Episode Transcript
[00:00:05] Welcome to Prompting Curiosity, a podcast for the AI curious. No coding background required. I'm your host, Dr. Shantae Cofield, also known as the Maestro, and I created this show to explore what these AI tools actually are. Really, though, are the files in the computer, how to use them, and what they might mean for how we think, work, create, and move through life. Whether you're skeptical, intrigued, or already experimenting, you're in the right place. All that I ask is that you stay curious. All right, let's get into it.
[00:00:37] Hello, hello, hello, my curious people, and welcome to episode 55.
[00:00:44] Zero of prompting curiosity. Episode 50. Y'. All. Let's take a moment, a quick moment. I'm gonna take a quick moment to celebrate because it's dope. I'm grateful, and it's important to celebrate your wins. Right? Thank you for real. Thank you for being here. Thank you for the ratings and the reviews. Thank you for listening, thank you for reading, thank you for responding, and just thank you for being willing to be on this journey with me. Right. It's pretty wild to see how much has changed with AI since episode one of this podcast, which was, you know, a little less than a year ago.
[00:01:16] The, the, uh, the AI advances are slowing in some regards while AI capability is rapidly increasing. Uh, and honestly, I said it a billion times, but it's true. I really do believe that this tech is here to stay. It's not going anywhere. So here, I guess, is to another 50 episodes. So, like I said, welcome to Prompting Curiosity. I'm your grateful host. This is our traditional intro. I'm your grateful host, Maestro, and today we are talking about to celebrate 50 and to celebrate the changes in AI, we are talking about the impending AI price hike. So this price hike is more speculation, extrapolation and common sense than it is some sort of insider knowledge that I have. But I have brought this up like 11 billion times before in the podcast, and I'm going to bring it up again today and dedicate the entire episode to it, because it's not a matter of if, it is a matter of when.
[00:02:14] This episode, fully inspired by Microsoft's recent rollout of Copilot coworkers, uh, they rolled it out to the general public and they announced with that rollout that users would be billed for Cowork on a usage based basis, which is determined by the tasks that they run.
[00:02:40] So for those of you who don't know, uh, because you don't use it, Copilot is Microsoft's LLM and Cowork is their agentic version of the model, which means that it can actually execute tasks. Asks for you. Right. It doesn't just answer questions. It runs in the cloud, not locally on your computer. And it works within the Microsoft suite.
[00:03:01] So this usage based pricing, it's a huge shift from the standard pricing model utilized by basically every AI uh, company which is that predictable flat fee per month.
[00:03:13] And so to clarify, right, regarding again, this is Microsoft's copilot Cowork.
[00:03:22] Users will pay for both, right? Users will have to pay a flat subscription fee just to have access to Cowork.
[00:03:28] And then on top of that they will be billed based on their usage of Cowork.
[00:03:33] My prediction is that the rest of the companies follow suit. Right. Anthropic already has usage tiers. I've spoken about this before. I think it's a good thing. They have usage tiers, they have usage limits.
[00:03:43] And for a very, very brief stint there I never covered it because it like, it came and like went away very quickly. But for a very brief stint there was talk about them removing Claude code from the pro tier altogether. Right. The pro tier is what I'm on. It's the lowest of their paid subscription tiers, additionally tying into this prediction of the rest of the companies following suit. If that name Cowork sounds familiar, it's because it's exactly what Anthropic calls their agentic model. That's the one that they built for, you know, the everyday person.
[00:04:15] They have Claude Cowork and it's not a coincidence. Microsoft worked with Anthropic to create Copilot Cowork and they integrated the technology that's used for Claude Cowork. Right. So you can kind of summarize that as Microsoft's cowork is built like on top of Claude called cloud code and Claud co work. Right. They working together here.
[00:04:38] All right, so speaking, actually before we go into the next part, speaking of Anthropic, right. Because I did a little last week's episodes about this. Speaking of Anthropic, uh, like those are the, the creators, the owners of Claude Fable 5 is still down, right? It's still uh, what is the word that I want to want to use? Suspended. And there's no reports on if or when it will be live again. I did read something about mythos, mythos 6 and kind of a leak, but like there's no specific word on, you know, when this thing may, may be allowed back into the public.
[00:05:13] Um, but when anthropic did release Fable 5 back when it worked. Right. And we had access to it. When they released it, they announced that it would be paper usage billing. Right after that 2 ish week trial period, right? They were pushing, you know, use it, try it, try it. But On I believe June 22, it was going to go away and then it was going to switch to pay per usage billing. All right, doing, they're doing to being a good drug dealer, right? Get you a little try before you buy. They get you a little taste, get you hooked and then charge you. All right, so like I was saying before, cowork, same name, cowork, right? We got Microsoft copilot co work, we got Claude cowork. Why? Because they are using the same thing, right? Microsoft worked with Anthropic for that new model. Uh, and because they're working together, when one company sees another company doing something, even they're not working together. When one company sees another company doing something, especially if they're working together, it feels inevitable to me that other companies won't follow suit. They're like, oh, you can make more money, of course we're going to do that as well.
[00:06:20] So when will the other companies switch to the usage based model? Again, I have no idea. This episode is is largely a compilation of speculation, extrapolation and common sense, but it's absolutely coming.
[00:06:32] I want to introduce you to a term that further speaks to this impending price model change, and that is token maxing. If you've been in the AI world or on the threads, things like that, you've seen it before. But token maxing is the word that's used to describe the most tech bro, bro chacho way of using AI, which is simply burning as many tokens as possible per interaction with AI for no fucking reason at all besides just using the most, right? As a reminder, taking it back to the way, early episodes of this podcast, tokens are the foundational, quote unquote language of LLMs, right? Words. We type them in, they're broken down into sequences of characters and those sequences of characters are referred to as tokens, right? Tokens also serve as the billing unit of AI models where usage is measured in tokens, right? So token maxing means being as wasteful as possible with AI. It is a very, very, very brochacho approach to using AI and it is officially coming to an end. Why?
[00:07:34] Well, the biggest offenders of this token maxing seem to be enterprise users, meaning people whose companies paid for their AI access and their AI usage.
[00:07:45] Many of these companies get this folks. Many of these companies, they gamified usage, right? They had like Leaderboards and, and they were just pushing their teams to use the best models and move the fastest and use the most tokens, costs be damned. Because it was a flat rate.
[00:07:59] Right. CEOs though, of these companies, they have started singing a different tune this year as that all you can eat pricing has been replaced by token based pricing, which when you pair that with the advent of these, these more advanced and token hungry models, this has had companies like Uber burning through their entire 2026 AI coding budget by April.
[00:08:25] By April. Not going to lie, it feels like a bit of poetic justice that Uber got Ubered. You know what I'm saying?
[00:08:35] So what we're seeing now is these companies, they're doing what they should have done in the first place and they are encouraging their employees to be mindful about usage. No, right. Uh, this is the approach that I have always recommended and that I will continue to recommend and suggest for you.
[00:08:50] Honestly, I think it's as a good approach to life, like why use more than you need? Right?
[00:08:54] Additionally, I've said this many times, but the most basic model is very likely more than sufficient for your needs. All right. As a Claude User exclusively, Sonnet4.6 is my daily driver and it's what I use for Vibe coding.
[00:09:07] I don't need anything higher than that. They try to push, you know, use opus. Use Opus. That's just token hungry as. No, thank you. I don't need it. I, and I don't want to speak for you, but my guess is that, you know, it would be, you know, you using these higher models would be a bit like having a legit military grade Hummer when all you do is just drive down the street. Right? That's driving to the mall. And yes, I'm saying this is someone with a Jeep Wrangler. I don't have a Hummer though.
[00:09:34] Right, I'm aware, I'm aware.
[00:09:37] So you know, Anthropic already has usage based limits.
[00:09:43] Uh, and the per usage pricing model does exist for all tiers as it currently stands.
[00:09:49] Um, but that's once you've exceeded your standard usage total. So you pay your standard fee $20 a month. Once you've exceeded that, you can pay to have access to more and you can do it on a per. Per usage basis.
[00:10:05] Um, and so I think that they're going to expand on that. I think that there's many ways they can do that. They can throttle things and even if they do it gradually and they throttle things gradually, we're like, man, this same task is seemingly taking up A little bit more of tokens than it did before.
[00:10:20] Um, I think that, you know, I think it's just a matter of time before this per pricing, per usage pricing becomes, you know, the default becomes more standard. So why not let's get ahead of it?
[00:10:32] I get ahead of it, meaning let's use less, let's need less, and it'll cost you less.
[00:10:39] I will also say I'm not opposed to paying more for this tool. I find this tool incredibly useful for all that it does. Right. Especially when I compare it to the other software that I use for my business.
[00:10:53] Um, and you know, to me, you know, paying like a hundred dollars a month wouldn't be unreasonable. I am quite annoyed that all these other companies don't just keep raising their prices and they don't, they're not like giving me any benefits or changing anything. But for all that I do with AI, you know, I'm comparing the pricing of these other things that I'm m not even using that much and what their price point is. And I'm like, oh yeah, like I use this thing like daily or, uh, you know, multiple times a day or every other day. Like I'm using it a lot. And I'm like, yeah, this would not be unreasonable to, to have the price point cut up. But, you know, what are your thoughts on this? I would actually really love to hear your feelings about cost and if you'd be willing to pay more for, you know, this LLM and if so, how much more?
[00:11:34] So to summarize, the AI price hike is coming, especially as the VC funding dries up. When is it coming? I don't know. I don't know.
[00:11:46] But that money, man, AI money continues to be funny money, circular deals. Circular money, right. Anthropic. Did just file for, for an IPO XAI purchase cursor, which I think is terrible. I think it's like a very big deal and it's very bad.
[00:12:03] Um, and you, you probably have no idea what I'm talking about. If, uh, you care, you can to learn more, you could, you could search it, but I'm seeing very little to no talk about it. And I'm like, this is actually very bad.
[00:12:15] Um, you know, the valuations for these AI companies is through the roof. This is crazy.
[00:12:23] But the price hike will happen, right? So, you know, why not do your best to quote unquote, get ahead of it in any way you can. All right, all right, last things. The last how I used AI this week. So if you're new here, each episode I share a quick episode. A quick episode. I share a quick example. Wow. Quick example of how I. Or in this case, someone I know, which is you find people used AI this week. So this week I'm talking about my guy John.
[00:12:50] Uh, he goes by J Code, and he actually used AI, more specifically Chat GPT to save his Mac. So J Code got himself in a bit of a pickle trying to use an external hard drive as an internal hard drive. And, uh, basically everyone at the Genius Bar told them that there was no hope. They were like, everything's corrupt.
[00:13:08] Like, just, you're gonna lose everything. Just.
[00:13:11] Just accept it and move on. I think. I think he went to another place as well.
[00:13:16] Uh, and it was just like devil's, like, you're gonna lose your. It's. It's done. You can't.
[00:13:20] So my guy J Code teamed up with Chat GPT and he spent a few hours in the computer. Uh, this was not like a five minute task. He was like, you know, doing things, entering Code in the back end of things. Like he was. He was putting in work.
[00:13:33] Uh, but he got everything fixed. He didn't lose anything. Uh, and his Mac is now fully functional.
[00:13:41] He shared the story on Instagram and he gave me permission to share it here because he's the fucking best. We talk all the time. Um, but, uh, while, you know, while your mileage may vary, this is a damn good use case for. For AI. So, yeah, there's that. All right, that is all for today.
[00:14:00] Hopefully you found this episode helpful if you did consider leaving a little rating or review. We are actually at 31 ratings. Let's go.
[00:14:10] And.
[00:14:11] And we got a new review from, quote, unquote, marketing listener. That's the person's name. Uh, that person said, I have a couple of AI podcasts that I follow. Sometimes the voice on those other podcasts is so monotone that I get distracted and miss everything they said. This show is a lot like talk radio. Very common sense, not overly technical, but always learning something new every episode. Thank you.
[00:14:35] Thank you, marketing listener, whoever you are.
[00:14:38] I read this and it really does. It makes me. I smile so big. It really makes my day. Right? Podcasting is such a unidirectional medium. I'm just sitting here talking to the screen. Every now and then, maybe you hear something and it's usually Rupert, and I'm throwing him off the desk. All right? But I'm just talking to the screen and it really, truly means so much to hear from you and know that this stuff is landing and it's helpful and just, you know, that you're listening, so just thank you, truly. Thank you. Thank you, thank you.
[00:15:04] All right, I'm gonna wrap it up there. As always.
[00:15:07] Endlessly. Endlessly. Let's give it to you one more time. Endlessly appreciative for every single one of you until we chat again next Thursday.
[00:15:16] Stay curious.
[00:15:18] Sam M.
