Ep. 31: Are We Addicted to AI?

[00:00:00] Foreign. [00:00:05] Welcome to ChatGPT Curious, a podcast for people who are, well, curious about ChatGPT. I'm, um, your host, Dr. Shantae Cofield, also known as the Maestro, and I created this show to explore what ChatGPT actually is really, though, are the files in the computer, how to use it, and what it 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. [00:00:38] Hello, hello, hello, my curious people, and welcome to episode 31. I am your grateful host, the Maestro, and today we're asking the question, are we addicted to AI? And I'm saying we and lumping myself in there because we don't throw stones around here from this glass house that I am sitting in. So we're going to jump right on in. I am using Claude, right? I am using LE Claude and I am trying to make the switch to using it. And this discussion right here, this little point right here is somewhat tangential to the main topic. Um, but it really. [00:01:16] But it's important because in trying to make the switch and in looking at anthropic, right, Looking at using Claude, switching to using Claude. And why am I switching? If you didn't listen to last week's episode, there is a proposed boycott of OpenAI. [00:01:29] Um, you know, they're the most vulnerable to, uh, a boycott and people are talking about boycotting it, you know, and I'm going to get into the reasons, but I have since started using Claude and like I said last week, and I still feel this way. [00:01:47] It feels like out of the pan and into the future fire, right? Switching from OpenAI, switching from ChatGPT to Claude 100% feels like out of the pan and into the future pyre. I said this last week, I'll say it again, I still feel that way. OpenAI, yes, we know they are objectively lowers their co founder, Greg Brockman. He donated like $25 million to Trump, which worth noting was back in September, this didn't just happen. [00:02:13] Um, but they are also like, you know, they push back against AI legislation. Sam Altman's fucking weird. [00:02:18] Their software is being used for ICE application screenings. [00:02:22] And, you know, they, like I said, they are the most vulnerable to a boycott. Majority of their revenue comes from paying customers and paying customers are only like 5% of the total customers they have by most people using the free tier. But either way, that's where they're making their money. By making money, I mean, they're losing a ton of money. But if and when it comes to voting with your dollars, trying to send a point, you know, send a message. [00:02:44] But boycotting OpenAI would be the most effective. [00:02:49] However, I do feel that it is from an ethical and ethics perspective, it is out of the pan and into the fire. Why? Because Anthropic, while they don't have any Trump donations, they are big time in bed with Palantir. [00:03:05] So Palantir. Excuse me, Anthropic partnered with Palantir in 2024. This isn't even new. 2024. [00:03:12] Right. And they uh, allowing Palantir's government and enterprise clients to deploy Claude models within their Palantir. So those. For those of you that don't know, and maybe you've heard this name, and my guess is that if you've heard it, you haven't heard it in a good way. Palantir is a data analytics company. They build surveillance and intelligence software primarily for government agencies, military and, wait for it, law enforcement. [00:03:37] So why, if you can't tell already, is a partnership between Palantir and Claude controversial? Well, let's give you a few talking points. One, ICE contracts. Palantir provided software to ICE that's used in deportation operations, including tracking and arrested undocumented immigrants. They are. They're. This is worse than just screening fucking resumes. [00:04:03] Military applications. [00:04:05] Palantir works extensively with DoD intelligence agencies and foreign militaries on targeting and surveillance systems. [00:04:12] And what are they going to be using, Claude? Predictive policing. This is so fucked up. And this is probably possibly. If you've heard about, uh, if you've heard about Palantir, it's through this. [00:04:24] Their software has been used by law enforcement for predictive policing programs criticized for amplifying racial bias. We think you might maybe commit a crime. We're going to do something about it. That is bad. Anyone seen minorities report? Yes, this is. This is bad. Uh, data integration at scale. The core product connects disparate. That's the. I'm just reading this verbatim this part. Data sources to enable mass surveillance capabilities that weren't previously possible. Mass surveillance capabilities. [00:04:53] This Palantir. [00:04:56] Lastly, lack of transparency. Palantir operates with minimal public accountability about how their tools are used by government clients. Is bad. [00:05:04] And Claude is all. Or Anthropic. They all up in bed with them. Um, so leaving ChatGPT to go to anthropic, that 100% feels like out of the pan and into the present fire. [00:05:18] And I call it a future Fire. When I say it, when I've said it before, because folks don't seem to be aware of the anthropic palantir relationship. So, like, when they become aware of it, then it's like, oh, shit, this is bad. So I don't feel any kind of way about. I'm not going to use the word. I'll say flip flopping. Like last week I was like, hey, if you want something that's, that's immediately better on paper, I guess it's going to be anthropic. If you want to join the boycott and things like that and vote with your dollars and look to send a message and probably have more of an impact, it is leaving OpenAI. And if you're going to move somewhere, moving to anthropic. [00:05:57] But like I said last week, right from the ethics perspective, from the take a step back and look at, like, what is actually happening and where your dollars ultimately go. And like, are you just kicking the can down the road? It's this, right? Leaving ChatGPT to go to Claudette is out of the pan and into the future. Fire. So what am I going to do? I don't fucking know. [00:06:16] Am I taking all of this as a sign to change the name of my podcast and everything? Because if OpenAI is anything like CrossFit, they're going to come for me at some point because I'm using the word chatgpt. [00:06:26] Uh, yeah, for sure. And I have a few ideas in the, uh, in the old noggin. Perhaps you were thinking when I said, am I taking this as a sign to stop using altogether, that would be the right thing. But, like, you know, as it relates a good segue, as it relates to the bigger question of what am I going to do? That is a good segue into what I want to discuss today, which is, are we addicted to AI? The answer is yes. Yes, yes, yes, we are. And by we, I mean me. Um, um, I've been using Claude exclusively for like two weeks, like I said. And there's a super interesting feature that I basically want to talk about exclusively today within Claude that is called usage limits. ChatGPT, their paid tier, their first paid tier, the $20 tier. You can basically use it as much as you want. $20 a month gives you pretty much basically unlimited, uh, usage. [00:07:19] And I've talked about this before, like, you never get throttled. It's just like, okay, you just one time it was like, hey, you've been using me a lot. Do you want to maybe take a rest and I was like, no, I'm coding, right? But it doesn't stop you and it doesn't charge you more. [00:07:34] And that, my friends, should be a big red flag to all of us. How the fuck are they paying for this? The answer is that they're not. They're not paying for it. The investors are the investor. Money is going towards this. OpenAI is projected to lose 14 billion. That's with a B. $14 billion in 2026. Why? Why are they doing this? So they can get you addicted and lock you in and recoup their losses down the road. It is straight out of Uber's playbook. I have said this in previous episodes, right? Uber took 14 years to become profitable. They were founded in 2009, and their first annual profit was in 2023. In 2022, they had a $9.14 billion loss. [00:08:17] They spent billions of investors money subsidizing these rides and making them below. Below cost. Right? Some cities were paying drivers. I do the math on this. Some cities were paying drivers 1750 an hour while charging their customers $9.10 an hour. [00:08:34] I know that it goes per ride, but if you do the math and do all the efforts, things like that, paying 1750 per hour, charging customers $9.10 per hours. The math. O math. [00:08:44] How did they turn a profit in 2023? [00:08:47] Well, they cut driver subsidies, they raised their commission rates, meaning how much they took of what the money was coming in. So the money pay goes down for the drivers. They did improve operational efficiency, which I think means they did some bullshit. Uh, they did diversify into Uber Eats. So worth noting, something else. And they've raised their prices, and people kept taking Uber despite the price increases. Why? Because they were hooked. They were locked in, addicted, uh, dependent on it. OpenAI is doing the same thing. [00:09:21] The reason I want to focus on Anthropic's usage limits is that they have them at all and they enforce them. [00:09:30] I think this is a very good thing. So on desktop, if you're using claude, you can go to. On the desktop, you can go to Account, you go to Settings, and you go to usage, and you'll see your plan usage limits and how much you've used of those limits for both your current session either, uh, you have a current session limit and you have a weekly limit. All right, yes. There are two current session limit and weekly limit. [00:09:54] If you're using one of the more expensive models within claude, yes, that allotment will get eaten up faster. And I had seen people that are on the the, the more expensive tier. A lot of the coders talk about this and then the new models come out, kind of like, you know, with um, your iPhone. As the new models come out, they take these better pictures, but they are massive and they're massive videos, right? It's better quality, but it's massive. Same, same for this. New models come out, but they use a lot of tokens, right? They're really expensive models. [00:10:22] So yes, you can pay for more usage. You can upgrade to the max plan which starts at 100amonth. You can also enable it so that when you hit your limit, let's say you have your own 20 plan, you can enable it so that when you hit your limit you switch to what's called consumption based pric and that's at a standard API rate and you can continue doing your work and stuff without interruption. Um, so like I said, I've been using Claude and you can literally see your allotment get eaten up in real time. Like there's a little blue progress bar thing and you see, you're like, it's getting filled up. And I am not mad about this though it would be nice to have that little bar on the main page, not like hidden inside of settings and gets really annoying. But either way its existence I think is a good thing and has prompted three big, three big thoughts for me. So number one is the title of this episode, are we addicted to AI having this little bar there? I'm like, wait. And I'm paying attention to stuff like am I addicted to this thing? [00:11:18] Two second thought, it feels like Anthropic is positioning themselves very well from a financial perspective and actually charging people to use this thing and creating limits. I think it's a good thing. I think it's a good model. I support that they have this. [00:11:33] The third thing that uh, this usage limits made me think of was when is OpenAI going to raise their prices? I've talked about this before. I really clearly, I really like talking about this stuff in the financial and economic side of these things. But they gotta raise their price. They have to do it at some point. And I could see this coming in a few different ways, right? You could, they could significantly throttle every tier that they have, the free tier, the paid tier and the like higher pay tier. [00:12:05] And then you feel like, you know, you need to because you're throttling. When I say throttling, it meaning like you hit. They could actually have limits on OpenAI, right? They don't have limits right now, but they could impose them and, you know, it wouldn't be like, they can't not allow you to do that. So they can impose these limits, you see, with the free tier, right, you could make it so that, like, it locks you out much sooner. And there's been rumors, I see on threads of people talking about that. Claude does that, um, that it's just like suddenly people are seeing, they're like, before this used to use way less, and now it's same query and same prompting and same actions are using more. But we could see that from the free tier perspective where, you know, you get capped much quicker and that makes it feel like you need to pay in order to use it in a way that's, that's enjoyable. They could put limits on any of the, any of, any of the tiers. Now, this happened to me with YouTube and their ads. Like, to me, it just felt unusable to watch it. And I really like YouTube and I was like, you know what, I'll pay. And they came out with like a lower price tier, which we saw OpenAI do, right? They came out with their like, uh, chat GPT go. When I, whatever I think it was two or three episodes ago, I talked about chatgpt rolling out ads and a new tier. [00:13:11] And you can see this. They'll just throttle things. Maybe the ads become so pervasive that you're like, man, I'll just pay to not have these things. [00:13:18] Um, they could throttle the paid tier and they could introduce a new higher tier. So right now, you know, you have the $20 tier and there was like a pro one that was like $300, and I was like, that's too big of a jump. But they could introduce like $100 tier that has significantly, quote, unquote, significantly increased usage limits. And that would push the folks who are already paying $20 up to that next tier. [00:13:42] But either way, they're going to hit us, hit us in some way, shape or form. And yes, like I said earlier, clearly I very much enjoy the business finance side of this. [00:13:52] Um, but to circle back to the question of are we addicted to AI? [00:13:57] Putting usage limits on something is the easiest way to tell. When you say you can't use this, suddenly it's eye opening and you're like, but I want to use it. Wow. [00:14:08] So for me, I really love using it for coding, building things. I love, um, learning coding. I really, really love it. [00:14:16] Right? And I love using it. I will say, as just a thought partner that I can word vomit to. It can keep up. I speak really quickly I slow it down for the podcast so that people actually listen. [00:14:25] Uh, but I know people like to put things on, like, 1.5. X and they're like, I cannot do that with you. So I try to slow it down so that you can put on 1.5x and, like, you still understand it. And. But like, I speak, I speak quickly. I end with ChatGPT. I never have to repeat myself with Claude. I know I never have to repeat myself. Although. Ask a few more questions. They are different, man. Like, I will do an episode about, like, you know, switching and, like, my experience. They're different. And sometimes it's not a bad thing. They are different. [00:14:50] Um, but, you know, as relates to, to chat. [00:14:54] Yes. I love using it for coding. I love using as a thought partner. It can keep up. It never asked me to repeat myself. It quote unquote, it knows me, you know, which became really apparent in switching to Claude. And it doesn't know me in the same way that chat does. This is a familiarity. Like, it is a computer program. I understand. I'm not like, oh, this human, this, this thing that I love. No, Like, I understand that the computer program has stored the data and the, the numbers that it associates with me. And I'm like, yes, but the addiction is real. Right. And I don't necessarily have any suggested action items here. And that wasn't really my intent with this episode. Like, to provide, like, and here's what to do. Ah. This was what prompted. This is again, just me being curious and sharing the things that are in my head. Right. Of course, when I do have action items, I, I will share them, and I do share them. And I think that even noticing this and asking ourselves this and thinking about this is a valuable first step. [00:15:44] Um, you know, I, I, the addiction is real. I'm out here with a whole ass podcast, making a whole, you know, podcast episode, trying to figure out and discuss, be curious about which is the least shitty of two really shitty companies. [00:16:01] When ultimately I know the answer is to use neither. But when you're addicted, you're like, I don't. Just don't want to quit it. I need to, like, ease off. [00:16:08] Right. The answer is to use none of it. It's to walk away from LLMs altogether. But I can't do that. Yeah. Because I'm addicted. I am human. Human. Which I should also note, being human means being resilient. Right. Humans are adaptable. We survive before AI. We can survive without it. Uh, and I, I don't want this to sound like I'm addicted. I never, and I could never do anything else. And like, no, I'm just being honest. [00:16:30] I'm wearing my headset right now. Right. I don't, I don't want to see this technology go away, but I do think that the pendulum will swing back the other way with people using it less. Right? It's a dopamine hit. And as, as soon as the dopamine becomes less dopamine, you're like, it's great. And we'll see like it's just being used like a million in all the, in all the ways for the people that are actually, actually using it. But I, I say I foresee the pendulum at the other way, you know, with less usage. And yes, I do think that regulations are a very good thing. [00:17:01] A very, very good thing. All right, so to that end, Anthropic's users limits is in my opinion one of the best second steps. Like the first step, uh, the first action item we talked about before in terms of this addiction is awareness. [00:17:14] Uh, but the second is having usage limits because when you have to pay, you pay attention and you're not using it all willy nilly. Which is why I'm um, a stand 10 toes down on the fact that OpenAI is 100%, 100%, 11 billion percent trying to create that addiction and subsequently that lock in so they can make money off of us in the future. [00:17:37] Right, but you don't use Anthropic in my opinion. You don't use cloud all willy nilly because you can see that blue usage bar getting filled up, which again, I know this isn't like a finance podcast, was really interesting to me from the business side because you know, folks talk about LLMs replacing Google searches and it's like, not if you have to pay. Google's free. Yeah, I know what's, how it's happening is these ads, I get it, all right. But it's not going to replace it. So I find a whole, this dynamic there and it's like I really, you know, in switching to cloud I was like, holy shit, this is different. This is the most different thing. Not oh, doesn't know me and like how it speaks or whatever. I'm like, it has limits. [00:18:15] Right. And I think perhaps we'll say this question for another episode, but I think a fun question to think about is how much should it cost per month to use an LLM? [00:18:24] How much am I willing to pay per month to use an LLM? [00:18:29] Right. These things are out here being thought partners, CMOs, CFOs, designers, coders, strategists, therapist for $20 a month. Like, come on now. [00:18:38] Yeah, I'd be willing to pay $100 a month if that money wasn't going to the worst people in the partnerships with terrible organizations or whatever. If it was like going to paying electricity bills and going to helping the environment and fixing things and improving things. Like, yes, we should pay for what we use and be mindful of what we use. Yes. [00:18:57] All right, but the flip side of this and saying this is something that I expressed in 2022 with the video that I did on, on, um, I did as a YouTube video when chat GPT debuted and rolled out is that. Which is that I don't want this to be a tool that is only for folks who can afford it. [00:19:15] I do think this is the most remarkable technology of our lifetime. And I, I've said it before, it's like again, it has the ability to democratize so much. [00:19:27] And so yeah, I do get that like an ads tier would be one way to make it possible and allow this to stay free. Which I know is like the marketing spin that OpenAI is taking. When in reality they should be, we need more money. But they're like, we want to get it to everyone's hands. And I know that that's a lie, right? These companies aren't concerned with making this technology universally available. They'll say that it's great marketing, great optics. They're primarily concerned with profit and shareholder value, right? And AGI, artificial general intelligence, because they do believe that it's winner take all, which they care about from a profit and shareholder value perspective. It's all about money, right? They say it is said there is no ethical consumption under capitalism. And it's right. That is a true statement. [00:20:08] So the answer solution is from a tech perspective, if you want to keep these things around, is open source models, it's local models, it's smaller models. These things are really already exist and I am not knowledgeable enough yet to be able to speak to them as solutions. But I'm gonna throw it out there so that this episode isn't all doom and gloom. And yes, I will get on to learning the things and trying the things, I will. But also I am a human. [00:20:33] I only have so many hours in my day. But again the episode more thoughts and musing than actions and conclusions. But it is something that I think is definitely worth asking yourself, right? Are we addicted to AI? And what does that mean? What do we do? What can we do? [00:20:50] All right, final section. And Then we're done. How I used Claude this week. So if you're new here, each episode I share a quick example of how I use whatever LLM I used Jesus that week. This week I used Claude to code a sales page for a new offer that I'm beta testing. You can check out episode 20 for the rundown of that. When, um, I say run down of that, I mean how to use an LLM to code sales pages. I, uh, I don't know where you folks listening are at in your journey with things if you're coding anything like that, but I, I really enjoy doing that. [00:21:22] Really, really enjoy doing that. Um, um. So no, it's not making the copy for me. It's literally just making the structure and what it looks like. And I love this. Being able to move things around and change things and just being. It's like. It's like driving a stick shift. If you've ever driven a manual car, to me is the difference between driving manual, driving automatic. Right. And not being like, oh, I'm driving manual in traffic and uphill. No, it's like, um, I'm in control. I'm on fun roads and I'm in control of this car. And it feels really, really, really fun. So I will link that in the show notes if you wanna listen to that episode. It was episode 20, using ChatGPT for your sales pages. [00:21:58] Uh, and I will say that the process, the process, if you're in canadia that I did with using, uh, Claude was not as good as ChatGPT. [00:22:09] And I did nearly max out my session allotment, my session usage. But growing pains. Right. And, but also, who knows what ll I'm gonna stick with what I'm gonna do. I don't know. But I'm here to keep sharing the things with you. All right, that, my friends, is all for today. Hopefully you found this episode helpful. If you leaving a rating or review. I'm not sure if we get 25 yet, but as I didn't get to look today. Um, but you guys are great. And those of you have left reviews, uh, and you've left ratings. Thank you. You can only do it on Apple. I did have someone reach out and I wanted me to make a point. If you're listening on Spotify, I don't know how that works. So I can. I know for sure. This is what I think that I'm looking at is on as an Apple. But you can also reach out, let me know that you're enjoying it. And if you found something helpful that's it's. It is really cool to hear from you. It's a very unidirectional, you know, medium running a podcast. So, uh, feel free to text me 310-737-2345. It will be green, but it is me. It's my sideline. [00:23:14] All right, Last thing, don't forget, I do have a companion newsletter, the Curious Companion, that drops every Thursday. And it's basically the podcast episode in text format. So if you prefer to read or you just want a written record, join the lose newsletter fam. You can head to chatgpt curious.com forward/newsletter, or you can check out the link in the show notes. As always, my friends, endlessly, endlessly, One more time. Endlessly appreciative for every single one of you. Until we chat again next Thursday, stay curious. [00:23:49] It.