← AI For Humans: Weekly AI News, Tools & Trends
AI For Humans: Weekly AI News, Tools & Trends
Midjourney Built a Body Scanner. AI Science Just Got Real.
00:29:09
Loading summary
Transcript75 lines
- [00:00]
A
Is the AI changes science era. Finally here midjourney, the AI imaging company just revealed a new piece of hardware that actually benefits all human health.
- [00:11]
B
To be clear, I maintain I had the first anime babes to medical device pivot in history and it also made users very wet. So I understand how that sounds. Let's hear from the founders.
- [00:26]
A
We show off the first new whole body medical imaging modality in 50 years. We kind of dreamed of something as powerful as an MRI and as casual as a trip to the spa. We'll have details on why this full body scanner is such a big deal. And also maybe wonder who's going to clean out that liquid.
- [00:44]
B
Oh, you think they're just going to team to pour hydrogen peroxide on it
- [00:48]
A
to get the algae. Exactly. So the algae growing.
- [00:51]
B
Oh, hey, by the way, Claude, Fable 5, the model that changes everything. It might be back very soon maybe.
- [00:58]
A
And Snapchat just unveiled specs their new AR glasses. They're pretty cool, but they're kind of expensive and also kind of ugly.
- [01:06]
B
Okay, whatever, square. I see you, I see the weather. I got my slack notifications and I can monitor the algae in my anime babe tank. You see this?
- [01:16]
A
This is AI for humans, everybody.
- [01:18]
B
Guess what I'm. Guess what I'm doing right now. Gavin.
- [01:20]
A
No, I don't want to know what you're doing. I don't want to know what you're doing at all. You keep that to yourself. Ugh, gross. Welcome, everybody, to AI for Humans, your twice a week guide to the wonderful world of AI. I'm Gavin Purcell. That is Kevin Crater. Kevin, we have a weird story that I did not expect. Midjourney, the company that originally was kind of the leader in AI imaging, has fallen behind a little bit, but maybe we know why. Yesterday, Mid Journey founder David Holes, who is kind of a weird, interesting guy. Midjourney, never raised money, and they're doing pretty well for a company that never raised money. Unveiled a new hardware device, and it's a series of devices that are coming out from that. But this, Kevin, was surprising. It is a full body ultrasound scanner, and when you watch the video, it really does feel like kind of something from like almost a Minority Report kind of sci fi world. But I think we should go in deep because this is actually a pretty big deal. It is a pivot, but I think this is a big deal in the world of AI and might get us closer to that really interesting future we've been talking about.
- [02:24]
B
Yeah, a couple things. First of all, was I alone that when I saw the video of this, I thought, oh, they're just teasing like a new video model that.
- [02:33]
A
Yeah.
- [02:33]
B
And they're going to say like, oh, yeah, because it was very stylized. A lot of orange glow on someone getting lowered into like a vat of liquid. And I was like, oh, the, the physics, physics effects are actually pretty good on this model. Good job. Mid Journey. And it was like, Mid Journey Medical. But first of all, David has a background in all sorts of wild stuff and complex mathematics and his father was into medicine, I think on the dentistry side. But it's like not completely out of left field. It does feel like that for Mid Journey, but not for David himself. And what little that I've, I've read about him. But, you know, using the profit. Because they are not a publicly traded company with investors to worry about, they can take the Mid Journey war chest o cash and use it towards good. Like this. Yes. If you haven't seen the video or looked into how this thing works, they're kind of promising. Again, what remains to be seen, but they're promising a full body ultrasonic CT scan. So there's no magnets, there's no radiation. It's complete. Thank God.
- [03:31]
A
Because, you know, radiation is doing so much terrible things to us, Kevin. All this radiation I get hit with all day long. I can't stand it. I've been pilled by Radiation World. Anyway, keep going, keep going.
- [03:41]
B
Yeah, no, do you want to put on your aluminum hoodie for the rest of this podcast? Are you okay?
- [03:46]
A
All right, let's keep talking. I'm fine now. I could be.
- [03:48]
B
Okay, so what this means is like 60 seconds, according to their pitch. You get dipped into a vat of liquid and it has a bunch of rings around it that have a bunch of transducers on them, like hundreds of thousands of them. And it fires these ultrasonic waves through the water. And what it's doing is checking the readings back from them from this whole ring. Imagine like you're kind of Almost like a 3D print is layer by layer, it's scanning your body in the same way. And then eventually it's going to use AI because hey, imaging company, it'll use AI to read the images that come out of it and produce this 3D map of your internal tissue. Very early days, I've seen a lot of shade thrown at the sample images on their website against the business model about the fidelity of these scans and the usefulness of it. But if it pans out, I think Even, you know, 50% of what's being promised today it could be a game changer in health, early detection and for
- [04:42]
A
ease, using the cloud. Using the cloud. Word game changer. You didn't write that with Claude, did you? Claude loves game changer. Yeah.
- [04:49]
B
Buckle up, buckle up. This isn't medicine, as usual. This is a disruption of your. No, I don't. This is. I'm literally shooting from the hip and I hate that I am a generic LLM.
- [05:00]
A
I think the important thing to understand here is we have talked about, for the three years we have done this podcast that eventually AI would allow us to do kind of weird moonshot technology, right? And I think the important thing for everybody who's listening to this to understand is if you've been following us for a while or if you're brand new, welcome. But also, this is the kind of thing that we would hope that these companies that are amassing a fair amount of capital would actually do with their money. Right? Like, this is a real hard new physics.
- [05:27]
B
Let's go.
- [05:28]
A
Exactly, exactly. And this is like the idea where, again, midjourney has never taken venture capital, so they don't have to, like, go to their investors and say, like, hey, we want to do this. But I do think this is the idea that if. If companies like Anthropic and OpenAI weren't like, kind of racing to go public right now, I do think ultimately those companies may end up doing this kind of stuff as well, too. But the most important thing here is that, like, there are promises of a future where this kind of technology can help us. And I do think it's really important to kind of say, there's a couple caveats here. And to Kevin's point, he just, he mentioned, you know, there's people that are out there kind of shading out. Hank Green had a very long, quick take on this. And I'm a big fan of Hank Green. We like what Hank Green does in general. And he's been kind of like, over back and forth on AI in general and has very good takes, I believe. Like, if you're going to follow somebody who has kind of sometimes negative takes on AI, Hank is a good person. Hank's take on this was very, like, positive, but also giving a quick caveat that, like, the idea is that this isn't gonna, like, scan. This doesn't replace every scan you wanna do. Like, really? This is specifically designed to replace MRI scanning. Right. Which is a lot of, like, you know, what's that guy's. Who's the blueprint guy or the guy who's.
- [06:36]
B
Oh, Brian Johnson.
- [06:38]
A
Brian Johnson, who was in the crowd at this thing, like, Brian Johnson's gonna adore this. Cause it's gonna be really good to check your health. And the idea here is if you got scanned on a regular basis, you would be able to see tumors over time, very quickly. And instead of waiting until they've grown. Right. But Hank pointed out this idea that, like, look, there are other types of scans. This is not replacing, this is not a panacea for all scans, that it doesn't scan your brain or that there's other things that like different types of technology can do better. That said, he, a skeptic of AI is still pretty excited to see this sort of thing happen. So I do think it's an important thing to kind of like lean in and be like, yes, finally there's somebody taking what is essentially a pretty big risk to do something that is an audacious kind of goal. And that, that feels exciting to me.
- [07:25]
B
Will you share your body scans with the Internet a la Brian Johnson? Can we, can we check your density?
- [07:31]
A
Listen, my density right now at this particular time is off. I will say I'm in my, I'm in my peak moment of, of I go. If you've watched this show, you may see my face grow and shrink over time. My entire life has been like a fight with like 20 pounds on and off. So I wouldn't share them now. But Kevin, I am in, I am on the right direction again. So I'm on my way back down. So I think we love this. At some point I would. At some point I would. Yes, I do.
- [07:55]
B
Here to catch you when you fall.
- [07:57]
A
Thank you.
- [07:57]
B
Mostly because you'll be lighter. Yes, mention that. But let's mention that because I was going to get to this and I think you're going where I'm going, which is like we say, oh, this is a mid journey pivot. But this doesn't mean that they've announced they're done with image models or more traditional LLM stuff.
- [08:10]
A
Right? Yeah, that's right. There's. And by the way, they have like six other things are going to announce. I think that they really, they know the AI imaging part of their business is what makes them money. Like they've, you know, banked quite a bit of cash from that. And you know, one thing that's tricky with midjourney, as we've said before from an image model perspective is like you and I kind of both felt like the last one, I think it was seven or eight, I can't remember which number it was. Wasn't like a leap. But a lot of people really do love midjourney still because it. It delivers like, very artistic image generation. I think they're going to continue to move in that direction, but they've also. They have these kind of mysterious devices that we don't know what they're going to be. David mentioned this idea of, like, another one for your body that's coming and another one that's for other things. This is kind of weird imaging of, like, them showing off what's coming next.
- [08:57]
B
He. Well, he was. He was the founder of Leap Motion, if I recall, like, eons ago, and then did, like, hand tracking in VR. So if you look at like, and maybe we'll put it on the screen, there is a clearly, there's a shot of some lips. There's a shot of, like, a person looking down. There's a weird, like, community of maybe people or a portal that goes into something. But there's clearly a human hand on there. And so he's no stranger to doing stuff with, you know, hand tracking and VR. I wonder if that will have AR applications or whatever. Maybe it'll have to do with robotic hands. I don't know. I love that they're getting into it, though. And I think, to your point as well, it's like, I expected this sort of from an Alphabet or from a division within an open AI, but it's clear that they're all focused on, I don't know, crushing coding benchmarks so they
- [09:39]
A
can get to IPO making the money. I will say one last thing about this, which is really important. I mean, it's not. This is me projecting. One of the things in science fiction that people have talked about forever is this idea of slicing up the human body to scan it, to ultimately, Kevin, be able to upload your entire body into something else. And the first step of that was always, in science fiction, was always, okay, you die. And they slice your body into these tiny little slices along the way so that you can be recomposted somewhere else in the world or in the universe. This feels like a start of, like, doing that without having to die, which is a very interesting thing. If you can have interior imaging of your entire body, including your brain, eventually. And this is a baby step one. We may be talking about the idea of, like, you know, this idea of uploading yourself to something, whether it's to the Internet, whether it's to, you know, the lawnmower poor man's world, or it's whether it's a way to kind of travel around the universe. This could be a little bit of that baby step to that. So all my sci fi fans out there, you don't want it. You want it.
- [10:48]
B
Like, I need to back up my liver. Do you know how gross that is? Do you know how broken it is from my 20s? I don't need that.
- [10:56]
A
You've been living healthy for a while.
- [10:57]
B
I don't these lungs. I had a vape sponsorship for a minute. I don't need my popcorn lungs, you broken liver.
- [11:04]
A
Were you really sponsored by a vape company? Oh, my God, dude.
- [11:07]
B
I would blow thick clouds, bro. All right, enough about. Enough about backing up our bodies. Let's talk about Fable. They finally released some gameplay footage of it, and I do love the franchise, but I don't know that I need to worry about getting married and buying a tavern. Did you see the new footage, Gavin?
- [11:24]
A
Because
- [11:27]
B
she has, like, one aspect about you. Well, maybe you got to go to buy a house. It just seemed like the gameplay demo had the final battle at the end. Anyway, I'll probably preorder now. Let's talk about Fable from Anthropic supposedly coming back. Gavin.
- [11:40]
A
Yes, following up quickly on our story from last week, Fable is still not in your hands. But according to an Anthropic spokesperson in Seoul, Korea, doing an event, they have said that they're confident in the next couple of days, Fable will be back. There's been a continual conversation around, like, what the government did, who's talking to who with all this. But I think the most important thing to understand is there's no way that Fable won't be back eventually because, like, the progress of the world marches on. There's a lot of really weird rumors going around about, you know, GPT 5.6. And I guarantee you that, like, Anthropic is in a place right now where they don't want to be leapfrogged while they can't do stuff. So they are going to be signing some sort of deal. We have no idea what that's going to look like going forward, but people are feeling relatively confident that it will come back, which. Kevin, as following up our last conversation, I have kind of paused my AI projects that I've been doing only because it does feel a little bit. And I've seen a lot of people say this online that, look, Fable could do these things. I want to do so much faster than why would I fight with 4.8 to get these little bug fixes? It does feel like we're in that moment where you're like, well, I'll go outside, I guess I'll go touch grass.
- [12:49]
B
A lot of the engineering groups that I'm in, they're all shifting towards maintenance routine, pruning, focusing back like a backlog stuff, cleaning up some PRs here and there. None of them seem to be implementing like major foundational stuff or refactoring important components of their, of their builds because they're, they're all waiting in the wings. They're like, oh, they gave us. Then they taketh away. But the moment Fable comes back, then they'll tackle the more ambitious stuff. And that's really, I mean it's wild from less than a year ago when it was like these vibe coders, these aren't real coders, they're not shipping anything they can't even read or understand. And now they're like, well, I don't want to do any projects until the magic box is back in my hands. But by the way, in the meantime, GLM 5.2, not to get weedsy for the folks out there, but you know, another model out of China that punches above, at least according to certain benchmarks, above GPT 5.5 but just short of Fable. Well, if you can't get access to Fable, it's now the best game in town. So this pause, you know, I don't know the long term effects if any for anthropic or for US foundational development. But if you're OpenAI, yeah, you've got a slightly better relationship with the government, to put it mildly. But do you release your big foundational something or do you wait and just slowly iterate until you see where the Fable cards fall?
- [14:08]
A
Yeah, you know, I'm really curious about this. There is a little piece of news out of OpenAI that kind of is an interesting thing that Dean Ball, who previously worked in the Trump administration on the AI platform, he's not like a hardcore Trumper. He's a guy that like came from the AI space and kind of went to this Trump administration to kind of help shape. This is now going to work at OpenAI and that's a big deal. I think the other thing, Kevin, that's really important to talk about when we talk about role switching. The other thing that happened this week was Noam Shazir, who is one of the original authors of the attention is all you need paper from Google that was the original transformer paper that kind of set off all this AI revolution, has left Google again and is going to OpenAI as well. So I do feel like it's kind of like Pokemon collect them all. Who can get the biggest names in AI? Because we know Andrej Karpathy went to Anthropic not that long ago, and now you've got Noam Shazir at OpenAI as well as Dean Ball. Like, it does feel like people are starting to. Your point about this idea of, like, how do you get the best AI engineers that you mentioned last week about, like, how do you get those people to come work for you? OpenAI and Anthropic are in this kind of battle with each other for that exact thing.
- [15:13]
B
I love recruiting talent, Gavin, and I think there's no better talent than in our community, our viewers, our supporters, and. And this is basically a. I wouldn't say a cry for help. We're gonna say a call to action to each and every one of you sitting there going, man, these guys are just so amazing. I can't get enough of them in me. I wish I could dip myself into a vat of their liquid. Oh, no.
- [15:37]
A
Oh, no, they don't.
- [15:38]
B
No, please, God, don't do that.
- [15:40]
A
But also think about Starship Troopers. Just raise your hand and say, I'm doing my part. You put that as a comment in our YouTube. You like and subscribe. I'm doing my part. That's what you have to do to help this show grow and get better. We are not putting you in a vat of our liquid. Never do that. If you would like to surprise us and although now, hold on.
- [15:59]
B
New tier unlocked.
- [16:02]
A
We have to talk about Snapchat Specs. Snapchat specs are the new AR Glass device from Snapchat, which, you know, has been kind of struggling for a while. This feels like Specs has kind of been the direction that's really driving the company. I'm sure there are a lot of people out there who still use Snapchat. It's funny. I have two daughters who are both, you know, older teenagers. They're in college. They never use Snapchat, which is interesting. So I don't really know who's using it on a daily basis. These level.
- [16:26]
B
I was going to say anytime I'm chatting with older teenagers, it's. It's usually on Roblox.
- [16:31]
A
Yeah, that's also not okay. This is a show where Kevin has
- [16:34]
B
said too many things.
- [16:35]
A
He has to. Older.
- [16:40]
B
It's research. Snapchat or Snap Kids creating and shipping, like, a lot of interesting stuff that usually gets stolen by Instagram very quickly. We know that they. I mean, they. They've been playing in the wearable space for a long while. Yeah, I, we, we say this time and time again with hardware. Typically version three of something is where it actually gets good. And it doesn't matter if it's, you know, cell phones or rockets. Usually version 3 is where everything is humming along. This looks like a version one product. Even though they've been playing with it for a while, the price point has some people balking like, it's upper.
- [17:19]
A
It's more than 2,000, 2,100 bucks.
- [17:20]
B
Yeah, I mean, still cheaper than the Vision Pro and way more capable, at least on the surface, than like the meta Ray Bans. So it is kind of priced in between. I don't, I don't. It's not for me. It's for maybe a younger me with like, who cares? Disposable. Disposable sort of attitude. Yeah, but, but for me now, I'm like, I do like the idea of the navigational aspect on my eyes or getting real time translation without having to pull out my phone. But four hours of battery use when you're in the wild, and that's like the good case scenario is a little tough. But I don't. How are you on wearables in general? As, as a youth, I thought I would love it. And now that I'm like old man shaking fist at everything, I'm like, I don't want cameras on my face or on everybody else's face all the time.
- [18:07]
A
I think we're in that in between time right now. I will say one other thing quick, about the Snapchat specs. There's no Puck, which is pretty amazing. They put all the hardware into the device. It looks a little bit goofy, which people have made jokes about. Very funny picture from Ian Chan where it's like doubled up and made significantly larger. I think in general, I think we are in that in between space where eventually something like this will work. I think we are a ways away from it. There is definitely that wearables tax, which I will call almost like a friendship tax, where when you see somebody wearing it or somebody mentions the fact that they have a wearable, there's a lot of people are like, please, no, I don't want to be recorded. I don't know how we get over that per se. It might just be where everybody has them at one point and it's just not that big a deal anymore. I think that we are probably. What's funny about VR and AR and you have followed this world forever is that we're probably still like five to 10 years away from these things being a thing that people use. But I feel like we've said that three times over the course of our career in some ways. So I think the Apple Vision Pro failure was a big kind of like, oh, watch out. And this feels like another moment of like, well, Snap is a public company. Their stock went down when they premiered this. It's not trading very high. I just think you might have to have another like, kind of like Palmer Luckey type moment where somebody just comes out with something and it like blows everybody away. That isn't a public company. I don't know. We're in the middle of everything. I feel like with this.
- [19:29]
B
Yeah, bless them for trying and I'm sure some people will get them and they'll be very good for them. I got more excited, Gavin, by not even a game developer, but by the Unreal Engine.
- [19:40]
A
Yes.
- [19:40]
B
Updates that were announced, which, you know, some people, of course, big mad that whether you like it or not, AI tools and AI technologies making their way into the core engine that's going to power a lot of games. So now when developers say, like, well, I didn't use AI. I mean, I did build it in Unreal though, so maybe they used AI, but I didn't use AI. But let's maybe talk about it before I get too.
- [20:02]
A
Yeah, I mean, I think the big thing that I'm very excited about, and I can't wait until Fable comes back to try this, is that they unveiled an Unreal mcp. So that's a thing that we've talked about in the show many times. It is basically a backend that allows you to connect your AI systems, whether there's Claude or OpenAI to the Unreal Engine. And I am very excited by the demo they showed where basically they talk about game developers having a bunch of assets to kind of build with their assets and using AI tools to write the code and build those assets in the game. And it's very cool looking. The thing I'm kind of excited about as somebody who doesn't have a bunch of assets, is like looking at what I can use that's open source out there and trying to build what essentially is an Unreal Engine video game pretty quickly with AI, which I think is possible now. I do think the thing I always think about when we were talking about last show, about how the Unreal people were using Nanobanana and GPT Image 2 in their actual AI stuff. Do you think? I just cannot imagine this is the case. Do you think that the actual Engineers at Unreal. Now, the actual coders, do you think they're shunning using Claude code or using Codex? I would imagine they're all using it right now.
- [21:08]
B
100%. Yeah, I would be, I would be. I, I don't have insider knowledge. Let me be very clear about that. I barely have any knowledge. I would be shocked if they weren't all using these tools and sprinting and iterating with them. I just, I mean look, they, if they're showing off within Fortnite that they've got their own custom tooling which leverages image models and whatnot. I mean, yes, yeah, it's kind. There's a world where maybe there's one, one guy surrounded by 20 sided dye and mountain Dew in a closet somewhere that is handcrafting all of the code that uses someone else's AI. But I'm willing to bet they had a big internal huddle like most companies and said how do we unlock 10x productivity by using these tools?
- [21:50]
A
Well, and this is not to like get anybody in trouble there or try to incite some sort of like you know, flame war on the Internet. But just to be clear, like they're all using AI coding tools. And to also be clear, you may be mad about GPT image 2 and nano banana and all the stuff that you know, they trained on. Guess what? The coding tools come out of that same training database. Like there is no separation. So if you are using coding tools and you think they're great, you're in the same pot, you're in the same liquid vat that you all we all got dipped into. So you can't be hypocritical and say one thing and not the other. I just want everybody gamers. If you're going to be an organic programmer, I get that. If you're going to make a game that says no AI was ever involved. But I think we are just crossed the rubicon already. If you are a programmer engineer and you are not using AI tools, salute. Congratulations. But it seems like you are making a mistake at this point.
- [22:40]
B
And again, not like if you heard MCP and went well, I don't even know what that is. Model Context Protocol. As Gavin said, it allows your AI tools like your Claude code or your Codex or if you're Gemini, whatever. It allows the AI to interface software or another system. Right. And so what people are already doing is hooking these up to their AI agents like Hermes or openclaw and giving the agent a task. Go build a game that does X, Y and Z a top down shooter with a blah, blah, blah. And the agent will be relentless and it will connect to the Unreal Engine and build the world. It's not going to be AAA quality yet, but the fact that now the agent could directly talk to the engine. Exactly. And again, it's going to come down to taste and a whole bunch of other stuff. But quickly, Gavin, I think we should leave some folks with some real takeaway. And I know that they're like, oh great, where's this bit? But no, like some real new things that you can use and implement now because every week new stuff comes out and we should just quickly shout it out.
- [23:39]
A
Especially because I will say especially because the Unreal Engine is going to use a crapload of tokens, especially if you're connected to Fable, Fable comes back. So yeah, let's talk about Headroom.
- [23:48]
B
Yeah, so Headroom is released by a Netflix engineer shout out to Tony Simmons, who at least brought it to my attention through their, their X post. But what it does is it basically compresses your input and in some cases the output from an LLM so that it uses less tokens the more you type in. If you throw a giant log file at something, you're debugging something or whatever, it can have thousands of lines of repetitious stuff or even just removing spaces, like blank spaces. It takes all of that out. It compresses it all. You don't typically lose accuracy with it. And that's like one of the things that I still have to stress test myself. I just wanted to shout this out. I have not used it just yet, but the early reports are coming back great. It also keeps a real time log of the unadulterated prompt next to it. So let's say you're throwing a 10,000 line file at it. It's going to compress it, try to get you the response. But if the agent's like, hey, I really need to see explicitly what happened on line 74 and you cut that out because you found it repetitious or not related to the task, it still has the original file and it can go back and grab it. So there's some cool stuff there. But basically it's promising massive, massive savings, which is something that maybe the AI companies don't want you to have just yet, but you can plug any tool into it and yeah, you can. Really?
- [25:05]
A
You don't think they want you to have that? I think this is like a little bit of like. I think having savings on tokens would mean people would spend more on it. I would, I would suspect right now, because they're talking about how they're worried that budgets are being cut because they're not getting enough value out of the tokens that are being spent. I kind of feel like that's a world where they would want it. But the funny thing is, we don't know. People are speculating about what the anthropic or OpenAI margins are on their tokens, and this is very in the weeds. But whatever they make or don't make based on a token being delivered to a user is a really unusual question in the API. So if they are making a lot of money, they wouldn't. But if they're not, and they want to find a way to make people more efficient, they might. So that's all.
- [25:46]
B
Yeah, I mean, look, maybe. Maybe everybody's happy all around. I just gotta. I gotta feel like, what. What did I miss?
- [25:54]
A
I'm just laughing. Maybe everybody's happy all around. That's a great dismissal. It's like, sure, maybe everybody's great. Maybe everybody's utopian and happy.
- [26:02]
B
Maybe, maybe everybody's delighted that end users are saving this, this, all these tokens. I just have to imagine that, like, even when the headlines come out about all the unprecedented Uber spend on AI, they're like, well, they're probably still going to buy more tokens. They're probably going to buy more. And if they have to buy a couple more seats, oopsie doodle. I guess that's going to help with our ipo.
- [26:24]
A
But, Kevin, how do I convert. How do I convert tokens into a better me so that I can make me better? Because that's the most important thing is
- [26:31]
B
this, Gavin, you're benchmaxing your body now. Yes, you can. I guess benchmaxing also works for the bench press. You can turn Claude into your ultimate personal trainer. And this is a guide that Hawks Ox posted over on X. And basically, it is a series of prompts that you can use. You can set it up as a project within Claude, the desktop app. It'll also work in Codex. You just have to set it up slightly differently. But it's literally a series of prompts that you can copy and paste and tweak based off of your needs. And instead of having to pay for a fitness app, this will turn Claude into a very cool. Yeah. It'll suggest routines for you, it will track them for you, it will adjust them for you based on your feedback. And I just, like, did a glance at the prompts. They look really solid and it's so simple. You don't have to be a coder. You just have to be able to copy and paste and wanna put some stuff in there. And honestly, if you don't have the willpower to do that, you probably don't need the personal trainer app.
- [27:28]
A
By the way, this is really funny. This is an anecdote, but like I mentioned earlier that I'm like trying to like, cut calories back because I have a tendency to like, just, just I. If I don't watch what I eat, I'll just keep eating, which is a fun experience of my body. But I use the I anytime I do this because I. I'll do it relatively regularly just to get myself back on track. I use an app called Lose it, which has been around forever. Right. I never Lose it tries to always convert you into a paid user. I never used it. It was a very simple, straightforward kind of calorie, calorie tracking app. It had a lot of my foods in it, other people's foods in it. I have found this app now almost unusable because a. I think they're just trying to convert everybody. But there's something going on where there's just like an insurification of that kind of software I am really now seriously considering. And it's not like I couldn't have done this before, but like, that personal software thing is like, I guess I could probably just do this myself and it would just be easier to like, have my own version of this rather than trying to use this thing that exists. And I think that is the world where we're probably going to get at. We are now close enough where a personal trainer app is not like you have to go give somebody in the, you know, a vibe coder who created this thing on their own in the App Store. You got to pay them 10 bucks a month, which has been a business for a lot of people for the last two years now. You just make your own, right? You just make your own. You pull in this series of prompts and it can make your own. You're not paying any money. You're paying the money out to the AI companies. And like, this feels like where software is going in a lot of ways to me. I agree. All right, follow me on my fitness journey, on my Instagram, on my Tik Tok. I'll be talking all this stuff. And we'll see you next week, everybody.
- [29:00]
B
We'll see. Thank you for Everybody's happy all around. Goodbye.
- [29:04]
A
Everyone's happy all around. Goodbye. Everybody's happy all around.