ChatGPT Mass Exodus and Switching to Claude
There has been a lot of buzz lately about a possible ChatGPT mass exodus, with many people wondering if they should stop using it. Headlines about OpenAI, Anthropic, and government contracts have sparked conversations about switching to Claude and other AI tools.
In this episode of Run Your Private Practice with AI, I break down what is actually happening and share my perspective on whether switching tools is really necessary. I also explain a simple way to move what your AI knows about you so you can easily work across platforms like ChatGPT and Claude.
AI Tips You Don’t Want to Miss:
- What started the recent ChatGPT controversy, and why people are talking about leaving the platform.
- My honest take on using both ChatGPT and Claude instead of choosing just one.
- A simple method for switching to Claude by transferring your AI “memory”.
- Why creating a brand guide helps any AI understand your business and voice.
Mentioned in this episode:
– Thera AI Hub: https://kymtolson.kartra.com/page/AI-Tools-for-Therapists
– How to Train Claude if you want to switch: https://claude.com/import-memory
– AI prompt for nonpartisan information: I want a strictly factual, nonpartisan summary of this topic: [INSERT TOPIC].
Follow these rules carefully:
Start by briefly stating what is known as fact (verifiable events, dates, direct quotes, official actions).
Clearly separate facts from interpretation, speculation, or opinion. Put facts in one section and interpretations/arguments in another.
When there is disagreement, neutrally summarize the main arguments from each major side without taking a position.
Use neutral, non‑emotional language (no loaded or persuasive wording).
Rely primarily on:
Official documents and data (laws, court cases, government reports, company statements).
Peer‑reviewed research and reputable reference sources.
Factual reporting from established outlets.
Avoid: opinion columns, editorials, campaign sites, and social media posts except when quoting them as examples.
For every important factual claim, provide at least one citation or link so the information can be checked.
At the end, add a short “Uncertainties and open questions” section that lists what is not known or where sources conflict.
Format the answer with short sections and bullet points so it’s easy to skim.
Connect with me:
- Instagram: @thetravelingtherapist_kym
- Join the Clinical AI Club: www.clinicalaiclub.com
Sponsored by Alma: www.helloalma.com/kym
Sponsored by Berries:www.heyberries.com/therapists
Sponsored by Auralink Pro: Start your free trial and use code KYMTOLSON for 20% off www.auralink.com/pro
Transcript:
[00:00:00] Hey everybody. Welcome back to another episode of Run Your Private Practice with ai. So this has been a week. Every week is just, even more crazy than the week before in the world of ai. So right now there’s a big fight going on between Anthropic Open ai, the Pentagon, over whether AI can be used for mass domestic surveillance and fully autonomous weapons.
[00:00:24] Okay. That’s terrifying, right? I, I don’t really trust AI to send an email for me without me checking it first and for good reason, because sometimes it’s like putting weird things in there. Now I know that this AI would probably be more advanced than the type of AI that I am using, but still, it’s a little bit scary why this matters for us as clinicians using AI tools at our practices.
[00:00:47] Well, it depends on your feelings. I mean, there’s a lot of political stuff that’s coming up with this, you know, and I am not ever gonna get into that. I’m not gonna let politics or news cycle persuade me one way or the other. If something’s helping me and I find it useful, I’m probably gonna use it. ’cause I don’t trust any news source really.
[00:01:06] It’s all a bunch of just like propaganda in my opinion. I don’t wanna get into that too much. But anyway, people have been asking me this week, are you gonna switch from ChatGPT? And I’m gonna get into why. But no, the answer is no. I’m not gonna switch. I’m gonna keep using ChatGPT. I’m gonna keep using Claude.
[00:01:22] I’m gonna use the one that’s best for me right now because I don’t even know if half of this is true. Anyway, let me tell you the rest of what’s going on with it so you can decide for yourself or go research it if you want to.
[00:01:33] Now, I pulled all this information from a lot of different sources. I have this prompt that I use with my ais and I’m like, I want you to like go out find information about this, and I wanna make sure it’s not being influenced by the right or left. I want it to be like middle of the road, just like, what is the word? I’m trying to think of the prompt I use, but just like not partisan. Like I don’t, I don’t want information from either side. I just want the middle of the road facts so I can make my own decisions about this.
[00:02:01] when I use prompts like that, it does go and pull quotes and it’ll show me, well, this is from this side, this is from this side. Here’s the middle ground. So it does help me with that. I use AI with that a lot, but I’m trying to just like sift through all the noise of the news media.
[00:02:15] But anyway, let me tell you what’s happening. Okay. So originally Anthropic, which is also Claude, they were in talks basically with the Pentagon, the Department of War, to let Anthropic or Claude be part of the AI system for the Pentagon essentially. What happened down at the last minute before they were gonna sign anthropic, Claude refused the contract that required, this is in quotes, all lawful uses, including potentially mass domestic surveillance of Americans and fully autonomous lethal weapons.
[00:02:49] Now, I kind of laugh at this a little bit ’cause I feel like we’re already being surveilled. Like, I mean, our phones are surveilling us constantly, but that’s beside the point. Right after that happened, after the refused a contract, the administration came out and said, we’re blacklisting anthropic from government use and we’re gonna give anyone using it in a government setting six months to like wean off of it and that sort of thing, and get it outta your work systems and everything.
[00:03:11] So that’s a big deal, right? I mean, I don’t know how many people are using Claude in their government jobs, but maybe, maybe a lot are. So that happened, and then Anthropic basically said, here’s our red lines. We’re not gonna do any domestic mass surveillance of US citizens, and we’re not gonna put our AI in any fully autonomous weapons because current AI isn’t reliable enough.
[00:03:33] Which I totally agree with. I mean, like I said, I don’t even let it, send an email without me double checking it. And then what happened right after that, OpenAI signed a deal with the Pentagon, but added language in it, saying, and I’ve seen this in multiple places, I double checked it.
[00:03:47] They added language saying it’s tech cannot be used for mass domestic surveillance or for autonomous weapons. That’s what the contract said. Now the contract still says that the Pentagon can use open AI for all lawful purposes, which a lot of the legal experts out there are saying leaves a lot of gray areas.
[00:04:06] So even though they’re saying this, we don’t know if that actually means that they won’t use it for those purposes. So that’s where the big controversies come up. So inevitably what’s happened because of this, there’s been this mass exodus from ChatGPT. A lot of people leaving ChatGPT. I’m seeing it all over my feeds.
[00:04:25] I’m getting emails about it. People in the clinical AI club are asking me about it, like, what are you gonna do? What should we do? Oh my gosh, we can’t use open AI anymore or ChatGPT. And I’m like, I, like I said in the beginning, I’m gonna continue to use it. I don’t even know all the details of this. So I’m not gonna just jump ship with something that I’ve spent years building out and is like integral into my business systems.
[00:04:46] I’m not gonna do that. However, there are lots of people that are very like decided on this already and they’re like, I’m not gonna use ChatGPT ever again. I’m like, okay, fine. So what’s come out of that is people asking me how do I switch to Claude? And I did have an episode of this podcast a while back, I don’t even know what episode number it was, but kind of walked you through a rudimentary way to do that.
[00:05:09] Things have changed a lot since then. So what happened in response to this? And I’ll put the link down below in the show notes so you guys can go and do this if you want to. And I recommend doing it anyway just because Claude’s amazing. Claude has got so many cool things going on right now. I’m using it a lot more than ChatGPT just because it’s better for my workflow right now.
[00:05:31] But with ChatGPT, I’m still using my custom GPTs. I’m still using, some of the GPTs I’ve gotten from other people over there. I, I use it every day still. So I’m not, not using it, but I am using Claude a lot more because they’ve got the Claude Chat, they’ve got Claude Cowork, which lives on my desktop now, and I’ve been using it almost every day to help me with things.
[00:05:51] It helped me create viral LinkedIn posts, which I’ll probably talk about in another episode sometime how I did that, ’cause it’s pretty cool. The results that I got from it. And it’s got cloud code, which I haven’t even gotten into yet. So the link down below that I’ll give you under this episode, if you go to it, you open it up and basically Claude has given you a prompt.
[00:06:13] It tells you to copy the prompt, take it to ChatGPT and pop it into ChatGPT. And basically what it does is it summons ChatGPT to pull all of its memories about you. Put it in this really nice, what they call markdown file for you, so you can copy and paste that or download if you want to, into a .md file, download it, and then you could take it back over to the same window where Claude gave you that prompt and you just upload it right there.
[00:06:39] Claude takes it, puts it into its system, and now Claude knows everything about you that ChatGPT knew about you supposedly. Now I did it and I was kind of surprised, Chad Chip t’s memories. First of all, they were sort of old. They were like, not really as relevant. I’ve had a lot of new products come out since then.
[00:06:57] I have a whole new membership, the Thera AI hub, and I was really surprised that it didn’t know that, but, but a lot of it, it did get right. So what I did before, of course, I hit send to Claude. I went through and edited those memories. I took out some things that you know weren’t relevant, that sort of. And, and hit the button and uploaded it, and now Claude’s like, wow.
[00:07:16] It, it’s like eyes are open. Like it just bet me, you know, which is kind of funny ’cause I’ve been using it already for at least a month and a half pretty intensely. So I would think it would know more about me by now, but it didn’t. So you can do that too. And then if it’s not Claude, you wanna use, if it’s Gemini, if it’s Manus, if it’s grow, if you wanna try one of those instead.
[00:07:36] Just take the markdown file, take it over to the ai and in the chat box say, I wanna give you some new information about me. And almost all these programs in the back, there’s a setting where you can just go to settings and, you know, train it. You’re basically training it, it’s all called something different in each one, but you go in the back end and it’s like, what do you want to tell me about yourself?
[00:07:55] Or what do you want me to know about you? You could upload this stuff directly and put it in there. The other thing you can do if, if you don’t wanna do it that way. Is make a brand guide. We do this in the Thera AI Hub. I’ve got a whole process outline for everybody how to, how to build out your brand guide, how to download it, how to keep it as a PDF, so you could upload it to any AI that you’re using.
[00:08:16] You just take the brand guide with you, you update, update it as things change in your life, that kind of thing. But you could do that too. So create your own brand guide. You can even say to chat GPT or whatever AI you use. Like, I want you to create a brand guide. Write down everything that you know about me.
[00:08:31] Double check it. You know, make it up to date, that sort of thing. Add things in like your website, your brand colors that you like to use. If you know, you have a certain topic or niche that you specialize in, make sure it knows all of that stuff. And that’s essentially what a brand guide is. You just want to know everything about your business or you, depending on what you’re trying to do.
[00:08:50] If you’re a, the, like me and you’ve got lots of other income streams. Mine is very long. I think it’s like 40 pages or something. Super complicated. It’s got links to everything I offer. It’s got sample writings, it’s got, you know, websites. It’s got everything that I offer. It’s got links to all my blog posts that I’ve written, that sort of thing.
[00:09:10] So it really knows me when I upload that brand guide. But that’s a way to do it too. So if you’re concerned about this open AI thing with the government and you wanna switch, these are the ways that you could switch right now and do that. Something else I’m exploring, I don’t know if you guys know Jonathan Mast, J-O-N-A-T-H-A-N, Jonathan Mast White Beard Strategies.
[00:09:32] He’s one of what I would consider like my AI mentors. I take a lot of courses from him. I, I’m obsessed with him. He’s right where I am with ai, so every time I’m trying to learn something new, he’s on it. So he’s a really good training source in the, in the clinical AI club. I try to keep you guys up to date too, so you can definitely get a lot of this downloaded information from my brain and our two monthly mastermind sessions that I do over there.
[00:09:59] But anyway, Jonathan’s fantastic. And one thing I just took a class on is the skills.md files. MD just means markdown, right? Markdown is a type, it’s just like a text file or a doc or A-P-P-T-X, that sort of thing. So I’m learning the skills we’re going, in fact, in the clinical AI club, I think I’m gonna teach this method that he taught us, and I’ve seen other people doing it.
[00:10:23] I like the way he, he puts this stuff together and I’ve made a few myself, but these, these, these skills are like, basically, they’re essentially like folders. On your computer, in their MD format, markdown format. And the reason you do it in markdown is because almost all the ais, they recognize markdown as like a universal language for the ai.
[00:10:45] So that’s a good file format to use that you know, you can just transfer across to all the platforms. So you put in the folder, not just a file in the folder, a bunch of different pieces of content to teach it essentially. A skill, which is really what a custom GPT is, or a gem or a clawed project, you know, they’re called something different or a skill in Manus, they’re called something different in all the other a different ais.
[00:11:11] So this is just a you know, sort of a concise way to do it in a markdown format where you’ve got this folder, you could just take it and upload it to the ais and it knows the skill. Now some of the AI is like Claude, right? If you upload one of these skills folders to Claude it, integrate it, integrates it into its system.
[00:11:29] So it could just, it’s like that smart, like it’ll just go remember, oh yeah, you uploaded that skill that you created about, you know, how to make YouTube thumbnails. Let me just pull that and I’m gonna follow that format and we’ll do that instead. So anyway, I. Working on that. And if, if you’re in the Thera AI hub and you’re listening to this and you’re like, Kim Thera AI hub has a ton of GPTs, I don’t wanna use GPT anymore because of the, because of this thing that OpenAI is gonna work with the government.
[00:11:57] Don’t worry about it. If you’re in there, you’ve also noticed that I’ve got gems in there, and eventually I’m gonna have projects in there and I’m thinking, I’m gonna make these MD files, these skills that you could just have done for you. Take ’em, upload ’em to any AI that you want to, so. That’s where we’re right now with all of this.
[00:12:15] It’s kind of crazy and wild. Let me know if you have questions. Of course. Like I said, I’ll put the links down below so you can you know, follow up on this information on yourself. Come up if, if you’re in the world of politics at all, come up with your own little prompts. Maybe I’ll pull one of my prompts and put it down for you in the show notes.
[00:12:32] On just how I just get these like reports that are not left or right, they’re just like literally factual so I can make my own decisions and not be swayed by. Propaganda from either side really. I’m just a middle of the road kind of person that helps me a lot, like get through the noise. And you could take it to different ais and see the different ways that they address it, but a lot of ’em will stay in middle of the road if you ask ’em to with just like facts and quotes and like just a rundown of what’s actually happening based on pulling a bunch of sources from different places.
[00:13:03] I hope that helps you and let me know if you have any questions, and I’ll talk to you guys later. Bye