasts that with OpenAI’s consumer/aggregation focus and notes users report stronger flow and output using Anthropic products. Transcript: Brandon Whichard Well, I think, you know, it’s really interesting because they’ve had quite like, I almost call it like the comebacks, not like the way where, but I think people kind of like wrote them Off. And I think what they figured out and maybe by accident, maybe by strategy, is like the enterprise play really seems much more compelling suddenly than the consumer play. Right. Because it’s like, you know, this little Ben Thompson, Ben Thompson is like, you know, open AI needs to do a, an advertising model. Whereas Anthropik is basically like, we’re going to sell you tools, right. That you can use in your enterprise and you start using them, whether it be cloud code or co-work and they’re awesome. Right. And you feel, I kind of joke, it’s like, if you start doing this in a certain way, you feel kind of this intoxicating flow of like, I’m getting stuff done. I’m building stuff as fast as I can think I can like see code and documents being created. And it’s sort of like seems completely separate from open AI. I know they have codecs and their own products as well, but it’s just like they don’t seem as good. Like they don’t seem to be working on that as much. And it’s like that combined with this kind of natural positioning of like, hey, we’re, I’ll put it in good, quote unquote good, or using AI for good has been a really natural kind of contrast To OpenAI, right? Who’s sort of like constantly. And it’s like, it has really been amazing to watch. And it’s like, I don’t know, going forward, you know, I think when I first originally saw open AI, the thought I would pay $100 a month for it was crazy. And now the thought that I’m only paying 30, I’m like, I’m not getting enough out of this. I probably need to pay 100. And I’m using exclusively Anthropic for all that work. I mean, it’s amazing. So it’s like, it’s going to be a real interesting, like, kind of a stratechery Ben Thompson kind of thing. It’s like, were we kind of all focused in on the next Google, right? But really, it was the next set of enterprise tools. And Anthropic is kind of just owning that market. (Time 0:19:12)
  • Anthropic Faces Production Cost And Infrastructure Limits
    • Anthropic’s rapid productization raises financial scrutiny because inference and chips are costly and Anthropic doesn’t own chips or data centers.
    • Aaron warns investors will watch burn versus revenue closely as Anthropic scales enterprise sales. Transcript: Aaron Delp Yeah, it’ll be, it’s going to be interesting to watch. I think they’re now going to be the next one that’s going to be on Revenue Watch, you know, every month, every quarter, like how much are you making? What are you forecasting? Because, you know, OpenAI, you know, obviously, you know, AI is expensive. Inference is expensive. Chips, you know, training is expensive. You know, they’re now a little bit stuck in the quagmire of like, okay, everyone’s doing the math because they’re like, okay, they’re going to eventually go IPO. And, you know, they’re talking, I mean, they’re still talking about like, we’re going to burn through $650 billion, you know, by 2030. So in the next four years, we’re going to burn through, you know, more money than anybody ever has. And Anthropic really hasn’t gone down that path, although they did just raise another $30 billion. I mean, their one kind of huge flaw, not flaw, but weakness at this point is they still don’t really own the means of production. I mean, they own the model. They can train the model, but they don’t own their own chips. They’re dependent on somebody else for chips. They don’t own data centers per se, right? They’re dependent on, so, I mean, the two of the hugest, you know, things that will impact their cost, they don’t really own. So they’re going to be an interesting sort of, you know, follow the money, follow the burn versus the revenue kind of thing. And, but (Time 0:27:10)
  • AI Shortens Idea To Demo Cycle For Individuals
    • AI enables bottoms-up, individual-driven product creation: people can prototype working solutions with models and ship them before asking for funding.
    • Brandon and Aaron describe signing up free cloud tiers, building a V1, and walking into meetings with a working demo instead of memos. Transcript: Brandon Whichard As we play it forward, I’d love to get your take on this. It’s like, for a while there, I think it was open eyes, like, Hey, let Anthropik go public, sort of like be the market test. Right. And then they would do it. And now does that flip a little bit? Is it, can Anthropik go public in a way that maybe with like a slightly more, this is crazy to say, a slightly more rational cap table and, you know, accounting statement. And like, is that a way for them to actually kind of like potentially jump open AI, right? So it’s almost like the first to go public and the first to kind of like show that, like we do have growth in the enterprise. Like, does that sort of often like does money start to funnel into them because they’re public quickly while open AI sits on the sidelines? So I think that’s a really interesting story is almost like the IPO goes from like, if you will, you were going to be the test bed to like, not only could you go first, but you could jump and Become the leader. So I would be worried about that if I was open AI. It’s like, if they get out there first, do they become the leaders in this, or the perceived leaders in AI? Aaron Delp Yeah, I think there’s a chance that Anthropic does a better job of painting the story that Wall Street has to live under, like painting the structure. Because again you know, you’re going to be doing sort of the thing that Amazon did back in the day, which was like, wait a second, you guys want the stock to go up, but you’re not talking About any revenue. You’re talking about revenues, but you’re not talking about profits. And they’re going to be at a loss for a period of time. It’s just, you know, how long do people believe in their vision versus being like, okay, how long can you lose money? You know, how fast are you growing? And yeah, I think right now they are doing a better job of finding people that are willing to, you know, I mean, I feel like every day there’s somebody evangelizing about being like, here’s What I did individually. You know, it was me and a month, you know, basically what you were talking about. Like I got into flow state for a month and I built this thing. And I feel like I could do this for everything now. And you’re just, you’re starting to see that more and more of those stories come up. The other hard part about that, which will be interesting to watch is I feel like every one of these stories are an individual. Like they are, it’s very little of like, hey, me and my team figured out a way to jointly do this. It’s sort of like I built my own team. My own team is like me and 25 agents and i and i wonder from an enterprise perspective if you’re like you know it’s it’s the sort of classic like you know enterprises are built around teams And the idea of like you know you have the best people and then you have other people and this becomes like okay how am i going to manage this you know like how would i manage this thing which Enterprise thinks about as much as anything else is like how do i manage it because people come and go it still feels like that like ai is a very individual sport at this point it’s very Much not like a team sport yet and i think that that’ll be the next really big thing to figure out if somebody can crack that a little bit or at least you know start to talk about like hey how Does the team work around this stuff? You know, and get, you know, or does that force multiplier even matter? Brandon Whichard Yeah, and I kind of, I always say it’s like, it really is a bottoms up technology, right? That’s where the leverage is. And I do think, you know, it’s kind of back to this IBM statement for a while. It’s like, I don’t know, you know, whether IBM can or cannot, you know, modernize COBOL. Well, it certainly can. First of all, I’ll start there. But it’s really just like, it may just be easier at this point. And kind of I was joking before, it’s like, you know, you just get a couple people doing it, right? You don’t need like all you don’t need the massive team and structure of like a services company coming in and doing this anymore. And it’s just easier that way. So as I go forward, I always think, you know, the joke I always think about now is like, listen, if I’m going to the AWS meeting where we’re about to review my project, like I’m not bringing The six page memo. I’m bringing a working version one. I’m like, here it is. This is what it does. And here’s how it works. And like everyone can start using it. Like, why not? Like that, it’ll probably take you as long to write a memo as it would to actually sit down and use your favorite coding assistant and build out like a very good version one. So I think that’s like, but that’s gonna be very hard for like, we’re just using AWS example. That’s gonna be very hard for a company to kind of grasp because most of it’s like, show me a proposal, I’ll fund you, there’s checkpoints versus like you just walk into the room, like I’m done. Like I have version one, like I’m ready to launch this or I want you to, like, it’s really not asking for money. It’s like, I need you to now help me get customers. I want you to help to assign the Salesforce to do this. I’m not asking for anything. And I think that’s just executives just like their minds break, right? It’s like, where was the, where was the meeting? How did we do this? (Time 0:28:16)
  • OpenClaw Proved Personal Agents But Raised Security Fears
    • OpenClaw (OpenClaw/OpenClaw variants) popularized the personal AI agent idea but alarmed users by requiring wide access to local files and systems.
    • Both hosts describe the project as a security boundary test that proves the concept but pushes users to demand safer productized versions. Transcript: Aaron Delp Like that. Yeah, no, it’s a, it’s, it’s an interesting time. I think the last, the last big news item that came out this, I don’t know, when the last, it came out in like in the last month, but it’s kind of continued is, is OpenClaw. This idea of, I forget what it was originally called. I think it was called MoltBot. It was. ClawBot was the first one. ClawBot. It was a spinoff of Clawed. And now it’s called OpenClaw. Have you played around with this thing yet at all? I’m just right about it. Brandon Whichard As much as I like to try new stuff, this was like one step too dangerous for me, right? This idea that I’m going to install it on my Mac and just let it go to town and do anything. But I do I understand. I think, you know, I think it’s going to go down as sort of like it’s proven the concept of this personal AI assistant that just kind of lives on your machine and can do anything. Clearly, there’s something there and people really want it. But there’s also clearly there’s like even I even I like to be wide open. You know, I’m not the person that usually wants to. But like this is a step too far. So I think the next step will be like, can they figure out a way? And this is, I think, why Anthrop, or not Anthropic, why OpenAI hired the founder. It’s like, can you find a way to package it up in just a little bit more simple slash secure way that gives people a little bit more control over what’s gonna happen? But it is certainly popularized the idea of like, it’s just like, why do I even sit at the computer? I just have OpenClaw do everything that I would have done at the computer anyway, which is, again, back to kind of where we originally started. It’s just like, how far back in the chain can you go? It’s like, do I even need to be in front of the computer to do the stuff? No. OpenClaw will do a lot of it. And people love that idea. So I’m excited to see like how it gets, if you will, productized going forward. Aaron Delp Yeah, and for anybody who hasn’t heard of this, I don’t know if we mentioned on the show before, it is an open source project. It’s essentially a personal assistant, an AI agent slash personal assistant. I think the one thing, so as are many things on the internet or Reddit, you know, it’s doing all kinds of crazy stuff. The one piece of it that kind of spooks people, and I think you are kind of highlighting it is, you basically, you install it on some piece of, you know, some piece of computing. It could be your own laptop. It could be a separate Mac Mini. It could be whatever. But you basically have to give it like full access to everything. And I think that’s the part that scares the crap out of people is they’re like, oh, wait a second. No, no, no. I mean, it’s the equivalent of like, yeah, just give it root access to your life. And then it becomes your personal assistant and tell it what you want to do. There’s some people doing some interesting stuff with it, you know, whether it’s like, it kind of cleans up my desktop, it does some things, whatever. But yeah, I think that’s the piece that scares folks. This one, this one feels a little bit, I guess, if we’re putting it in a historical context, like we’ve seen, this one feels like it’s right on the edge between being like, is this really Important technology? Or is this like somewhere between novelty, maybe, and being like, no, no, no, to move into the next realm of this, to really be ahead of this thing, this is a mind change. This is a separate mindset that you just have to adopt and you have to be willing to give up on certain things in order to get the other things. I don’t know which one of those it is, but we’ve seen some of those in the past where it’s like, you know, this is, you know, people would call this like AI agent native and everything else You guys are doing is sort of like hybrid AI agent because you’re still human in the loop. You know, it feels like it feels like at least a data point for these, these sort of psychological things of like, are you going to commit to doing this or are you still kind of like one foot, You know, on the other side of the line? Brandon Whichard Well, we talked about earlier, but I think, you know, cursor was sort of like, in retrospect, I think it was playing this transition. It kind of got us comfortable. And I think OpenClaw has sort of opened our eyes to be like, I don’t need to sit at the computer. It can do everything for me. But it’s maybe just a little bit too much. And so who knows, maybe they figure, find a way, like put it in a VM or constrain what it’s going to do or like make it ask. Like that’s sort of the next thing. And they just put OpenClaw on a foundation, which I’m sort of like, well, I feel like that means that’s the end of OpenClaw. Right. It’s sort of like it’s not the end of OpenClaw, but it kind of means like the new ideas and the fun stuff is going to go elsewhere. And I think that’s right. And that’s why I think when they hired an opening, I hired the founder. It’s like, and of course, a million people are trying this. It’s like, it’s sort of like, this is a transition. Open call maybe showed us the idea of what is possible, but massively insecure. And now a bunch of people are going to go be like, here’s a way to do it. And dial back in a way that makes you feel more comfortable. And I think that next six months to a year is going to be a bunch of companies doing that. (Time 0:36:50)