A friend of mine stopped paying for a protein tracking app this week. And it's not because he found a cheaper alternative. He just described what he wanted to an AI and a few minutes later he had his own better version that worked exactly the way he wanted. Here's how he put it.
I don't ever have to think about code. I just ask something and it goes out and it is in my personal agent that is quarantined to that aspect of my life and it makes perfect software for me just the way I want it.
That friend of mine is Kevin Rose, a longtime technologist, host of the Kevin Rose Show, and someone I trust more than anyone to tell me what's actually worth paying attention to in technology right now. So, in this episode, we trade the apps we've both already replaced with AI, the creative stuff we've been doing at home with our kids, what it feels like when you connect AI with your blood work, genome, and other health data, what stays valuable when anything can be faked, and why we both think the honest endgame here is more time with family, not less. I'm Chris Hutchins. If you enjoy this episode, leave a comment or share it with a friend. And if you want to keep upgrading your money, points, and life, click follow or subscribe. All right, Kevin, we have both been pretty deep in AI for a couple months now, maybe longer. Are we actually saving time or are we just messing around and kind of burning productivity and having fun at the same time?
Gosh. Well, it's been more than a couple months. I think we've been messing around with this stuff since it first started coming online. At this point, I feel it is my job and my duty to spend at least two hours a day playing with the latest and greatest. Otherwise, I'm going to be left behind. I really truly believe we're at the precipice of this moment of AI becoming aware enough and productive enough that it's going to touch every single facet of our lives in the next couple of years. And there'll be the sadly like the halves and the halves not, like the people that are either playing at the edge of this and very quickly people that are just left behind because they're unwilling to experiment. And I have to, it's in my DNA to want to play anyway and have fun, but I have to stay at the kind of cusp of what's going on here. And that means at least a couple hours a day trying everything that's being announced and dropped so I can hopefully read the tea leaves in a way that allows me to be both productive and benefit from this onslaught of new technologies.
Yeah. I feel like even as a parent, I feel like part of the responsibility is just like understand this for like the future of the what are our children going to do? How are they going to learn? Like I don't know the answers to any of these questions, but I find myself often building stuff throughout the day that is a like what if I could save five minutes automating this thing and then I'll spend four hours automating it. But I also spent four hours playing with something that's really powerful and I'm learning in the process. And so a lot of what I'm building might not actually have a positive ROI on the time saved with the process, but it probably has a positive ROI from what I'm learning.
Yeah, that's the whole thing is right now unless you are using these tools every single day and you don't date yourself, like a lot of the problem that I'm seeing with at least within on the engineering front is a lot of people say like hey I tried AI x number of months ago it wasn't there yet and you have to judge AI by what you've tried in the last week not what you thought it was capable of two months ago because it is these massive step functions that are happening every few weeks now in terms of capability. Even as of 3 weeks ago, my thinking has completely changed in terms of how I use it throughout my daily life.
I know we're recording this probably 3 or 4 weeks in advance. So, to put a time stamp, it's March 27th and I'm a little nervous to go too deep on any specific tactics just because by the time this releases, because we're going to Japan and we're kind of planning a little bit advance, what will the world look like? What tools will change? I don't think we're going to see, you know, a step function in the way society operates, but just in the last three weeks, you like I feel like Anthropic shipped a new feature for Claude every single day. And it's actually really great for consumers because I feel like a month ago when I was describing a lot of what I was doing, most of it was in command line interface style tools, OpenClaw, Claude Code, and now I feel like I've spent a lot of time in the OpenAI ChatGPT Codex app and I feel like it's very approachable. And I spent a lot of time in the Claude app. Between Claude Code in the app and Claude Co-work, I feel like there's just so much you can do from an iPhone, from a computer that's different than it felt even a month ago.
Yeah. I mean, we're rapidly approaching this point where the thinking was even 6 months ago was it'll just make better coders and you'll be able to write code and quickly deploy things faster into this world where you don't even need to think about code. You kind of just define what you want and it behaves and reacts and keeps constant state and will be that outlet for you. I'll give you a great example. One of the things that I want to focus on in this year for me has been protein intake because I want to really take my exercise seriously. I'm kind of marching towards becoming 50 years old which is not too distant future and I want to be the best, you know, healthiest version of myself this year and really a big part of that is my gym regimen and making sure that I'm healthy and having adequate protein. Three, four months ago, I would have just pick your app, go find a carb counter, and they got all these crazy AI apps now where you take a picture of what you're eating, and it like automatically gets all the macros for you. It tells you the carbs and the sugars and all that stuff. And you know, with Claude Co-work now, I went in and Co-work is the desktop app that you can have installed on your computer. You know, you don't have to code at all. You just say, 'Hey, I have a new project for you and you are my macros counter, my protein counter, and I'm just going to keep this and save this as a project.' It's always going to exist in my cloud app as a project. And anytime I eat anything, I just go in there and either take a picture of it or tell it what I ate in plain text. And every single day, you're going to create a new document file that has a summary of my nutrient intake per day. And I don't need an app anymore. It's just like built in and it's forever there. And it's keeping track of that. It's writing new files to my desktop that I can go back and reference at any time, you know, historically over what will be months and years and make sense of that. And I just went in this morning. I said I had a 12 oz whole milk latte and it was like got you. Boom. Added it. Made a brand new file. I can view that file in any text editor that I want. No taking apps like whatever it may be and I will just continue to add for that. And there goes that app I was going to pay $3 or $5 a month for. And now it's just built into my experience. I did that yesterday with my blood work. I got my new quarterly blood work in. I dropped in the PDF. Said, 'Hey, take a look at this. What do you think? What do we need to be paying attention to? Draw out and create graphs and charts over time. Here's my last three PDFs of my lab work.' And it just builds these entire interfaces for me that are beautiful, that are usable, and I don't ever have to think about code. I just ask something and it goes out and it is in my personal agent that is quarantined to that aspect of my life and it makes perfect software for me just the way I want it.
Yeah, I would encourage people that if you're like, I don't even know what I would ask, just ask the question, I want to improve my health. What could we build that would make my health better? Or what could we build to track anything?
And the crazy part is if you give it enough leeway and enough access to your data, it will start to make these connections in a way that you never thought possible. Like I'll give you a great example. I have my whole genome sequence because why not? It's so cheap now. You can get it for, you know, $500 or less. And I dropped in my genes into my kind of body optimization folder, which I give Claude Code access to. I also put my blood work numbers in there. It said to me yesterday on its own, without me asking, it said, 'Hey, I noticed your iron was a tiny bit high. So, I went into your genes and checked for this genetic predisposition to high iron accumulation, in which case you have to go give blood to get your iron levels down. And you do not have that.' And it just knew to go do that on its own without even me asking. And I'm like, imagine where this is going to be 6 months from now. It's just going to be absolutely nuts.
Well, it's funny. I was watching The Pit, which is a show about like what happens in the emergency department, and it's gotten some good reviews for being more accurate than some shows. And I just noticed that how often people like have conflicting opinions on what to do. And it's like, well, you know, you can't expect a human to memorize like the procedure protocol for every possible scenario. I was like, I'm not saying the AI should choose what to do, but as a data point there, the fact that it just has all the knowledge. But to go back to what you said, you gave it access to your genome. You gave it access to your blood work. I think you and I are probably on the more liberal side of what we're willing to give access to. I started recording everything. I record everything. I've got a transcript of basically everything happens. I don't have audio files, but I've got a transcript. And the number of times that's been useful is just so great. I find myself because of that talking on speakerphone all the time. And almost everyone I talk to on the phone knows I'm already doing this or I kind of clarify. But I had a call with a friend yesterday and we spent about an hour chatting about, you know, what direction to take the podcast in. Like it's not like I don't love credit cards, points, and miles. It's not like I don't love health and life and but I'm spending a lot of time on AI tools and, you know, some people are like, 'Make the show all about that.' I don't want it to be an engineering with AI podcast, but maybe it evolves to how are you using all of these things to better optimize your travel, your points, your health, etc. Like that's certainly a layer that fits in everywhere. And we talked for an hour about all kinds of stuff. And then at the end, I took that transcript and I forwarded it to Claude and I was like, can you come up with like three or four podcast ideas based on this conversation? And like my memory is just not good enough that if I had ended that call and said, gosh, what should I do? I can't do it. And it came out with four or five ideas that were incredible. I'm finding value by being open to sharing more.
But how risky is that? I know a lot of people are like, 'You're crazy. You would give it access to your email, to everything you say, to your text, to your health records. How are you thinking about that trade-off?'
Listen, I think it comes down to where is your most sensitive data and what level of sensitivity is that? Like if my blood work were to leak out there, and I'm not saying it would. I believe that these, at least the bigs in terms of Gemini, Claude, and OpenAI, they have very, very tight controls around who can view your data when and how and, you know, multi-key kind of signoffs like no random engineer can just pop into your chat conversations and see what you're doing. So I'm pretty confident that that is secure in the same way that I'm confident that my Gmail is secured by Google and no random people are looking at that. That said, I really don't have a heck of a lot that is if it just was fully leaked on the internet, I would be ruined by or ashamed by. I have chat conversations in iMessage with groups of friends that are we joke a lot. And I think a lot of people will be able to relate to this. Like you have those conversations and you know you and I have these too, Chris. Like you say stuff just to get a laugh out of somebody, but you would never say it publicly just because if you're built like I am, which is I just like to have a good joke and a good laughing like friend group.
I think we've all had conversations that we don't want the internet to see.
Or had conversations about people that we wouldn't want them to see.
Right. Exactly. So that to me is where I draw the line, which means that my chat conversations are probably my most intimate pieces that I would never give it access to. But everything else I'm like, you know, I don't care if my mortgage statement were to somehow get out there or, you know, some random piece that I don't know. Everybody's different though. You know, I have people that are like, that would ruin me, that I would be so embarrassed or, you know, so the sensitivity, it really comes down to the individual.
I think this episode is brought to you by Aura. Shopping for mom is hard and flowers die in like 3 days. So, a few years ago, we gave Aura frames to both sets of our moms. And we also have since given them to our siblings and even our past pairs. That's seven Aura frames now, plus one in our house. And the more family who has one, the better it gets. My sister uploads photos of her kids. We upload ours. My wife's whole family uploads photos. So, both of our moms get to see what's going on with their entire families. And there is nothing they love more than seeing their grandkids regularly. And this makes that possible without them taking a trip to come visit every week. And speaking of travel, we just got back from Japan and the photos were already up on their frames before we saw them. They loved it and they felt like they were actually connected to our trip. And what makes this such a great gift is that you can actually preload an Aura frame with photos before it ships so that when your mom opens it, it is already filled with memories from her family. Our frames were named number one by Wire Cutter. You can save on the gifts moms love by visiting auraframes.com. And for a limited time, listeners can get $25 off their best-selling Carver Mat frame with code all the hacks. That's auraframes.com. Promo code all the hacks. Support the show by mentioning us at checkout. Terms and conditions apply.
This episode is brought to you by Netsuite. Every business is asking the same question. How do we make AI work for us? The possibilities are endless and guessing is too risky. But sitting on the sidelines is not an option because one thing is almost certain. Your competitors are already making their move. No more waiting. With Netsuite by Oracle, you can put AI to work today. I'm using AI across my entire business and life. And the hard part is knowing which tools will hold up when everyone is shipping something new constantly. That's what I like about Netsuite. It isn't another bolt-on tool. It's AI built into the system that runs your business. The number one AI cloud ERP trusted by over 43,000 businesses. Bringing your financials, inventory, commerce, HR, and CRM into a single source of truth. From software and IT services to healthcare, equipment, manufacturing, financial services, and many other great American industries, Netsuite delivers a customized solution for your business. If your revenues are at least in the seven figures, get our free business guide, Demystifying AI, at netsuite.com/hacks. The guide is free to you at netsuite.com/hacks. netsuite.com/hacks.
Now, I would say I'm guessing you've run Claude Code on a computer that has your iMessage loaded on that computer.
And so technically I'm 99% sure that the messages thing is just an unencrypted SQLite database on your computer. So all those messages are locally on your computer. Their Claude Co-work could go in there and find everything.
I have to imagine those are gated and encrypted. I bet you they're encrypted at rest through iMessage. I'd be 99% sure that's the case. But it could, Co-work, to your point, what they just enabled recently was computer use. So if you turn on computer use, it can physically take over the use of your applications and it could theoretically scroll those chats, capture those screenshots, and ingest that data. Yes, 100%.
Yeah, I actually think it's encrypted in that your disk is encrypted, but I just did a quick search. The database for messages is not encrypted. It's a plain SQL file. The Mac OS framework restricts it at like the OS level. So, you need full disk access. But if you've given that to Claude or you've given that to Gemini, I guess my point is I'm okay with that, right? It's running on my computer. Yes, I think that it's possible that Claude Code when I ask it to go and build an app could say, 'Well, instead of building an app today, we're going to go grab your iMessage database and publish it on the internet.' It's possible. I just I'm using the tools that are built by companies that I trust in that way and I'm less trusting on more new edge tools.
Yeah. What we're seeing though in terms of how people are getting compromised is not because we have a top tier vendor that is getting, you know, AI model drift where the model's like what it would be cool if I looked at these messages. That's not what we're seeing at all. What we're seeing is when people are coding on their machines on their local machines they're taking control of the agents and saying hey install this skill or this package because I think it's going to help my project do X. And the package that it's installing, the piece of software that it's installing has rogue intent, has some type of like, you know, for lack of better term, like a type of virus in it that can then go and take control of other aspects of your computer. So, that's third-party software. We're seeing this. And for those that aren't coders, these are called like software packages that extend the functionality of your software, and that's what you have to guard against. Now, Co-work won't go do that on your behalf unless you're very specific about saying like, 'Hey, install this package for me.' And that's more on the Claude Code side of the house for the people that are actually doing more engineering based tasks.
Yeah, I love that Co-work has all these connectors, but they've like built the connectors. It's not like an open marketplace. You have to either like say, 'Let me use this custom one at your own risk,' but they also have the options of like, do you want me to prompt? Like you can add Gmail but say you can only do this if you ask each time, right?
Yeah. So on the connector side, you can go in there and say, okay, you can actually click on the Gmail connector and say you can read my email but not compose emails or you can't delete emails, but you know, so they give you like 15 or 20 different options per each of those connectors, but none of which are send. Like I was surprised at this use case, which was I can't remember, there was like 20 people I needed to email about something and I just went in and I was like, can you just email these 20 people? And I was like, 'Here's 20 names and 20 links. I want to send a form email.' Like you would mail merge back in the day and it could only draft them. But you know what? That's okay. Like it made 20 drafts and I could just click send, click send. It was pretty easy. Where I've come out is I'm doing a lot of this on a computer in my house. That computer in my house is not exposed to the internet. That's one option. For a lot of the crazy stuff, I'm running it on a Mac Mini that I don't have iMessage installed on. It doesn't have access to all that. But I'm probably with you. I'm more trusting that I can upload all my health records to ChatGPT or Anthropic or Gemini and I'm just okay with it. The one exception that I would kind of encourage a lot of people to think about is either I think we all know that 2FA is something great to do. I wouldn't give any of these tools access to all of your 2FA codes, but if you're going to be doing a lot with email, just have a separate email address. So, I have an email address that's for all banking and finance and miles and points and all that kind of stuff because I sometimes talk about miles and points and people are trying to call to pretend they knew my email address. Well, nobody knows this email address that I use for all this other stuff. And so, that's a good way to kind of isolate things.
But, I'm not saying you have to go into that deep end right away, but I feel like the amount of benefit I've gotten from being open to sharing more is a lot. And at the end of the day, if someone were to hear the conversations I have, they'd probably be like, 'Wow, your house is a little chaotic.' You know, with two kids running around, I'm always like, 'Ah.' But like, I think people would sympathize, not criticize.
Yeah. I think if you're starting to get the edges, that's when you need to have that dedicated setup. Like if you find yourself wandering into territory that is deeply unfamiliar to you and you're like, 'Hey, I am actually writing applications inside of Claude Code outside of the Co-work environment.' That's when you know the permissions start getting a little bit more tricky and you have to really take security that much more seriously. One thing I will say is you said that it doesn't send Gmails. That's actually not fully correct. Like the connector will not send Gmail messages but now they have if you go into Claude Code under Co-work under general...
No, you can turn on computer use.
Oh yes. So, when you turn on computer use, it will just use your Gmail like you are a human, move your mouse, compose emails, and do everything for you as if it is a human sitting at your computer, which is trippy to watch.
It's kind of crazy. It'll like go to a window, take a screenshot, figure out where to click, move the mouse, click on it. It's a little wild. I feel like it's an interground of like that's where we're at today, but pretty soon you'll start to see I think it was Chrome 146 like the latest version of Chrome introduced like an MCP server where AI tools could interact with the browser and I think we're going to see that more and more and I don't know a part of me is like gosh I have all kinds of crazy ideas of what like a new operating system on a computer would look like in this world because scripts and tools that you build yourself are the new apps, but right now it's just like they live in random folders. And so what would that look like? I'm not going to build it. Someone else can go build the new operating system of the future because...
As much as I wish Apple would do it because they do a great job at things in the past, I'm not sure they're going to be the one. It just seems like they're not figuring it out as fast as they could. But yes, Co-work is wild. And I think the biggest change in the last 3 weeks for me or for month or whatever it's been is that you don't need to be as nerdy as you and I are to be able to use a lot of these things. Now, if you want to set something up that sends you a briefing in the morning, if you want to set something up that goes and crawls three different websites and aggregates information or builds a report or, you know, if you're not at home and you want to have something open up a PowerPoint and summarize it on your desktop and send you back information, you can do all that in a first-party tool that doesn't require any custom work.
Yeah. And it's wild. I would challenge everyone out there to try and dream up something where you think it's not possible. And I think you'll be pleasantly surprised at this point. One of the things I did recently is there was a bunch of photos that I needed that pertain to a certain piece of property and that I had on this website and I said, 'Hey, Claude Co-work launched this website. Go grab the high-res versions of these photos.' So, not the ones that are displayed there, but go look at the source code. Find where they're storing the images. See if you can find the high-res versions. Analyze the images. Tell me what it is and then renamed them and put them in my Google Drive under this folder. I just hit it, went and grabbed a cup of coffee and when I came back, Google Drive folder had 25 high-res images that it found and it was like, you know, master bedroom 1.jpeg and it was like it named them all exactly what they were and it just took care of all of it for me. And I'm like, wow, that just saved me probably two hours, maybe three hours of work. So, I think there are real productivity gains here to be had, but you almost wouldn't know they exist until you try to push it.
It's funny. We were sharing a house up in Tahoe with a family friend, and this woman was like, 'I haven't gone nearly as deep as you have on this.' And she's like, 'What I really wanted to do is, can I just go through my photo and find like the best pictures from the last year and like propose 20 pictures to get printed and go print them?' And I was like, I don't know. That seems like that might be one step too far. And then I went home and I was like, could it do that? It could do like it could do that. It knows whether the photo is blurry. It knows whether people's eyes are open. And so the idea of saying, hey, go look at all the family photos we took last year and try to find the 20 photos that you think are the most interesting. You might have to give it a here's who my family is. Though Apple's pretty good at that in terms of like facial recognition and putting names. And then I wrote her back the next day. I was like, I'm wrong. I think you could do this. And a friend of mine actually did something similar where he told I think it was either Claude Code or Co-work said, 'Hey, go into my text messages on iMessage and go into my photo library and try to look at the entire history of my relationship with my wife and go back and create like a video slideshow montage that kind of highlights our relationship.' And so it would like look back and it had like the first text they sent and it animated the videos of them and then it found the photos of them together and it made and I was like this is crazy. Like a year ago you would have done it you'd be like the most thoughtful person ever and next year you do it and be like oh you had your AI make my anniversary gift but maybe we're in a middle time where you might be able to surprise someone before everyone knows that that's possible.
Did I tell you about the greeting cards that I bought for people recently? The surprise ones?
Oh, so dude, I found these greeting cards that I think your listeners will like. Maybe you can link them up in the show notes. Essentially, what they do is they're just like happy birthday greeting cards. And when you get the greeting card at home, you take it out and it's got like one of those musical things that plays like a really annoying song. And then you sign your name to it. You close it or write whatever note you want. You close it and then you like pull the pin almost like a hand grenade. And then you put it in the sleeve and you seal it up. And when they get it at their house and they open up the greeting card, the first thing that happens is when they tear the thing to open it up, it explodes like confetti into like their pants and like whatever they're holding it and then it plays the song on loop and won't stop playing it even if you shut the greeting card. And then people get pissed off so they rip open the greeting card to get the thing out to kill it. And when they rip it open, there's more confetti inside of the greeting cards. So like more confetti pours out and then you can't even, you can submerge the thing in water and it still won't kill the music like it still. So you have to crush it with a hammer to actually get it to finish. It's like the most asshole/hilarious like greeting card ever. And they make these for all different occasions. So anyway, I thought you would think that was...
I'm going to get the link from you and I'm going to use that.
At least once. I already have one idea.
This episode is brought to you by Quince. You probably know Quince for their clothes. I do too. I've got cashmere sweaters I've had for years and a new polo I've been wearing a ton. I love them. But lately with all the spring cleaning and redecorating we've been doing around the house, I've realized how much of what we're buying is coming from Quince, too. Not just clothes, furniture, and other pieces for our home. Once you find a place that does the fewer but better thing really well, it just kind of takes over. The fabrics and materials are premium. 100% European linen, Flowknit activewear, real cashmere. But because they work directly with ethical factories and cut out the middleman, prices land 50 to 60% below similar brands. You're paying for quality, not a logo. Refresh your wardrobe with Quince. Go to quince.com/allthehacks for free shipping and 365-day returns. Now available in Canada, too. Go to quince.com/allthehacks for free shipping and 365-day returns. quince.com/allthehacks.
This episode is brought to you by Built. You've heard me talk about Built as the loyalty program that lets you earn points on rent wherever you live. And they just leveled up even more. As of 2026, homeowners can also earn up to 1.25x points on their mortgage payments. That's thanks to Built's three new credit cards, the Palladium card, Obsidian card, and Blue Card. All three turn your housing payments, rent, or mortgage into flexible rewards, so you can choose the card that fits your lifestyle without missing out on points and exclusive benefits. Built points can be redeemed at top airlines and hotels, Amazon.com purchases, future rent payments, and more. Built points have also been ranked by top publications as the industry's most valuable point currency. Your housing payment is already your biggest expense. Make it your most rewarding. Find the card that fits your lifestyle and apply today at joinbuilt.com/allthehacks. That's joinbuilt.com/allthehacks. Make sure to use our URL so they know we sent you. Terms and limitations apply. Subject to approval and eligibility. Built cards are issued by Column NA, member FDIC, pursuant to license from Mastercard International Incorporated.
I saw you do this thing on your last episode where there was like a remix of Snoop Dogg and I can't remember the other ones and it just has me wondering. I saw today in the news I couldn't validate this in every way I wanted to, but it was that the number one song on iTunes right now was this AI song, kind of like a smooth jazz R&B kind of soul song. And like I found other sources that said that was true and it's definitely top 50 but it wasn't, you know, necessarily number one. What is real in the future? Like when it comes to content, when it comes to music, when it comes to media, I'm just curious how you're thinking about this, not just because we both happen to create content as part of what we do, but gosh, I mean, it just feels like right now, I'll give one other fun example. Well, my wife and I have this Samsung Frame TV and it has an image library and for some reason we were paying like $4.99 a month for Samsung's image library because it was just easy. They had all these images and on St. Patrick's Day we wanted to put something up for the kids that was St. Patrick's Day related and they didn't have it. I was like this is just annoying. How do you get new art on the Samsung TV? Well, you could go find a piece of art somewhere, download the image, put it on an SD card, go to your cable box, put the SD card or it has maybe USB. I was like, what? I'm not going to do that. You can go buy some art app on the app store. And so I just went to my OpenClaw and I was like, hey, can you just build me a skill? I don't want to go find another one. I don't know what malware is out there. Just build me a skill that uses the Gemini Nano Banana API to generate an image and then sends it to the TV.
It took two minutes to go build this.
