Enemies in AI: Why Sam Lost the Race

Nahshon Cook-Nelson: I think Sam Altman has a massive credibility problem. The consumers as well as business owners are like they don't trust Chat GPT for business use cases. You and I know Nishon, we say this all the time. Never make enemies. We're always friends with people. A lot of non-technical people are paying for technical solutions now and they don't know how it works. Dario embraces like the mad scientist, right? He's got the crazy hair, he's got the glasses, he's always fixing. Like, Claude has a massive problem because they're addicted to Claude. Block already has the internal agent. We don't need to go help these software engineering companies. AI is not replacing jobs, AI is creating jobs. Welcome back to the Return on Intelligence. podcast where we're all about helping you save time and money when implementing AI. I'm your co host, Nishon Cook Nelson, joined as always by Andrew Eman. How's it going? A A and Nay nay. I'm gonna keep doing that for every episode. I like that. It's I like that. It's easy to remember like a like a basketball duo. Yeah. There we go. We still have to play our basketball game to prove out who's the better shooter, but I would definitely be the ⁓ the rookie that would come in and surprise you. Yeah. Well, I I think I'm the rookie now because you've hyped it up so much. You said like you've never lost a game. That's such a really big claim. I've never lost a one on one, but there's two factors of that equation. One is I'm pretty good. The second one is I get to pick who I play against to decide if I'm ever gonna lose one. That's that's fair. That's fair. All right, so we've got a really jam-packed episode today. One of them is a guy who loves staying in the news, Elon Musk. You hit me up and told me about this. He just made a move this week. ⁓ can you tell me a little bit more about that? Yeah, I mean the news on the surface looks like during the court case that he's fighting with Sam Altman and trying to reduce open AI's credibility, he invested alongside or with Claude and Anthropic to ⁓ increase their capacity of GPUs. So he's giving them GPUs to to operate their cloud instances on. And what the immediate news story said is everyone's code is going to have two X or three X, I can't remember the number, but you just have more time using Cloud Code, which is a huge unlock because as you know, you get blocked every day. Yeah. Yeah. That's pretty cool. I think one for the Elon fans, of course, they're like, yes, like bowing to him even more now because he's like Alleviating some of their blockers. But ⁓ also I think it's it reminds me of the Lego movie, is something I've been thinking about lately. Have you seen it? I've seen all the Lego movies and they're classics. Yeah, the songs can be stuck in everyone's head and they start singing it right now. Yeah, the master builders. Like everyone's a builder in the Lego society, but then there's like the master builders, and I think he's helping those master builders by alleviing alleviating that. But you saw this as like a jab. against open AI, you know that they're they're fighting with each other. Like why would he do this? I look at it actually from a CEO standpoint. And when you go to therapy and coaching and I work with a lot of CEOs ⁓ and I travel with a lot of them too. You'd start to notice patterns. The CEOs that have been around for a really long time, you respect and you enjoy and I would, you know, bend over backwards for a lot of CEOs if they get into a dire straight situation. They would call me up or text me and be like, hey dude, like, you know, my company's not doing well. I need, I need some help. With X, Y, and Z, I'd I'd bend over backwards for them because I know they're solid people. They've told me honestly through the years how they've run their companies or how they managed it. And so I'd help them. I think Sam Altman has a massive credibility problem. And this has been brought on by him. This is not something that the news has created. This is not something that other AI companies are falling into. Because you would think that, like, if if he has this problem because he's running an AI company, then you know, maybe Darius has that problem too. It doesn't seem like anyone else has the negativity that Sam gets brought towards him. For making up claims, for putting false things forward, for saying that AGI is here sooner than later. He he seems to have this credibility problem, but he also seems to have this problem where he's conniving and he's trying to gain power. And it's it's very tangible, like it's palatable to us watching him when he makes these decisions. You're like, dude, you just you're just trying to accumulate more. So he did Sora and he did an image generation, and then he did the greatest thing to Slack, and Slack was gonna be dead, and then video generation was gonna be dead, and TikTok was gonna be dead, and he was just announcing. Product after product after product that he was going to rule the world with. And that approach, I think, has a lot of bad taste with people that run certain companies. And so if you're stepping on everybody's toes so quickly, and then you're just kind of like rubbing them in the dirt while you're doing it, you're gonna lose friends quickly. And when he's not credible and he's not giving solid answers when the news is asking him, people start creating enemies. And I think he's created enemies, whether it's with Elon, whether it's with other companies. And so Elon looked at the situation and said, I can help. Claude, increase their capacity while they're in ultimate growth mode in the middle of the news cycle, passing GPT for usage, tokens, and basically just by building a better product, I can really step on open AI while I'm also battling them in court. And he doesn't even have to put that much money forward. He has to use some of his resources that he built. So of course he's going to take that. That's a prisoner's dilemma solution where you're just like, I win in all situations if you're Elon. And the person that loses the most is Sam. But I don't think this happens to open AI if Sam was a better business manager or relationship manager with some of the other competitors in the space. And as you and I know, Nishon, we say this all the time: never make enemies. Every single person that comes at us and is like, hey, I love what you would do at nine two three. I'd love to have a similar product. Be like, cool, here's how you build it, here's how you do it. We're not like competing. We're always friends with people because of the reason. One, I want to learn from them. And two, I wanna see them grow. Like it's fun when you see someone call you back. Like I was just in the car and somebody called me and was like, I am successful today because you told me three years ago how AI worked. I'm like, I love that. He's in my space. I'm competing against him, but he's in Canada. He's not gonna compete against like specific customers. If he does, I'd be like, I'm so proud of you for beating us at our own game. Now I need to step it up. But if you have the Sam Altman approach, I think you're gonna get hit and stepped on and pushed over when the people start surpassing your product, which is finally happening. Yeah, it seems like you're hinting at trust as well. Because people don't trust Sam Altman. They like you said, they ⁓ even the consumers as well as business owners are like they don't trust Chat GPT for business use cases. And something I've noticed is transparency. We trademarked full transparency development. And throughout every conversation with the client or prospect, they always love that because a lot of non-technical people are paying for technical solutions now and they don't know how it works. If I'm a CEO of a construction company, I don't know how it works to like build a customs solution on a pine cone database. And I don't care to know. But what I do want to know is that I can trust the vendor or the person or the model that I've hired to do it securely. And I think a big part of that is transparency. And it seems like Sam, at least from where I'm sitting, hasn't embraced that. And you look at Anthropic and the way they did MCP. Like they were the ones who like coined that and they put out a whole thing, like, hey guys, this is how it works. Yeah. Everybody go bananas. Yeah. MCP. And it was like, ⁓ you're making the overall ecosystem better. That's the vibe I get. Yeah. It like Dario embraces like the mad scientist, right? He's got the crazy hair, he's got the glasses he's always fixing. Like when you look at what the product is they're building, it like it fits him. Now, Claude has a massive problem because they're addicted to Claude. If you've talked to people internally, I have a few friends that have worked there and one that works there. And they say like they trust Claude in a lot of situations. They'll be building something and then all of a sudden they'll get in an argument human to human and they'll ask Claude for the answer. So there is this part where they believe in Claude a lot. And I think Dario's like done a really great job, like painting that picture that this company is just a scientific madman house. And like when something like all of a sudden, what was Claude Code? It was like gonna break some sort of system or hack it. And Dario's like, We're not gonna let that happen, right? Like it seems like he's on your team. He's like trying to go to bat for the Americans or the world. against this like AI thing that's being formed. Whereas like Sam's like being buddies with it. Sam's like trying to look at ChatGPT being like, this buddy of mine and I are gonna be king and queen of the world. And that that's the impression it gives when he's like, we're gonna dominate SOR, we're gonna take over Slack, we're gonna do all social media, we're gonna have an agent network. Like it just has that feeling of like dominance versus like scientific exploration. And I love the scientific exploration approach. Yeah. Yeah. That's interesting. That's values. That's values. It's let's go to the The like ⁓ you just hyped up Claude, but you feel like people who are using it in the wild every day, ⁓ might not know some of the risk. And Claude, even themselves, we've done this before. Claude has a lot of ⁓ spots where they say, Hey guys, don't do this, be aware of these risks. But usually everyone just says, Accept, accept, accept. I don't care, always allow and just see what happens. Yeah, like ⁓ but you think specifically for business owners and business leaders, if you're a dude on your own, whatever, dude. Good luck. But for business owners, you think there's something they should be keeping in mind? Yeah, I think two things are kind of happening quickly. Kind of mid twenty twenty five, we started realizing that we had this opportunity to build something for ourselves. And at the same time that we were building products that are now before us that most of us are using cloud code for, we're starting to realize that They are becoming part of our daily lives, whether we're using the application that we built, whether we're using the dashboard that we built, or whether we're just having the same conversation with Claude every day because that's just part of our job. And we're like, ⁓ can we just automate this a little bit? But that thing that was supposed to be fun and kind of replacing like the mundane tasks that you were doing became a big part of your day. And two things happen with that. Your cubicle mate next to you will look over and be like, Can I have that? And the second thing that will happen is your boss will be like, Hey, you seem to not be working that hard anymore. Can I give you more work? So you get more work and then you rely on the Cloud Code automation to run that part of your job that you used to do that used to take hours. And you should be creating more revenue opportunities for your boss while the Claud Code thing does the automation. I mean, that's the picture. That's what Dario has been presenting to us. So as that happens, you get into this pickle, especially if you work for a big company, is how do you deploy this? How do you get it in a place that it's stable, that you're not also the QA, the engineer, the midnight fixer, the designer? And also when a teammate gets on it, you're like the authenticator. You're like the person that has to make sure the login is correct. That's a lot of engineering. And there's really like two things that come of it, but like the first one is stress and anxiety and pressure because you're running your job and you're managing your agent. But the second thing that comes of it is just lack of security knowledge. When HubSpot builds HubSpot and you log in, you don't have to think that every time you add a contact to your email, that's going to be leaked to the internet. But in the last five weeks, how many systems have been AI ready? Lovable Vercel, like how many systems have been hacked and all the data been leaked? Everyone that built on Lovable has some information leaked at this point. Or if you just put your API keys in the environment on the internet, then you're in trouble. But like that leakage is a problem. And so if junior at a company and you're the boss or you're the CEO and you allowed that to happen, you're also at fault. So I've been doing classroom sessions and I always ask at the beginning how many people have used AI and then how many people have like used AI to code something. And a very small proportion raise their hand. And then the immediate next question I ask is, Are you 100% sure that the thing you coded is not on the public internet? And they're like, I don't know. Like, is it? And I say, All right, open up your Claude. They're usually in front of a computer. Go to the toggle that when you logged in, does it say that it's gonna securely hide your private data? Like does it say that it's not going to train its model? And a few of them are always like, no, it's on because it's on by default. And I'm like, well You just trained the foundational model of Claude on how you do your processes internally, which means every one of your competitors would ask Claude a simple question, whether it's in the internal or private, you know, market at that point, is getting your answer of how you do work. Are you okay with that? And they don't think of that. And so if you're the CEO and those people made the mistakes, what about your junior that's just trying to get through his day to like walk his dog by five o'clock and he's also doing that? Are you positive that that toggle's turned on for him? So the governance here and the knowledge. Of how to secure an entire Claude instance or an open AI instance or a Microsoft instance or a Google instance needs to be learned at the same rate that Claude code is being learned to deploy things. That also means if you're in Microsoft or Google, you're probably fine and you don't have to do much because it's already inside of your walled garden. But if you're using Claude or ChatGPT or Codex or any of those, there is a whole laundry list of engineering deployment skills that you need to learn to be able to deploy your product. Either go learn it, hire an agency, or don't deploy it. Or don't deploy. I think there's two things you're hinting at from my side. One is it. This is going to be the new expansion of what IT does because ⁓ a lot of companies IT is the ones like, Hey, here's your computer, Johnny, and here's your login. Let us know if you have any issues. And it's almost like IT is now owning these instances of Claude. But also to your point, if they're limiting what Johnny can spin up and deploy. Like he can't just be like, all right, yeah, I just deployed my own solution built on top of our data, proprietary data, and sent the login to my cousin who's on the other side of the country. Through a text message. Yeah. And so or even worse, he just put it up on the web and then someone you know, it's like fully out there. So it's check my Reddit post really interesting. My Reddit post has the keys. It's really, really interesting. And I like your solution. You told me about what you talked about drive and Microsoft because it already has governance. I didn't even know what that meant. It has like rules of who has access to what already built in. So if you have ⁓ these internal AI solutions that allow people to vibe code their own solutions, but it's built into the Google and Microsoft ecosystem, you're you're pretty much covered from what you're saying. Yeah. And I think companies like us are scary from the outside because we can do a lot while staying secure. But does anyone know that we have a VPN set up? That our development team has it. And we have DevOps. We have three people in DevOps. They're awesome. And whenever we deploy a Claude code thing, like our head of product is deploying one right now, it takes three nights, ⁓ three days for us to finish a Claude prototype. I think she's on version 49. ⁓ she gave that over to DevOps. Three days later, she gets a secure environment prototype that now she has in her nine two three account. That anyone in the company can now use. And we did that with our stack on AWS. We did that with our security that we have across our company. And then we deployed it on the intranet, but it's like a VPN that anyone at our company can log in with Google authentication. We deployed this the same way that you deploy consumereports.com or you deploy your Uber app. And we're a development company. So when you see this from the outside, you're like, cool, I can do that too. You forget the three days of a professional DevOps that has gone to school to just do development operations. And like, Has to deploy this thing in a very specific way so that it is secure. Now, did you do that as a company? No. Do you have that as a company? No. But when people raise their hand and say like they're coding these things, I ask, Do you have an engineering department and do you want to be an engineering company? And they're like, No, no, no, no, no. I'm like, well, then if you need that, hire DevOps, hire the right people. Because the immediate next question to Sean is the thing I put out on LinkedIn recently: Will you pay $2,500 for that stupid prototype you put in Claude? And most people are like, no, $2,500 a month. And they're like, no. So Kill it because it's not a 10x return. If you're not going to pay someone $2,500 to ri to keep it alive, then it must be in value. Like it must not be worth it. If you're getting $25,000 of added benefit every single month, pay twenty five hundred dollars for somebody to maintain this so you don't have to be the midnight fixer. And I think that's the most important point. Wow. Don't be a developer. Don't be an engineering company. Like, like if you're not, I think ⁓ something we talked about. This is for non-technical leaders. And there are CTOs and things who want to up level their staff and get them on the latest version of Cloud Code or get them trained up. We we help those people too. But majority of the people we love working with and can create the most value for are non technical companies. Because they don't have like Ramp already has the internal agent. Block already has the internal agent. We don't need to go help these software engineering companies. It's the one who's in manufacturing, logistics, healthcare. It's like we're healthcare professionals. We're not engineers. And they have IT departments An IT department is different than an engineering department developing code for ramp. Like that's the separation. Yeah. And it's like become an engineer within guardrails. Cause I don't want people to tick come to this and be like, ⁓ you're basically telling everyone to not vibe code. That's ridiculous. It's it's too big of a wave. You're not gonna stop it. You're telling people at your company approach it differently. Yeah. I think the last topic we have is just basically about the change of course that The CEOs are finally making of these frontier models that we've been preaching for the last two years. AI is not replacing jobs, AI is creating jobs. And I think finally the news cycle is going to come to a aha moment and realize that AI is the same thing that the steam engine did, the same thing that LED lights did to the world, the same things that compute did to the world. I I mean, it's just a pattern we've seen as society. We've been preaching this for a year and now it's here. Talk to me a little bit more about that. Like what do you mean by that? Yeah, so in economy terms, it's called the Jevons paradox. And it's this concept that was created back in the steam engine days. And from when I s when I was looking at it when everyone was complaining that AI is gonna eat the world, the ⁓ James Watt, when he came out with the the steam engine, he created Obviously, a more fuel efficient engine, and they were using coal at the time. So the coal industry is like, what are we gonna do? Like we have these, you know, all these ⁓ trains going around the world that are using half the coal that it used to have. Coal's gonna go out of business. You should short coal. Maybe coal went from 9,000 to 500, like HubSpot did. That period of time, what actually happened if you study the history of of trains, more demand. came from trains because the coal was so cheap that more entries happened into the coal space and and into the train space. And so more products were shipped across the country because it was easier and more cost efficient to ship a product. More users were created across the ecosystem of the train network. That allowed people to create other businesses around the train network. And I mean, the same thing happened with LEDs. I I remember fluorescent lights, and we would put 10 lights in our house. They were all fluorescent. They cost a fortune. And my dad would tell me to turn it off. Now we have LED lights. And where do we put I put LEDs in my stairs? I put it on like inside of my like house. I put it like every 10 feet on the walkway instead of like every hundred feet or just a light post. We add LED lights everywhere. And so the demand For LED went skyrockets. Like I think it's 100x in the last 10 years or something like that. And so those types of products are proof that the Jevons paradox is creating an opportunity for people to enter into the market. When AI comes into the legal space, the hospital space, the manufacturing space, whatever it happens to be, and it removes low or 90% of somebody's job and it automates that, what does it do? It creates the cost for that job to be so low. That new entrants can come into that space. And with tokens of cost being so high, we're starting to see the opportunity for people to use both. And so the expansion of AI is happening across the world, just like the industrial revolution, just like the light bulb revolution. And now, just like the AI and compute revolution, it's utilizing the resources that were there and expanding those resources so more people have the opportunity to do the thing they never did before. And Dario, for the first time ever, said that. This week at the JP Morgan conference. And like, I didn't take it lightly. I was like, hurrah like I stood up. I was like, yes, like finally, dude. Like you've been preaching that you're gonna replace the entire world and you go on Fox News and like freak everyone out that they're losing their jobs. And then I have to deal with this on my next like seven proposal calls. Like, finally, we're creating opportunities by augmenting, not automating. So that is huge. And I'm finally glad that they're saying it out loud. Walking away from this episode, you should have hope. You should have. excitement about what's to happen. And if you want to figure out how to be one of the first movers in your industry, then you should click the first link in the description and apply to work with us and get a free AI audit and we can point out some opportunity to you. Andrew, thanks for another great episode. Anytime. Awesome. Good to see you all. See you next time.