Anthropic, OpenAI, and the Rising Cost of AI at Work

Every once in a while, we want to talk about certain issues around Anthropic Open AI, but then they just kind of bleed into other issues. And today, we are realizing that there's so many different kinds of things to talk about with Anthropic and Open AI that Jayden and I couldn't understand. So we decided to make it a cornucopia grab bag, choose what you're interested in type of episode. So Jayden, when I was looking at this, a couple of things were really standing up to me. First of all, post-election, all of a sudden the markets are going all the bonkers around AI, we can totally understand why I think, but that just just just days ago Anthropic was kind of like starting to get up to a 1.5 trillion valuation and all of a sudden they'd be like, whoa, now all of a sudden it feels like Anthropic is on their back heels, which is hilarious because even a trillion dollar would have been practically like record setting, right, which is crazy. And then Open AI at the same time, I kind of backed off their IPO and then all of a sudden Open AI also in the news wasn't able to raise as much as they thought they would raise all of a sudden these companies that look absolutely invincible. And as we know stock prices based on future earnings, not to someone's happening right now, this growth that was happening in the summer suddenly is not growing in the same way month over month for Anthropic growth because people are being more careful because people are looking at ROI, but also Jayden, there are so many new features with Open AI and Anthropic. So I don't know, I kind of want to lean into the financial side of this first because I think it actually has implications about how people see AI and what happens with the future of the two biggest companies, not exactly on the planet, but close them in Anthropic and Open AI. So you tell me where do you want to go on this? Okay, well, there's a couple interesting new things that came out. First of all, the markets are in a rough spot right now. So this possibly is the worst time ever for Anthropic to do their IPO. Maybe things will will improve a little bit. We have Jim Kramer assuring everyone that the AI sell off is not another.com crash. If you know anything about Jim Kramer, basically everything he says the opposite happens. So everyone's retweeting this and saying like, we are very, very doomed for Jim Kramer. I don't know if you saw there's, have you heard of inverse Kramer? No, I love that though. Basically, there's like, there's like funds that track everything he says and they trade the opposite way. They'll short whatever company he promotes or they'll buy whatever company he's like crap talking. And the inverse Kramer is like, it's like, you know, like 100% return every year or something crazy. Like it, I think it was like 72% for the last year, vastly outperforming the market. So anyways, so it's a tricky spot for them. And yet we have a bunch of good metrics. And I think opening, I might be the one that's actually perhaps setting themselves up to be in a really good position for 2027. We had Greg Brockman come out and say that they believe they're in a good in a good state for 2027. They said that basically they were in a pretty bad financial place for Q1 and Q2 of this year. Greg Brockman did an interview. He said Q3, they turned it all around by coming out with dots and then their new AI model Astra and I think a lot of the progress they made with like chat GPT work, which if I mean honest, that's kind of the timeline for me going all in on Anthropic and completely forgetting OpenAI to switching back over was along that same timeline because OpenAI does have a lot more features. I mean, they have the audio video image. They had a lot of things that Anthropic didn't have, but at the end of the day Anthropics co-work was just unbeatable and their clawed code was just so good. So I feel like Codex and everything's brought back. OpenAI is now doing pretty well. And they said financially they've also made a huge shift. They've copied Anthropics playbook. And apparently Greg Brockman said that Enterprise now accounts for more than 50% of OpenAI revenue where in the first two quarters, it was primarily just everyday users that just had lots of random people that were paying $20 a month to use chat GPT. And Anthropic was getting more and more people, Anthropic went through a list of who their top buyers were. And they're like, we have X amount of buyers that are paying $100 million a year companies. We have X amount of companies that are paying $10 million, X amount of companies that are paying $1 million. And OpenAI did not have that huge Enterprise lineup. They were really just at this core of everyday users. And so they've now made the switch just over 50%. I think they're going to be in a good financial position. And I'm curious to see Anthropic is going to have to release numbers if they go public. So I'm really curious to see what their numbers are. They were on an incredible run. But I'm curious if they plateaued if the whole market as a whole grew or if OpenAI was able to thwart a little bit of their growth. So it's going to be an interesting time in a bumpy part of the market for sure. Yeah, it's it's also you know, it's pretty easy to sort of see when the market is flying up and down, right? And I was like very connected to the elections, obviously. And but with this, with the IPO stuff, it's a little stranger, right? Because I know that I think it was ramping. There was a there was a thing I don't have it open right now. I don't think but like TechCrunch talked about this. We're all the sudden they were realizing that again, the sort of like the month over month token usage where Anthropic was making most of their money. We're on the super users. They're as yours are referencing, right? They were it was it was these top companies just using a tremendous amount of this. But Jane just anecdotally and this means nothing. I'm not trying to put anecdotally over what we're actually seeing in the market or anything like that. But I will say this when we say that OpenAI is now over 50% enterprise usage, it's got to just be encoding, right? Am I crazy because when or software companies because when I'm working in so at AI mindset, we work kind of like across the board and we're working a bunch of software companies and all of those companies for the most part are using sort of a combination of cloud and codex. But when I'm just like talking to like really big companies at scale, like 20, 30, 40,000 people, they and even and I won't name I won't name the latest client which is like 100,000 people. But the but the but it is a well it would give it away. It's a it's a company that does have a lot of technical it has a 30,000 person technical staff for example. And when we think about that, they're all using co-pilot still and when I go into law firms, it's often a shootout between co-pilot and then some of the legal text of Harvey and Legora which are growing then wildfire right there huge. But everywhere I go, copilot, copilot, copilot, we just filmed and we have this whole AM mindset course. It goes out to a lot of people and just like last Friday, we just filmed with Microsoft, the new version of that because it's moving so fast, co-work is inside of that. But even inside that, people are getting worried because inside some of the best stuff inside copilot is co-work, which is very expensive. Anyway, my point is that when these companies are saying that it's over 50% enterprise, which is great and that's how they make their money, they're not as far as I've seen anecdotally, they're not giving clawed to everybody, they're not giving a chat GVT to everybody. That's not what I've seen. I've seen them giving co-pilot to everybody and that's just where I'm wondering these numbers are just so huge and it must be based on just when people are building it into product, when people are using it to code, when people are using it for a lot of other ways, they're just using, it's not that everyone is using it. It's a group of people are using it so much. Is that what you're sort of seeing as well or are you seeing something different? Yeah, I think there's definitely industries where it's super, super saturated at one interesting point to fit you brought up as far as who's giving it. There's an interesting stat that Apple is apparently cutting their budget, their daily spending cap for their engineering team, they're decreasing it from 300 to 200 dollars and then they're going to cut it to 100 dollars. So it's interesting that we're seeing some of these companies actually cutting back how much their engineers are allowed to spend on some of these tools, but you can imagine if they were spending, if they did have a $300 day allowance, they're cutting that to 200. It's going to hit the bottom line too and the graphic probably is who I'm assuming most of this is going to. Perhaps open AI, but I think we're going to, yeah, we're going to probably see, we'll see a pullback in usage, maybe a shift to open source models. If companies are really worried about the ROI and some of the spend, which admittedly can be pretty out of control, and I think another thing that not a lot of people think about is the fact that you can spend $200 a month on anthropic and open AI and like that's what I do. And I'm like, oh man, it's expensive compared to every other subscription I have. And yet, look at Apple, their daily spend cap for their engineers was $300 a day. If you're not on these subsidized max plans and you're just going straight for the API usage, which is kind of what like everyone of a lot of the big companies are doing, it can be insanely expensive. So I do think we're going to see people draw down their spending. They'll try to find ways to save money, but I don't think the usage of AI is necessarily going to drop. I think you get so much value out of it. I think people are going to try to find ways to maybe use open source models. Like there's just so much that you can do and it can accomplish so much for you. But yeah, to your point on like people using like co-pilot and other tools like that inside of their organizations, I would be curious. If I worked at a company and I'll be honest right now, I'm sure people maybe some people inside of companies will hate me for this, but you got to you have to believe that there's probably a lot of people with the same mentality. If I worked inside of an organization that said, look, no one's allowed to use open AI for security reasons or anthropic or maybe even we don't really care what you use, but we're only going to pay for you to use co-pilot for because that's just what we have sees for it. I would personally go pay for it. I would personally go get my own laptop and whatever I could accomplish with that and my own laptop, I would do it. It would help me do my job better. It would help me do my job faster. There's no question. That's 100% just what I would do. And I know I can probably a lot of people in that boat like co-pilot or you know, chat GPT work and anthropics co-work, cloud co-work, they can control your computer and literally do anything you need. So if your company isn't paying for those tools, I think a lot of people are going to use them. I'm curious when they say enterprise, what percentage of people are just paying for this themselves because $200 a month, if you have a job that's paying you 10 grand a month, why would you not do that to make your job 20% more efficient or save yourself. There's - The only thing I would sort of push back just of late is that co-pilot over the last six, seven months has gotten very good. It's, I think that co-pilot had a rough start. I think that they would sort of probably admit that too, just because they're up against these incredibly powerful models that had no risk. - Microsoft's your saying? - Microsoft co-pilot, yeah. - Is it, I control your computer? - Yeah, I mean, now it can't, I think that it's not able right now, but what it does have is the same agentic stuff that dots and things like that are building out. They're just building it out now. Again, Microsoft's Google, and we've talked to some of the top people at Google over the last six months as you're getting on this podcast, but I have to move slower. They can't be like wild Westie. Oh, hey, that didn't work. Let's pull back, right? Because they're integrating it into just hundreds of millions of computers that are running the enterprises of the United States and beyond. So they have to be careful. They can't do what Cloud Code does for me or Codex does for you or whatever, which is just take over your computer, which is the single best thing ever. But it's also, but, Jayden, just, and I know we're just about a time here, but can I just like pings that you have open on the screen right now that we're both looking at on TechCrunch? This is about the Apple thing. When you say like, or not you're saying, but the article is saying in October, Apple, lower the daily spend from 300 to 200 and then 100, maybe go even going down to $25 a day. It just, and I want to leave the audience just with this question. Just think about that for a second. These are the, these Apple, if you're not using AI a lot, just know that Apple is trying to restrict people from using $300 a day. Now, that's an engineering, which is a little bit different, but it really does start to beg the question of, are, is this divide going to continue to grow? People who are just using AI crazy versus people who are not. - Yeah, 100%, I think we're, I think we're definitely going to see that. And if your organization is wondering, what are the best ways for us to think about how much we should be using it? Or maybe you're like, look, our organization definitely should be using it more, but how do we really work that in? I would 100% recommend you check out Connor's AI mindset course that he has. He does trainings and courses. He onboards companies that have 10,000 employees that will all take it at the same time. And way, and much bigger, and he's done partnerships with Google, Microsoft, and all sorts of future organizations. Please, if you have not taken this already, bring this to your department, bring this to your organization. The AI mindset course from Connor is one of the greatest things I think you can do to help up skill and make sure that everyone in your organization is getting the maximum usage and value, and your company is going to make the most amount of money from AI. So go check it out. We'll leave a link in the description. Guys, thanks so much for tuning in. Leave the podcast, our rating and review, wherever you get it. We appreciate you all. We hope you all have a phenomenal rest of your day.