Live DEMO: We Build 3 AI Apps (In Under An Hour!)

15:24Grant:Exactly. So there's a lot of flexibility here, which is really cool.

15:30Corey:I think it's time to begin the quest. I don't know how many minutes we'll be in when this is edited, but I'd say we're about 15 minutes deep. So we've got some time to spare, and I think probably enough. We don't have to make them great. We just need to make a functional app that does the thing we say it can do.

15:49Grant:Yeah. And I'm going to try and do something kind of tough for this first one. So I don't know if it's going to be able to do it. And the reason it might not be. Yeah, the reason it may not be able to do it is it might need certain integrations and permissions that I won't be able to give it necessarily right away. So let's see how far we get. OK, you ready? Let's do it. All right. So what we're going to build first, this is something that Corey and I thought of minutes before getting on this call, is we want to create an inbox zero tool. So the pitch that we have here is inbox zero buddy. It ingests your Gmail inbox, clusters messages by urgency topic, spits out one click reply drops and a prioritized to do list and maybe connected via MCP.

16:34Grant:But we'll we'll we'll get to that part first. Does this look good to you, Corey? That looks good to me, Brent. All right, let's run it. So as you can see, we didn't pick any starting template.

16:43Corey:We're just giving it the idea down so they can see the speed the on the thinking tab. Where it opens that gray window. Yeah, right here.

16:54Grant:see the speed this happens with it's crazy so it's structuring the app right now it's then developing the application components and so it's got a couple of files that it's making it's developing the application core which i guess is just figuring out what is the core application and here it looks like it says well i can't see it now we're just gonna let it run but i'll go back and explain how this works now it's implementing the gemini integration so that's what makes this different from like a tool say like lovable which maybe we'll show at the end of this lovable is a vibe coding app where you can take the same concept in the same pitch and say like go make me you know a gmail inbox tool and it will go do it and it can spin it up on its own servers but it can't automatically integrate with uh say gemini because it doesn't have that built into it so you would have to go and get what's called an api key from gemini probably here in AI Studio, and then you'd have to add that in on the back end on Lovable.

17:53Corey:I find for making getting an API key really easy, too, since you're right there.

17:58Grant:Yeah, you can literally just click this and it'll set it up. But what's cool about this is you don't even really need an API key to actually use this, right? That's right. Yeah, it basically just gives you, essentially bypasses that step until you actually go and deploy it, and it handles all that for you. so uh for people who've been watching it's been coding and it's coded uh something called the app.tsx uh then it's got the components where it's got the email input area it's got the icons it's got a pulsing icon and a results display that then it's also got the index files html which is like the web-facing version and the tsx all right and so now we've got a preview so let's pull this up so inbox zero buddy paste your inbox so i actually don't know how it's proposing that this

18:53Corey:function so let's go ahead and ask functionality is it we'll say okay what the heck is this

19:00Grant:functionality please explain it to me in the simplest possible way yet completely explaining every function in a bullet point list because that's easy to read yes we can buzz through a bullet point list so it's we're gonna go ahead and let it think and explain itself like now i have to distill everything that i just came up with it's outlining the functions it's detailing the app workflow it's describing the app It's outlining the app functionality. It's now detailing blah, blah, blah. Okay, of course, I'd be happy to explain how Inbox Zero works. At its core, this application is a smart assistant designed to read a raw dump of your emails and automatically organize them for you.

19:50Grant:The goal is to quickly show you what's important, what actions you need to take, and even help you reply. So core components and functionality, then it kind of explains how each of these work. Do you have any specific questions, Corey? Because I have one. I think at its core is I don't even know how to paste an inbox.

20:11Corey:So perhaps you need to find another way.

20:14Grant:So it's like, okay, this is cool, but nobody really knows how one would paste an entire inbox into.

20:27Corey:I'll put$1 on that. It's going to give us instructions on how to paste your inbox. What did I say? The connection.

20:35Grant:So instead of pasting the inbox, how about you make it so the app works with Gmail instead? It's kind of a creative solution because I wasn't sure how it was going to solve this problem on its own. So it's like, OK, yeah, you just paste the inbox and sure.

20:58Corey:Yeah, we kind of threw this one in here knowing it was going to be the hardest of the three. like the other two that we have planned uh are probably way easier than this yeah presumably so yeah though though i have been wrong before at least once so let's say like uh you as a you as

21:22Grant:a viewer are trying to do this and you come up with a great pitch and you put it together and it comes back with something that is not at all what you were thinking of uh there's two things you could do. One, you can have it do what we're doing here, which is have it go back and rewrite everything. But that's probably not practical. Actually, what would be better is to say, okay, this is what happens when I don't give it all of the context it needs to succeed, which is where the term context engineering comes into play. And instead, what I would do is I would spin up another instance and I would try it again, this time being even more specific with the goals that you're trying to accomplish.

22:00Corey:Yeah, I think that's a really good call out. Like, you know, just like with anything else, it's only going to be as good as the context you give it. Because the truth is, if you're not giving it the context, it's going to make assumptions. And every assumption is, you know, I mean, a human can't read your mind. So there's no reason to believe that Google can't Well, maybe.

22:22Grant:Maybe Google Cat.

22:23Corey:Wrong example. If anyone, Google said. If any company can read my mind, it's definitely Google. I'm watching you, YouTube. Yeah. But you've got to feed it to context. That context is big. It helps it understand what you want to do. It needs to know why you want to do it, what kind of functionality you want, what kind of results you want, how you want your results presented. You know, that's all great information to be giving this thing. Yeah.

22:51Grant:And so if we were going to structure this prompt in a way that would probably be best for it to be received, we would want to give it like the goal at the beginning, which the goal is to create this. You don't really need to give it a role. I think that's probably handled by this UI that Gmail has made or Google has made. It's like it knows it's an it's an app building agent at this point. So you want to give it the goal, and then you'd also want to give it like explicitly how you think of it is going to work. Basically, you need what's called a, well, I'm just going to simplify it, a vision for how it works.

23:28Grant:So you need to have like a mental model of how you want this app to work and how you conceive of it working. And you can break that out either, you know, very, very granularly, or you could just kind of broadly break it out, see what it comes up with, and then adjust your plan from there. Like if you have, this is, we're talking if you have no software engineering experience at all here. If it was me, what I would do is I would start out with O3 Pro.

23:54Corey:And I would go in and tell it, I want to build an app using Google's App Builder. Here's what I'm thinking. Here are the kind of functionalities I want. Build me out a roadmap. of everything that I need to feed this thing to get it to build the app I want and have it build the prompt for you because the truth is it'll save you a ton of time and you'll probably wind up with a better app. If I have a good guess.

24:22Grant:Without a doubt, without question. And there's so many other software engineering tricks and programming tools of the trade that you can apply to improve. so this is just the tip of the iceberg basically here. So I'm gonna show two things real quick. First one, we have this prompt here that says something went wrong and it gives us the option to auto fix. So let's go ahead and auto fix that. This is rule number one for why you don't want to just have it redo everything from scratch because it's gonna create all of these errors. It builds all of this code and now it has to go back and reconnect everything to make it work in a different way.

25:03Grant:so it's a big mess so what we're going to do instead the second thing is we're going to do exactly what cory suggested i'm going to go to o3 and i'm going to basically o3 pro and i'm going to basically tell it exactly what i want to write and um what i want to create and then i'm going

25:21Corey:to have it write the prompt for me so i'd recommend regular o3 because o3 pro is going to take a half hour.

25:29Grant:Okay, so I've went ahead here and I've entered the prompt. I'm basically saying essentially a version of what you suggested. I'm working in Google's app builder and AI studio and want to build the following app. And then I pasted in exactly what we pasted into AI studio. And then I said, help me create a prompt that explicitly defines all this functionality and ensures it can connect to Gmail and the formatting makes sense to this app builder. And then I'm adding a search functions i'm going to say use search to make sure you understand the google ai studio app builder um look it up and how it best receives prompts before you write your prompt and this is a trick that i do whenever i am like referring to a specific platform i will have it go out and search and get all the up-to-date information.

26:22Grant:I'm usually even more specific than this. And I say, you know, as of, you know, newest as of July 23rd, 2025, so that it gets the most up-to-date info. But yeah.

26:34Corey:I do that with news gathering. One of the struggles is LLM training sometimes gets you funky results around times. So I'll go in and tell it like, hey, it's 622 a.m. on July 23rd. Get me everything from the last 12 hours on this subject.

26:54Grant:Yeah, that's a great way to do it, especially if you need something really timely. In this case, AI Studios app builder has been out for a minute, so I wasn't too specific. But if I wanted to say something like, give me the most up-to-date prompt advice for this as of 12 hours ago, then that would be where I would use that. And as you can see, it's just searching the web here. So while it's looking that up, let's go back to our AI Studio Builder. And it looks like it had another error. Let's see this here. It says...

27:27Corey:I think you start it from scratch when we go back.

27:30Grant:Oh, yeah, definitely. When you come in with the new code, I think you just scrap this

27:34Corey:and throw that in and see what comes out, because I'll bet what comes out is way better.

27:39Grant:Yeah, I think this functionally is dead on arrival at this point. Oh, yeah. So this is why you always start over. And I think this actually would be a good place to talk about this. Quick side tangent. Corey, are you familiar with the concept of context rot?

27:58Corey:No.

27:59Grant:Okay, I think you are. You just may have never heard it framed this way before. So context rot is when the longer a conversation goes on with the AI, the harder it is to focus on what's really important.

28:13Corey:And it kind of falls apart and the answers get worse. And yeah, exactly.

28:18Grant:And that is one reason why the first time you prompt the AI, you usually get the best answer you're going to get when you work with it. So yeah, it's why it's usually good to like, hey, instead of continuing a conversation where you left off with special exception, which we'll get into in a minute, it's usually better to just start over and start an entire new chat.

28:39Corey:It's usually isn't it usually going back at every message and reading the full conversation history above it again?

28:45Grant:That's right. So it doesn't remember anything getting huge. Yeah. So imagine starting with like a massive prompt that's like, you know, 30 ,000 words or something like that. Every time you prompted it, it would be hard for the AI to give you 100 % what you want.

29:02Corey:It would. That's a great place for it.

29:06Grant:So speaking of prompts, let's go ahead and see what it came up with here. I'm not even going to read everything that it wrote. I'm just going to go ahead and... Oh, it's gotten multiple. Oh, what's going on here? I'm going to say just give me the prompt, bro.

29:27Grant:And let's see if it can figure it out. Yeah, it looks like it gave you two prompts. You have unsafe changes. Are you sure you want to leave this page? Yes, because that's a piece of junk.

29:37Corey:We're going to start over. And we're still going to finish in an hour. That's what you need to know.

29:44Grant:Corey, do we want it? While this prompt is generating here, do we want to try our second attempt? I see. Let's go ahead and do the second one while we wait. Let's do it. Okay. So the second one that we had planned is meme me up. So the idea of this is you're going to drop any photo. AI spits out a captionable meme suggestion set and it will include the image. So let's see if it can make meme me up. do you want me to send you a ridiculous photo yes please always no questions asked okay

30:25Corey:here's a photo of a pager i've got a vintage phone i have no idea where these pictures came from let's uh are you gonna send it to me on slack i'm gonna send it to you in slack yeah perfect

30:39Grant:I'll download it and throw that in here when this is ready.

30:42Corey:If it makes fun of you and I, Grant. Oh, perfect. Like this.

30:49Grant:Awesome.

30:50Corey:Well, it would help if you have the one without the background. You have the kitty cat one.

30:58Grant:Oh, I do. I do. Let me see if I can pull that up. All right. So we've got all of the stuff we need pulled up. so we are going to try this out so first of all we have got our prompt for v2 of inbox zero buddy all right so i copied over the whole thing looks like it added some json which is this formatting structure here you might see this the more that you mess around with ai code stuff and so i'm going to go ahead and send both the prompt and the json and let's see what happens all right well we're letting that cook we're gonna go back over to me me up looks like okay got something here let's go ahead and working or let's drag our photo in so i've got the photo that you and i want to use cory and this is where we find out how bad it's gonna work it watch on spotify there we go all right there we go now let's generate a meme

32:05Grant:let it work generating witty memes okay we've got some suggested captions when the team meeting requires creative input okay Cory and Grant leveling up our professional game nice our faces when we hear let's circle back on that alright gotta rise out to Cory Yeah. Okay, I'll accept that one. Just two cool cats ready to tackle the week. New headshot dropped. Feel free to screenshot. Okay.

32:38Corey:I gotta say, those are pretty lame. Perhaps we tell it to add in some, you know, witty sarcasm.

32:43Grant:All right, let's say we need to boost up the sarcasm. Let's 10x that. These need to be funny. Sass. they need to be funny and potentially roasting the picture. Let's try that. All right, let's go back to inbox zero. Still coding away over here. Gotta love it.

33:10Corey:We're going to have to cry when this is over. It's going to destroy us, Grant.

33:15Grant:That's what I'm wanting. Okay, so it says an internal error has occurred. I'm going to hit retry. See what happens there. Okay. Let that cook. So now we've got it finalizing the prompts adjustments over here. I've cranked up the sass-o-meter to 11. Alright, Gemini. Alright. You set your expectations. Let's try this again. Click to upload.

33:41Corey:I cast vicious mockery.

33:45Grant:Let's see if we get some vicious mocks. Let's see. oh this is so much fun corinne grant the only make the only men who make catfishing literal their linkedin profile picture probably this is what happens when you let the interns design the marketing campaign breaking news adulthood has been canceled for these two when you ask for playful headshots and get this all right i like it i'm gonna add one more level okay yeah i think i think i think it needs to go to more like 20 okay that was good but we need the sass to be ratcheted up 100x we need you to make us burn like the heat of a thousand suns also um can we regenerate well this is another i want to do i want to have a try and actually put this in a meme format so um that'll be the next the final test there the final test i'm gonna go ahead and check on inbox buddy all right it's cooking okay i'm liking it still going yeah uh cory you you want to introduce us to the third uh app we want to create here and then we'll let these cook on the side okay all right so while

35:13Corey:grant's been working i have also been working and uh as has o3 in building me a prompt for this next app which we are calling the spontaneous side quest spinner this is where you will input your current location, your budget, and any interest you have. AI will then design a spontaneous microadventure itinerary with some food, fun photo ops, and odd landmarks, maybe a bar. You know, anything fun around to do that might make for a fun time. So I've got the prompt in here. It's lengthy, and we're going to see what happens when we hit go.

35:54Grant:Let's do it.

35:58Grant:all right whoa oh that's okay that's code's looking good that was the prompt yeah that was the prompt okay

36:09Corey:let's drop this down so we can watch it work at a speed that we would never reach in a billion years

36:15Grant:i know i know imagine if you were trying to learn programming and you were trying to build this

36:20Corey:yourself this would be so discouraging like why am i learning this well i mean remember last week

36:26Grant:or a couple weeks ago we did an episode on ai education right so if your goal is to actually learn coding then you do want to be doing a part of this yourself um and so that's you know a topic for the previous episode what do you see includes root element for the app till i have an error

36:48Corey:to win CSS CDN link and links to Google Fonts and a PWA manifest. Retry.

36:54Grant:Interesting. That's never happened before.

36:56Corey:Yeah, I've not had that one happen before. It's really not going to end my hour here, though. Outlining the structure.

37:05Corey:If this kicks out, I'm going to go back and simplify my prompt a little bit.

37:09Grant:Sounds good. Yeah, sometimes I feel like O3 will definitely overcomplicate the prompt.

37:15Corey:I was throwing it weird because it was giving me some of it as text and then some of it as SQL for whatever reason. And then a certain amount of it in loose text. And I'm afraid the loose text that it threw down there was poorly pasted JSON, which might be enough to break this and cause the JSON problems.

37:42Grant:I can see that being the problem.

37:44Corey:I'd be happy to build it for you this sounds fun and creative interactive form okay cool cool cool awesome let's go for it we're building out alright the real question will be how long does this take

38:00Grant:I've got an update on the meme me up if you want to see the

38:04Corey:latest version in that case I'm gonna take this one down and let Grant share up meme me up and we'll see how that one's going see if it is sufficiently shamed us and embarrassed us for our whiskers in that image.

38:17Grant:Alright, so I love the first one. It says, when your midlife crisis involves a cat filter and questionable life choices. The only duo whose fan club consists solely of their moms and a confused house cat. Pretty sure I've seen more convincing cat impressions from a vacuum cleaner. They say, never meet your heroes. Clearly these guys never met a mirror with this photo. somewhere a litter box is calling and who wants its whisker aesthetic back okay that didn't make any sense that's some ai slot there um obviously this is great but can you take the initial image and using imogen generate the memes in a meme format so text on image once the user selects the caption they like.

39:12Grant:So this functionality should work the same way it does now, but then add a button so that the user can select their favorite meme, favorite caption, and memify the image with imaging. Let's let it work on that. All right. I have an issue. What's happening on your end? You want to share your screen?

39:43Corey:Yeah, it looks like it's done, but I can't figure out how to get the preview to come up. Hmm.

39:52Grant:You know what you can do in that instance is you can tell it, let's load the preview.

40:00Corey:That's a good idea.

40:01Grant:Let's make sure there's a preview to review. So add some sort of HTML front end for us to look at and test it. Because I have the same problem on inbox zero, buddy.

40:18Corey:Did you? Okay.

40:23Corey:Interesting. Okay. It's putting us together a preview. Now we analyzing map integration. Let's see if we can tell what it's calling here.

40:35Grant:it might be that we need to tell it like hey build a front end for us to test it um that might be the way you need to phrase that it's integrating with google maps as as you would

40:46Corey:hope um developing map display component interesting uh i'll test it out and move it to the budget bar budget bar nice let's see i like the sounds of this yeah uh it's still rolling still cooking so we're in the wait and see phase this is really cool grant i can see a lot of uh a lot of applications here and i'm really impressed with what i've seen like the truth this, I am fully acknowledging the fact that I have no idea how to go in and fix this. And I think that's important to call out. Like, I could go in and I could understand how to get it to call a different model. I could make it do that.

41:36Corey:I could change colors and dimensions. I could drop in some new files, that kind of stuff. But as far as the nuts and bolts of what's making it work that's a thing that i as a non-developer and it's just a skill that i am missing yeah and uh

41:56Grant:transparently pretty much everyone who's trying to vibe code and build their own app is going to have that same problem and there's a lot of like i mentioned there's a lot of resources for people showing how to you know how to ideate and design the best uh tool possible like before you take it to code right there's a lot of um ways that you can improve your like initial prompt in order to create something that's going to have the best shot on goal in order to win uh there's also a lot of um ways that you could potentially improve your skills by reading other developers who are experienced in this space and they're actually trying to figure out what is the best way to code with AI in this way.

42:40Grant:And also just kind of try to teach yourself on the fly. Like, let's say, try to build something really, really simple and then go back once it's done and it's working. Ask Gemini, OK, explain exactly how this works to me and show me step by step, explain and break it down and then go through the code and then ask, OK, for this piece of code, what does this code do? How does it work? You can even ask it to do this, let's say you're working in 03 Pro or Gemini 2.5 Pro in the chat model. You could ask it to help you design a program from scratch and all the features you need before you ever take it to code.

43:17Grant:So like say, okay, this is what I want to do. Let's break down feature by feature what I need to build in order to do this and then work with it backwards to reverse engineer it. So you're not coming up with one sentence prompt, take it to Google Builder and expect it to build you a full application. But instead, you are working incrementally in order to program something. I think that's the way I would go about it if I were. That makes sense.

43:43Corey:That makes sense. I can see why that would be beneficial.

43:49Corey:Yeah, I'm still having hiccups here. I'm going to figure this out, though, because we have a time limit, and I just refuse to not hit it. I'm going to go start over, and I think I'm going to just do this prompt differently than I did originally. Do you have any updates on your end?

44:08Grant:Same issues over here. I'm letting it try and work itself out. It looks like here it says, to provide a rich preview for the user interface without needing to make a live API call, I'll populate the application with some realistic mock data. Okay, so. Fingers crossed. You know, one thing I would recommend people do is if they're going to try to build a simple application like this, I would add some language that says, like, don't over-engineer it. Make it as simple as possible for a 1.0 version. Because these O3 prompts might be, like, really over-engineering it, to be honest with you.

44:47Corey:Yeah, and the truth is, yeah, they may very well. It's also possible that what we're dealing with is that the SideQuest spinner is not as easy as it sounds on the surface because it does have to access a map. It has to be able to recognize current events and businesses in the region and be able to tag that with a budget and have an idea how to be sure that they're spending the way you want to along with a calendar. so like they're uh this is probably actually it's it's a lot more impressive the second time

45:18Grant:watching it go yeah this is kind of cool to see it cook yeah it's it's absolutely cooking right now

45:26Corey:yeah maps spinner budget guard budget bar it's doing the map view there's a great well this is loading itinerary card there's a great uh video that i watched to

45:40Grant:kind of helped me learn how to use this tool. And it is by NoCodeMBA. And it's a step-by-step tutorial that shows you basically like how to get started on this tool, AI Studio. And it shows you how to debug and handle errors and like tips for that. It's got pros and cons. So we're going to go ahead and link to that in the description of this video. So if anyone finds this idea kind of fun and wants to check it out and try it themselves, you can totally watch that video. And we want to see what you know.

46:15Corey:You know, we want to see what you built. And also, I would like a quick mention here that if you're enjoying this, if you like the channel, if you enjoy the newsletter, please give us a like and a subscribe. Whether you're watching on YouTube, listening on Apple or Spotify, you know, we really appreciate that. We love to interact with you in the comments. We think that's a ton of fun and enjoy the insight you really bring to the table. And now a spontaneous side quest spinner seems to be functional. So let's check this out. We're going to put it where I'm in. We're going to say I'm in St. Louis, Missouri.

46:55Corey:We're going to say a budget of, we're going to say$2 ,000. That's a lot, but we're going to roll with it anyways. And we're going to say four days. It probably should ask us for a date in retrospect, but it's going to generate an itinerary. Planning your adventure. I love the little perfect forms. Wow. I gotta say I really think that this is a big step up from the first iteration I've seen on these so far. I think maybe the answer is going and letting a reasoning model build you out a really baller prompt. Oh wow. Oh wow. Budget tracker 157 of$2 ,000. View interactive day one.

47:44Grant:uh wait view interact does that work yeah let's close that let's see what it does

47:51Corey:oh you've got to be kidding me oh oh snap oh snap let's go oh my gosh there's no way that's the

48:03Grant:most optimal route though let's be honest no that's absolutely not the most optimal route

48:07Corey:i'm assuming it's based on what we're seeing here let me let me blow this up where it's bigger yeah it's starting at all the right places gateway national art gateway arch national park you obviously got to see the arch if you come tost louis got a courthouse city garden ted drew's you got to come have frozen custard if you're inst louis fun fact i did you did you did that's right yeah uh forest park is an obvious hit the art museum the zoo those are both in forest park so It's also understanding that those are in the same vicinity. Now, the Anheuser-Busch Brewery Experience is not in that region, but it's 15 minutes away.

48:48Corey:It's not far. You could absolutely still do these in one day. And I highly recommend the brewery experience. It's a lot of fun and comes with three beer. Say less. Day three, Science Center, Botanical Gardens, the World Chess Hall of Fame. Go to the Hill for a fancy Italian dinner. did it recommend a re-spin food stop. You can re-spin, Grant. What? Certain elements if you maybe want to try a different one here. It's the spinner. It's the spontaneous SideQuest spinner.

49:19Grant:Oh, yeah. What happens if you click it?

49:22Corey:We're going to find out. Re-spin food stop.

49:25Grant:Does it work? That function might not be ready yet.

49:28Corey:I don't think that function's ready. I'm still, okay. city museum you gotta hit city museum you gotta go to gus's pretzels and eat a pretzel dude this is this is a great if you were going to come tost louis had never been here before also i'd like to add that it's not exaggerating like that's 157 bucks of fun right there four days i mean 18 free free 10 free free free 15 yeah fun fact the zoo is free inst louis so

49:59Grant:if you're going to go. It is.

50:01Corey:If you park on the street and don't mind walking a little bit, you can park there free. This is grand.

50:07Grant:Grand. Yo, Corey, did we just solve travel planning?

50:14I absolutely did not expect something that advanced from,

50:20Corey:like, there's not a stop. I just one-shot that thing. That was one drop. You did, hey, Corey, you did it in an hour.

50:25Grant:You did it in under an hour. I did it in less time than it would take me to, I don't know, definitely drink a cup of coffee.

50:33Corey:So what did you do? What did you do differently? Less time than it takes to pour the cup of coffee.

50:41Grant:What did you do differently from the first prompt to this prompt?

50:45Corey:I'll tell you what I did differently. The first prompt, I grabbed everything that 03 gave me. And like I said, it broke it apart in two languages and there was text in the middle. Remember me saying, I think that's plain text JSON just dropped into a chat response, which looked bad. And I was hesitant to even try that. So I pulled it out and I came back and I just grabbed the prompt at the top, what was in plain text, dropped it in, hit enter. And I mean, we have software.

51:13Grant:Yeah.

51:14Corey:I cannot get over this, that it has gone. It has built this map with all of the directions. Like you would want to stick your hotel in here somewhere, maybe wherever you want to eat lunch. But you could run off this map most of the week.

51:31Grant:Even just this alone. So like I actually went to Missouri recently, met up with you in person. And yeah, it was a great time. But, you know, just having this little like map, like it would take so long to add all of these stops onto your own Google Maps. But now looking at this, you can just say, oh, my gosh, I know that I need a hotel somewhere in that, you know, X amount mile radius. And I'm good because I'm, you know, in the middle of everywhere I want to go.

52:00Corey:Yeah, I need a hotel here.

52:02Grant:Exactly.

52:03Corey:The other thing I would say, we were initially a little concerned that the map was maybe kind of chaotic. But what I'll point out is that one through four, like one through three are walking distance close, like like.

52:18Grant:Wow.

52:18Corey:Less than three blocks with with Ted Cruz Custard being a walk. Forest Park, Art Museum, Zoo, that's that's all parking in one spot. You're there.

52:28Grant:Right.

52:28Corey:Well, maybe maybe two spots. But you theoretically, if you're in the same park, your point is that these are clusters like this is these are clusters.

52:38Grant:over here is a day. That cluster at Forest Park is a day. Yeah.

52:42Corey:It has absolutely bundled them by region intelligently, which impresses me a lot. That is so cool. Like, this is probably the order I would hit them in. Oh, it did give us a new Italian restaurant, Zia's on the Hill. Zia's. Oh. The Hill is a very, very old Italian region of St. Louis that is covered in all the best handmade pasta around.

53:06Grant:Wait, that's interesting. So are you saying that when you hit the button to re-spin, it did end up spinning a new restaurant that just added to the Google map?

53:15Corey:Because it, well, let me show you this. Over on this side, it doesn't give a name for the restaurant. It just said, the hill, Italian dinner. Savor authentic Italian cuisine in this historic. So yeah, I did something.

53:31Grant:Click it again and let's check it again.

53:32Corey:Click it again and we'll go see. I might be crazy. Yeah,$50 of that budget is the Italian dinner. And that might be a little low for a good Italian dinner, if I'm being honest.

53:42Grant:That's fair. What's great about this, I just want to highlight, is that now as you're testing it, you can see this is a user design thing you need to fix. Yeah. Did it update?

53:55Corey:Yeah, no. No, it didn't. Yeah, apparently the re-spin food stop button doesn't work, but you bet your tail that's something I can fix in the end.

54:04Grant:Yeah, no, you can just, you can basically go to Gemini now and say, hey, so everything is working great, except the re-spin button doesn't update. And you might want to, you know, give it a very narrow objective in order to try and solve that, just that one functionality without breaking everything else. Yeah. But, you know, with a little thought and engineering mind, you could probably put that together and say, okay, what's the best way to solve this issue? and we're going to try something else let's do itst peter's one thousand dollars this is more

54:42Corey:the suburb area i'm in two days and see what it does i want to see what it gives us for this one oh let's i forgot to swap screens back over my apologies uh yeah i put it onst peter's missouri $1 ,000 budget, two days. This is awesome. This was the wild card. This was the last option on the list. Almost didn't make the cut as something we would try to do. This was really like I thought, I just thought there's no way it's going to do this well because that's pretty complex. Budget tracker,$230. View my interactive route. Where is it at? 370 Lakeside Park. That's a cool spot. 370 Lakeside Park is a, it's five minutes up the road from me.

55:33Corey:It's a park, little campground, some little stuff you can do. Nothing major. Shack, breakfast at the shack, okay. Serendipity homemade ice cream. I don't know them, but I'm going to. Bandanas, okay. Bandanas is good. Bandanas, Shamrocks. Shamrocks is the spot. it's a just a cool little bar and grill kind of place good food live music once in a while nice nicest peter's recplex really neat stefanina's pizza yeah if you're inst louis that's what the other one didn't do it didn't suggest eatingst louis pizza which is absolutely a thing i would have not suggested either so i think that it it gets it maybe okay go back up to how this works jj's i've eaten jj's okay maybe we would add uh an extra

56:24Grant:text field here that says like um something to the effect of wild card input from the user

56:32Corey:and yeah like i want to go a bunch of breweries or i want to hit uh all the parks i want i want to play disc golf too while we're there work on enter i want to only eat at the mostst louis places possible like i want to eat all of thest louis food yeah i want toasted ravioli and yeah

56:50Grant:oh this is sick yeah this is awesome how is uh how is yours going i'm gonna i'm gonna stop here we're gonna i'm gonna continue working on this after this but i'm i'm definitely calling that a win today i think what i like about this like okay obviously we're not applying you know coding best practices we're kind of just winging it and having fun but what i like about it is it just shows you like the possibility of what you can do um and it just sparks your uh creativity which i love yeah um okay so i'm gonna upload our picture one more time allegedly the functionality is built in so that now once we enter our picture we'll get a series of captions and then we can potentially generate a meme from those captions so let's see yeah okay so let's see the their spirit animals are definitely mild mildly confused house cat and overpriced artisanal beard uh cory i think that one might have been at uh directed at you oh that's a cold jab uh what do you order fierce on wish and get cory nolz and

58:04Corey:Great Harvey. Producer Corey Doles and Great Harvey is code for

58:08Grant:failed internet challenge. Okay, rude. The only thing sharper than those whiskers is the regret of whoever suggested them. If I peeked in junior high drama club was a picture, this would be their yearbook photo. That feels that hits for me. Let's do it the first one though. Okay, I agree. So now we're going to generate an aneem image. Okay. Alright. it kind of misses the point of a meme. You're supposed to be able to see the image and the text.

58:41Corey:I would also say that, you know, though, it's not a fail. It would not take much to make that work.

58:49Grant:In fact, you know what I'm going to do? So what you can do is you can take a screenshot, right? And then you can upload that screenshot and you can say, this is how the meme generated. can you fix the prompt for Imogen or the code, whatever is necessary and most elegant slash simplest solution. Can you fix the prompt for Imogen or the code, whatever is the most blah, blah, blah, blah, so that the text doesn't completely cover the image like regular memes. All right. And what's awesome about that is that because these models are multimodal, which means they are image based and they can also accept video and audio.

59:38Grant:It can basically take that image and then it can see the context of what you're seeing so that it can get better context for how to solve the problem.

59:48Corey:Exactly. That's awesome. Yeah.

59:52Grant:So, all right. So we got that one back at work. What happened to zero inbox? So we had an error. Basically, I was trying to get it to give me a UI element front end so that I could actually look at it. And in this case, it seems like there's having a lot of problems. This is not an easy solve, right?

1:00:13Corey:Go back to that original request you gave to 03 and grab what's in. It returned two boxes. is grab only what's in the top box and see if you have similar luck to what I did.

1:00:30Grant:Let's do it.

1:00:35Corey:I would say we firmly have two functional apps. We have one good one and one that's really, really close to being good.

1:00:45Grant:I just wonder if it's not going to be able to do it because it requires an integration with Gmail.

1:00:52Corey:That's definitely maybe a thing.

1:00:58Corey:You know, though, it's logged into your Gmail. It's logged into your Google account right now that you're using. It should be able. It's probably not a far leap for it to be able to make that connection, I would think, maybe.

1:01:15Grant:yeah i i think too another thing that we could we could do is we could try and show a couple alternatives while this is loading to ai studio if you want we mentioned them in the beginning why don't we go ahead and pull those up so one of the first ones is lovable so the reason i like lovable is because it hand it does it's as close to the google app builder um option as possible in that it integrates with a backend called Supabase. So if you build something on here that you like and you want to spin it up, you can spin it up with a backend database on Supabase.

1:01:56Corey:So with a click deploy, stick a credit card in to pay for any associated fees. Pretty much, yeah. Which you should be able to look up for, by the way, because fees on apps can get up there pretty quick if you're not, if say suddenly people are using it and you're not paying attention and don't have a solid monetization strategy, you could really quick have an$80 ,000 surprise credit card bill.

1:02:19Grant:Definitely. Yeah. So, you know, this works the same way where it's a build me an app that is a simple meme generator. You know, it's not going to work as well because it doesn't have the easy integration with AI. So let's do something even simpler. Build me a calculator app. And I'll just...

1:02:40Corey:Okay.

1:02:41Grant:I'm not locked in at the moment, so I'm going to go ahead and do that off screen. But another option, right, is Claude. Yeah. And so Claude has what's called projects or artifacts. And artifacts are cool. So what we're going to do is I'm going to make sure it's turned on. Okay, let's see. And I'm going to say, build me an app that works with Claude AI and is a chat app to chat with Claude in a computer. Let's say like a 90s computer, computer aesthetic using artifacts. facts let's try that see this word i like that yeah so there's multiple ways that could fail so we'll just let that run um in the meantime i'm gonna yeah and most of the ai tools now

1:03:45Corey:even in a normal chat window are capable of building something uh like you know most all of them you can absolutely do some level of this now spinning it up and deploying is where it gets tricky like you know um you know if you're building these in some apps you're gonna have to you know containerize it at some point and do all of the things required involved to in deploying an app uh

1:04:16Grant:but they can do the job um they can while this is while this is running i'm gonna go ahead and show you just a couple other options i've not i've used them before but marketly slower than

1:04:28Corey:Google too, by the way.

1:04:30Grant:You're right. So VO, this is from Vercel. Vercel is kind of like, what's the simplest way to describe it? It helps you easily spin up and run things and deploy them quickly. So I think it's basically like a wrapper for AWS, but if you're just trying to code something simple, you can do that with VO, and it works the same way as Lovable and Google does in that regard. Then another one is Bolt.

1:04:58Corey:You're probably buying space on their AWS instance and not having to go out and get your own and pay a fortune and set up all of the work and labor that goes into that. It's really a smart way to start with a small app, I imagine.

1:05:11Grant:Exactly. Bolts.new is another one. As we mentioned, I think Grok also lets you do this. Grok is XAIs from Elon Musk. That one, you can create code on here. but yeah so that's those are some of the different options that you have

1:05:30Corey:all right so that was that was a lot of fun uh you know i i was a little out of my element but it was surprisingly easy like i while i have a little cursory knowledge of coding and NCS, it's very cursory. And like nothing I did there involved it at all. I never clicked into the terminal and did anything other than just poke around and see what was there. Like there was no need for that. We literally just, we went to 03. I had it build me a prompt for what I want. I came We dropped the prompt into the tool and bam, we had a highly functional multi-step app that was calling calendars and making a budget and doing Google searches and planned an interactive route.

1:06:27Corey:Things like that. You know, it was awesome. Yeah, it really was. It really was. So we got we got two out of three that I'm willing to say we built an app. So something we wanted to show here real quick before we go, I'm going to open this tool back up. So there we are. If you want to deploy one of these, we wanted to show you what that involves. The answer is not a whole lot. So we'll click deploy to Cloud Run over here and deploy app on Google Cloud. It'll deploy your app as a Cloud Run service. It'll be accessible via a public URL, your API key will not be exposed, but it will be used, meaning it will charge your card, which is something that's important to keep tabs on.

1:07:13Corey:We'll talk more about in a minute. You have to already have a Google Cloud project built up, which we don't at this moment to connect it to, but we probably should do that soon. And basically that just makes sure it's able to build. But from there, you know, as soon as you have a project in place, you literally just hit deploy app and there are maybe a couple steps on the other end but it's it's pretty quick it's super simple and uh anyone could do it i literally believe right now that anyone is capable of doing this and that is a doubt i had going into today grant yeah is that this is genuinely a thing anybody could do but like i am i am confident this could be easily taught look can anyone

1:08:02Grant:build any application? No, not even close. But can anyone build a small micro application with a very specific use case? Yeah, we're pretty much at that point. And all it takes is trial and error, going in, figuring out what works, what doesn't work, maybe doing multiple different versions, you know, where you change the prompt, see what happens, make edits once you have something that you like, so that the functions work the way that you want them to. But yeah, Anyone can build an application at this point.

1:08:34Corey:That is so crazy. Like, that has not always been the case. That is a new development. There have been things, like Google App Sheets, been out there where you could get it to, like, you could have an app of a form or a spreadsheet turned into kind of a loose project management tool. But they were not as robust, as attractive, as simple to make massive change to. I think that's the key is that, you know, you can add a feature to this with a sentence while you, you know, run the dog outside.

1:09:10Grant:Yeah, yeah.

1:09:10Corey:Just take a break, come back to your desk, and it's there and it's ready. And that's awful huge. Grant? Yes. If you were going to start, say you were going to sit down and start from scratch, what should a person do?

1:09:24Grant:so i would say that you could apply this five-step quick start framework so you know you can go to ai studio you can pick a template pre-built template based on the type of app that you're trying to make uh then you would craft the prompt so that could be a couple steps trying to figure out you know first of all you need to outline exactly what you want to build second of all you could take it to a tool like o3 touch bt o3 uh and work with it to help you create the prompt

1:09:53Corey:Or 2.5 Pro's chat interface, for that matter, too.

1:09:56Grant:Yeah, technically 2.5 Pro is a little bit smarter, too, so on the artificial analysis leaderboard. So, you know, that's a perfectly valid tool to use. You can do it all in AI Studio if you wanted to. And I would recommend inside of crafting the prompt, take time to work with the AI and work out your own ideas and try to figure out exactly what features you want and which ones you don't and go back and forth with it. And then based on that conversation, have it create the prompt for you. So that's step two. Three is then you test it. So actually run it in the AI Studio Builder and see what comes up, what works, what doesn't.

1:10:34Grant:Just like we were doing. Yeah, just like we did today. And then iterate. So let's say like 90 % of the functionality works, like what happened with Corey's AI Trip Planner, but maybe like one of the functions isn't working. So then you go in and you iterate and you change it or maybe you want it to look different. Maybe you want it to, you know, function differently. So iterate with it that way. And then you can deploy it. And that's the final step.

1:10:59Corey:And you don't have to tell it where in the code you think the problem is. Just tell it feature X isn't working.

1:11:06Grant:Yep.

1:11:06Corey:Here's where it's happening. Here's what it's doing. Literally take a screenshot. From that. Yeah.

1:11:11Grant:Yeah.

1:11:11Corey:Take a screenshot and upload it just like Grant did. I forgot about that.

1:11:14Grant:Yeah. Yeah. In fact, we never showed the result of that, but I'll pull it up at the end once we go through this other part. So if you don't want to deploy it through Google Cloud, you can also download the code, as we mentioned. So it's really easy to do from there. You can download the code and then you can run that on another server that you prefer. I think another one you could use would potentially be Vercel. Another one is Supabase. I don't know if you can, I think that's more for database. So maybe instead of Supabase, you would use like Railway and we'll include links to both of those down below.

1:11:47Grant:Not sponsored. They just happen to be cool tools for this. Yeah, yeah. Or anything else, you know, just do some research. Have the AI look up what's the best tool for what you're trying to do and where to host it.

1:11:58Corey:Yeah. Yeah, it doesn't have to be AWS or Azure and cost a lot to be a large and be a large project. I mean, you could spin it up small very reasonably.

1:12:08Grant:Yeah. So then, Corey, what about as far as, like, governance when it comes to these tools? What do you do?

1:12:15Corey:Before we get back to kind of the money talk earlier, there are a few things I would recommend you really want to watch out for in doing this. Number one is I think you want to set some kind of a token limit within your app. You want your tasks that an end user is able to do. You don't want it to run away and pull out 2 million tokens and cost you$42 per user. You want to make sure that what you've got is streamlined. It has enough to do the job well. So you want to get an idea of what the token count of your thing is. That's a thing we can get into a little deeper. If it's something that interests you, let us know and we can talk about maybe doing a future episode on tokens, how they work, and token counts and token count tools, the various things to help you with that part of this.

1:13:09Corey:You also want to monitor its output regularly. Go use it. You know, presumably if you spin it up, you'll have access to some kind of a database that's probably saving most of what's running through it, where you can go through and look for anomalies, problems, make sure it's not recommending people go somewhere that, you know, is dangerous or inappropriate, maybe. You know, you want to really watch out for that kind of stuff because, you know, I would assume directly you're going to be liable for this app if you put it out into the world. So treat it as such. Third, and not necessarily most importantly, but one you really want to watch is cost controls.

1:13:47Corey:Like you can, with most of these tools, set maximum amounts of money for how much they can ding your card for, how much is running through, and have stopgaps in there to notify you and let you know that, you know, hey, this is getting out of hand. Because if you don't, it can just bill your credit card according to how many people are using it. And you could in a big hurry, you know, depending on the app you built. Like the truth is this one we're running would probably not be that cheap. It's pulling up maps. I mean, like those tools are free, but they're also data intensive. And if you're, you know, dealing with math and budgets, you're dealing with calendars, you're dealing with maps, that can get costly.

1:14:31Corey:And if you had suddenly, you know, maybe a few hundred, a few thousand, a million, God forbid, I mean, as great as that sounds, it could absolutely financially decimate your existence. And I think it's important to just make sure and keep tabs on that. I'm not saying don't do it. I'm just saying make responsible decisions with it. Be sure that, you know, if you're going to put something out into the world, you're ready to put your name on it.

1:14:59Grant:Yeah. And your checkbook. Yeah. To be really precise here, what are you talking about is putting a limit on the API number of tokens on the back end. So you can set a budget where you can auto-fill your card. You know, once you hit, you know, say you have a million tokens and it costs$6 from Google. Once you hit that million token limit that's been output, then it would re-up. Well, you can also put a limit where he says, okay, you can spend up to$200. So I'll make this free up to the first$200 a month and past that point, you know, it's cut off and you won't spend anymore. That does mean your app won't function past$200.

1:15:39Grant:But that's to protect you so that you don't make something that you think is cool that's connected to your API key, which means that you are the one fitting the bill. And then it goes out there. And then you root your house. Yeah, a bunch of people use it and then it's like you're toast. So you don't want that to happen.

1:15:58Corey:Yeah, you know, if it's an app where it makes sense to have a monetization strategy, you know, you should. There's a cost associated with it. Your time has a value. The tools you're using have a value. You know, it doesn't have to be much to go a long way. That's one of the beautiful things about AI right now is that, you know, I mean, even the most expensive models, you can run a lot for a million tokens.

1:16:18Grant:It's true.

1:16:19Corey:Like a million tokens is still a lot of tokens, you know, depending on what you're doing, obviously. If you're doing heavy data work, if you're building apps, that's a lot of tokens. But if you're not doing those kind of things, you know, if you're just building out, you know, a simple game or an app with a couple of cool functions, you know, you can run that pretty easily.

1:16:41Grant:I'd also say that there's a certain point where you might want to upgrade the tool that you're using. So right now this is available for free if you use AI Studio, but presumably at some point it won't be free. or in the future, you know, they'll pay wallet in some respect. And it's certainly not free to post it and have it be connected to your API key. There's a cost associated there. So keep that in mind. But then there's also, you know, other tools like Google offers Vertex AI, right, which is their like more advanced, you know. The non-training wheels version. Yeah, really complicated. You could do that for scale or for, you know, building your own custom model and running those.

1:17:19Grant:other people are going to tell you how to use that because that's quite complicated and not for this video but then there's also you know other coding tools so we'll do another video on this but you could do a command what's called a command line interface tool where you're actually building with the code on your computer so that's another tool that you can level up to at some

1:17:40Corey:point so if you go out after this if you watch this and you're like i'm gonna build an app absolutely do that and share it. If you don't share, you don't have to share the app. Record, you know, a quick YouTube short kind of video or something. And share it with us. We want to see what you're building. We want to see what you're doing, how you're using it. And I really think there's a lot of fun opportunity here. And if there are cool ones, we'll share them in the neuron. You know, there's always a chance we'll toss it out there.

1:18:12Grant:Definitely.

1:18:13Corey:But I'd say we have reached round robin time for this episode. And now today's question. What is the smallest, pettiest problem in your life that you still need an AI solution for?

1:18:32Grant:Wow, this is tough. There's a lot of things that I would want to use AI to help me with, but I'm not ready to give over either control. to the AI to do or access to the AI to access. So the perfect example is the Gmail inbox one. I do think that there is a place for an AI email assistant. I'm just not at the place personally where I'm ready to do it. So I'm still manually reading all my emails. And I think that that's probably the place where I really need some help. And I'm notoriously an email-aholic. Like I subscribe to everything. I never delete anything. I keep it all. So at some point, I'm going to need some help in that area, but I just haven't given over to the machine there yet.

1:19:22Corey:Boy, that's going to be a fun one to let go of, though. I so dislike email as a rule. It's so valuable, but it also just gets out of hand so quick, and it becomes a mountain. And I think for me, my smallest, pettiest problem that still needs an AI solution is that I'm tired of deciding what's for dinner. I think I've mentioned this in here before. Yeah. But every day, me and Melissa, we go through this thing together where I ask, hey, what you want to do for dinner tonight? She goes, I don't care. What do you want? And I say, I don't care either. What do you want? We do that for a couple minutes.

1:20:01Corey:And then inevitably one of us tosses out a solution and the other one's like, no, I don't want that. And then we'll do this back and forth. And it happens every day. And it's exhausting. What I want is a simple app where I can put in all of the foods we'd like to eat. These are the things we like.

1:20:24Corey:Pick debtor.

1:20:25Grant:Corey. you realize that you've built like the 90 complete version of that in the ai trip planner right like all you all you have to do all you have to do to to make your dream come true and solve this problem that happens to you every day is go back to ai studio and say okay i need a tool that will like randomly pick a a restaurant inst louis for me to order food from tonight That's right. Like, you know, just automatically decide. And maybe you have a little input where you and Melissa can choose what you want to do. And you can just go and build that.

1:21:01Corey:Maybe it can also be, do you want to eat at home or do you want to order in? There you go. Or go out. Anyone's got three options.

1:21:07Grant:And if you choose eat at home, then you have like, you take a picture. It says, okay, upload a picture of your current refrigerator right now. You go in, you upload the refrigerator. And then it, based on what it sees in the refrigerator, you have it like transcribe every item. And then think of like, what can you make from that?

1:21:25Corey:I love it. It's the most tiring decision. And the reason it's such a tiring decision is because it's at the dead end of both of our work days. We're both exhausted. We're tired. And all we really want is for food to appear in front of us. And I think that also is on the horizon. It is not unheard of that we could get an app to order DoorDash for us because not only am I you know too tired to go pick it up I could also be too tired to like go order it oh yeah well we ran a story about just have something here for food at five that's all I'm asking the uber ceo is

1:22:02Grant:is basically building that and then we've got uh here in glendale we've got cocoa robotics those little robots that are like coolers on wheels those are so cute yeah we got those driving around like delivering food to people so that's happening that's coming I had a run-in with a robot the

1:22:17Corey:the other night i was at a quick trip and they have these little robots that um sweep and mop

1:22:24Grant:the floor they just run around the stove is that marty but it's got the big eyes it might

1:22:28Corey:everybody hates marty with eyeballs yes people don't like it i think it's psychological because the eyes are so big it feels like he's spying on you it ran into me or probably that's why people don't like him maybe that's why i ran into him is probably more like that would make sense if that's him i know they said it costs like thirty thousand dollars i'm like that is a lot of money

1:22:50Grant:to clean your floors but uh well but it was it was pretty cool to watch because like it had the eyes and i just i was so weirded out by the eyes yeah no this is the thing people hate people hate marty

1:23:02Corey:yeah that's cold i mean it wasn't like it didn't do anything wrong it's cleaning the floor and i'm not having to yeah i'm not working marty people don't like the uh the ai surveillance state gory that is true that is true i have a video of him that i took in the store that i will uh i'll have to share let's throw it up on screen here this is the guy oh interesting

1:23:31Grant:he is the guilty robot that might not be marty that might be marty's like short brother like you know the movie twins with arnold schwarzenegger and uh danny devito

1:23:42Corey:that's marty's twin it says something robotics i wrote down what its name was i need to look at that but yeah this dude it just runs around the store and it mops the floor and dries behind itself so like you can't even tell it's been through when it's gone it just cleans which i I thought was super neat.

1:24:04Grant:That is interesting.

1:24:07Corey:And it blinks, it winks, it does things while it goes around the store.

1:24:13Grant:It's kind of like the next generation of Roombots. Like the original Roomba is like, what do we just vacuum? This is like vacuum plus like.

1:24:20Corey:Shampooer. Yeah. Yeah.

1:24:23Grant:Plus a friendly face.

1:24:25Corey:Yeah. I had a run in with him though and found it pretty cool. We were coming back from a concert and just stopped at the gas station and I was grabbing a snack and I looked down and I was like, oh, this is weird. You know, I get into these other ones at the grocery store they have now. You see more real world robots than people give credit for.

1:24:44Grant:It's true.

1:24:45Corey:At our grocery store, the robots are like inventory scanners. So they're like five feet tall, four or five feet tall. And they just roam up and down the aisles looking at what's on the shelves and presumably placing orders to get more products in where things are out, they're empty. But yeah, it's pretty common to go to our local grocery store and run into one of those now.

1:25:13Grant:Wow. Have you seen, so in LA now we have the Tesla diners. Have you seen any footage of that where you've got an optimist robot giving out popcorn to people and you can charge your Tesla and get a burger while you're there? Apparently the burgers are like$14, but, you know.

1:25:29Corey:I think we're going to need to need a robot episode soon and really nerd it out because there's so much fun. Like, it's just neat. There's a lot to learn. They're getting real.

1:25:41Grant:They are.

1:25:42Corey:And, you know, I mean, it hadn't even clicked to me until just now. That's two places in my everyday pathway that I now see them in the real world. And that's one step closer to having one that can do my dishes and make sure my laundry is nice and clean once I get off work.

1:26:01Grant:It's coming. Just wait. It's on its way.

1:26:04Corey:Yeah, we have an agreed upon price at the house as to when we will take the robot plunge. So I'm keeping my eyes open. And I know it's still a ways out, but there are just a few tests that we work a lot. We're busy people. We go hard all day long on purpose. It's just who we are. And we have some things we'd like to not have to deal with anymore.

1:26:27Grant:Respect. I respect it.

1:26:31Corey:We've just accepted it and moved on. Well, everyone, thanks so much for joining us and building along with us today. I hope you enjoyed today's episode. Please take a minute to smash that like button. subscribe, share it with your friends. Really, it means a lot to us and helps us keep doing what we are. Also, if you haven't already, please take a minute, subscribe to the Neuron Daily AI newsletter with a half a million other people who read it in the mornings. You should be too. I hope you enjoy it. It's really centered around trying to make the AI space digestible by normal people and help you understand that in the midst of all the crazy headlines and wild stories, like here's what you really need to know.

1:27:13Corey:here's what's happening, here's what's coming, and here's how it might affect you at home or at work. You know, our goal is to bring that fast-moving world to you at a pace you can follow and we can follow. So we appreciate your time and your brains. Farewell for now, humans. Have a great week.

1:27:31Grant:Bye-bye.

1:27:43Want to see a tank? Okay,

1:27:48let's feed all these skin Pac-8