Subscribe
Share
Share
Embed
Built This Week is a weekly podcast where real builders share what they're shipping, the AI tools they're trying, and the tech news that actually matters. Hosted by Sam and Jordan from Ryz Labs, the show offers a raw, inside look at building products in the AI era—no fluff, no performative hype, just honest takes and practical insights from the front lines.
Don't care about the jump out of the plane. Let's do it. You can simply put yourself inside your own videos. You know, here's me tanning leather. Here's me working on a oil rig.
Jordan:Check it out. I have done a ton of sort of two video.
Sam:Hey, everyone, and welcome to Built this week, the podcast where we share what we're building, how we're building it, and what it means for the world of AI and startups. I'm Sam Nadler, cofounder of Rise Labs, and I'm joined each and every week by my business partner and friend, Jordan Metzner. How's it going today, Jordan?
Jordan:Hey, Sam. Happy to be back. Another new episode. Lots going on in the AI world. So super excited to get started today.
Sam:Yeah. Of course. We'll jump into the docket. But before we do, just remember to like and subscribe. We have new episodes every Friday, and you can find them on your your favorite place to to listen to podcasts, YouTube, Spotify, Apple Podcasts.
Sam:So today, we're gonna talk about, a really cool, you know, product you built regarding parking tickets for Los Angeles, and I think you were inspired by something similar that was happening in San Francisco. And just about, I think, yesterday, sort of two came out, and we've been playing with it all day yesterday. So it'll be a fun second topic, and then, of course, some some latest and greatest AI news.
Jordan:Awesome. Yeah. Let's get into it. So, you know, just about a week or so ago, I saw someone on Twitter had posted a live map of the parking tickets being issued in the city of San Francisco. So he had gained access to the API that showed where the tickets were being issued, and he was able to show in real time where the meter maids were, essentially those who who write the tickets.
Jordan:And I found this incredibly fascinating. It it went a little bit viral on Twitter, and then really quickly the city of San Francisco tried to do whatever they could to kinda patch it up. I believe they took down the parking system multiple times in order to block his access from this real time real time data. And I think the city defended it as they were trying to create safety for their officers and that, you know, by giving this live data that it would that would be a threat or some type of danger to the officers. But the consumers or, you know, the residents of the city loved it because they saw where officers were, meaning, you know, they could probably park in an area if there was no officers nearby and then avoid a parking ticket.
Jordan:And, you know, parking tickets are one of those things that, you know, we've decided as a society that we're okay with. But, you know, everyone in society kinda hates them. And, you know, depending on where you are in the country or city, you know, you might find more aggressive or less aggressive means for attacking parking tickets. And once I saw this, I was super excited about, you know, kind of where else could I do this type of implementation. And luckily, the city of Los Angeles has a really good set of data.
Jordan:I've manipulated the data from the city before to looking at parking data at LAX and a few of these other APIs. And thanks to a company called Travis, which has the contract with the city of LA, there's pretty much access to the data through Excel, queries, as well as APIs. So I thought it'd be fun to take the LA parking data. I queried the API. I did this last week, so data might be a little bit stale.
Jordan:And essentially show a map of where parking tickets are being issued in LA. So I did that here. And here's the city of Los Angeles. And you'll notice there's some pockets kind of where there's no parking tickets, like, for example, in Beverly Hills or Santa Monica. You know, these actually happen to be autonomous cities within the city of Los Angeles, and so this actually makes a lot of sense.
Jordan:What you can see here is, like, the different officers. So, you know, if you kind of find the officer now this officer l four, let's say officer fifty three, you can see kind of, you know, he's been working the Valley area and, you know, as far north up till here, as far east and west all the way over here. But you can see plenty of pockets kind of where, you know, these officers are actually, you know, not not really penetrating. And, you know, this is Ventura Boulevard, so, you know, obviously, there's the most amount of tickets being happened on a on a Main Avenue. And we can go through kind of different officers and see, you know, where where they're giving their tickets.
Jordan:I also have like an officer leader board here, so you can kinda see which officers are driving the most amount of revenue. I've also actually added well, I mean,
Sam:you know,
Jordan:you gotta understand, these are officers writing tickets.
Sam:I'm just kidding. They're doing their jobs.
Jordan:Yeah. They're doing their job. Yeah. But, you know, obviously officer fifty six is in the lead here with 329 tickets, 55 with a 141. Again, this could be some data discrepancy just in how I brought the data in.
Jordan:But it looks like, you know, in general, officers are written, you know, I think in, like, over this time period about a 150 tickets or so per officer. And then I also have it segmented out by, you know, vehicle type vehicle color, excuse me. I have violation type. I have the different agencies, so we can see kind of, like, you know, does the downtown area versus other areas. And then I also have some date range selectors as well.
Jordan:You know, something interesting with the colors I started to notice, obviously, you know, black, white, and gray are the most ticketed cars, but they're probably also the most popular colors. And so I was trying to see, you know, are there any car colors that, like, don't get tickets very often? You know, brown, purple, and gold cars just don't get tickets very often, but they're also, you know, not very popular car colors. So, you know, I don't I don't know if the solution to avoid parking tickets is to get a brown car, but, you know, the data would lead you to believe that that would be true. So, yeah, this is called Tix Tix LAX, like I mentioned, all supported by LA parking meter data straight from the city of Los Angeles.
Jordan:And then, yeah, using Mapbox here, I coded almost all of it using cursor and codex. OpenAI is codex five. And, yeah, happy to answer any questions, Sam.
Sam:Yeah. What you mentioned that the San Francisco version was using live data. This is not live data. Correct? And what over what period of time is this data?
Sam:And could you utilize live data? So you could, you know, literally open up that application and be like, they're not in this area. I could probably get away with it for for the next hour.
Jordan:Great question. So when I looked at the dataset, first, I had some parking tickets that were actually given in the future. And so that meant to me that, you know, there were some missed dates or an officer had labeled, you know, writing written 2026 instead of 2025. So I actually found some parking tickets that were written for November, December of of this year, which hasn't happened yet. But when I looked at the data set, it was about two days delayed.
Jordan:So it looks like officers kind of write their tickets on the street, and then they upload it into the system. And so if I recall correctly, you know, I was first download this data, I think around September 24, and I think the newest tickets were from September 22. So about a two day delay in seeing the data online, which means, like, not ability to get live data. I tried very hard. When I saw that that there were the tickets kind of in the future, I thought, oh, wow.
Jordan:I mean, maybe these tickets are getting uploaded in real time. But yeah. Then I realized kind of like everything's about a about a two day delay. So not the same ability to kind of see where officers are, you know, right now or on a map or anything like that. And I think that's probably makes it, you know, a little bit safer for officers as well just if you're looking at this dataset.
Jordan:But, yeah, I think, you know, I just the city and the technology doesn't doesn't offer real time datasets.
Sam:Yeah. Well, just looking at the patterns here, I mean, looking at the valley, you could probably, you know, justify if if your goal is to avoid parking tickets, which it probably shouldn't be, but just for the for the sake of this conversation, you could probably, you know, justify parking just a couple blocks off Ventura. It looks like they pretty pretty much stick to Ventura Boulevard and, you know, avoid getting a parking ticket. But either way, I mean, it's a I think it's a really fun use case. Pretty amazing what you can do with available APIs for a bunch of different use case and and just represent data and and drive informed decision makings in a really, really quick and easy way.
Jordan:I mean, yeah, as you mentioned about Ventura Boulevard, you know, these are all expired meters. You know? So, you know, I filtered here by by meter expiration. So, you know, this is where the meters are located, but, you know, don't let your meter expire on Ventura or it looks like you're probably gonna get a ticket. You know?
Jordan:And it looks like there's a huge stretch here between Sherman Oaks and Encino where they just simply either don't have meters or they're not not highly enforced. But, you know, kinda once you pass this street right here, looks like this is kinda no man's land. So I think that's super interesting. Just like an interesting way to look at the data. You know, if I had more data, it might show some different behavior patterns.
Jordan:But I think this is just like the beginning of, you know, giving, you know, residents and citizens access to their city's data to to be able to look at it in different ways. And, you know, I think, you know, we have open records. We have Freedom of Information Act. And this is just one more way where, you know, residents can take advantage of leveraging the data in their own town and and see what the impact is. In this case, you know, parking tickets.
Jordan:And you can even see, I mean, this isn't even that much money. I mean, 20,000, 11,000, 12,000. I mean, I don't know exactly the time period this was over, but, you know, it's simply just not a significant amount of revenue for for a city as large as Los Angeles. So, you know, what is the purpose here? Is this is this even efficient?
Jordan:I believe I read a few months ago that the actual parking division was cash flow negative, meaning that they they generate more costs than issues they than tickets they generate in revenues. And so, you know, really, what are what are we trying to do here? If we just eliminated this parking department, would we end up saving money and, you know, have a better better quality of life and society, etcetera. So, yeah, hopefully, that's a a little background on my parking app, and I hope you like it.
Sam:Super cool. Yeah. Let's transition. Yesterday, sorta two came out. You and I have both been playing with it pretty much all day.
Sam:I would say it there's a a few things that make it really impressive, and then I personally, I think it trails a little bit in in to veo three. But, you know, for me, what I think is really, really impressive is the fidelity of, you know, the the face and and who's generating it. I also think, like, the social app launch included with Sora two was a really cool way to get users, you know, posting videos and sharing what they're making. But I feel, and this is, you know, one day of use, the the physics. You know, there's a couple kind of active videos I made, whether it's skateboarding or playing soccer.
Sam:Didn't you know, it was pretty wonky compared to to Veo three. But, yeah, I would love to hear your opinion, what you've built. Maybe let's build one live right now and then take it from there.
Jordan:Yeah. So I have done a ton of sorta two videos. As you see here, this is kind of the sorta two homepage on web. You know, lots of Pikachu, lots of SpongeBob, lots of Martin Luther King and John Kennedy, you know, dogs and cats driving cars. So we've seen a lot of really fun fun stuff.
Jordan:This is like security footage, you know, seen a lot of Sam Altman kind of stuff. You know, here are the things. Here's Hamilton. So it's been great overall. I think it's kind of like a really fun fun tool to play with.
Jordan:Sometimes the physics are great. I think in this case of this jumping on the trampoline, it looks pretty good. You can see the trampoline bounce. The you can see it kind of inflect and inflect back up again. But to your point, you know, sometimes the the physics have been a bit weird.
Jordan:I I was doing some clicking one and I had like three fingers or something, so it looked a bit strange. But why don't we just do a video here? You know, I have two dogs. One of them is a black lab. So, you know, let's do a black Labrador.
Jordan:And then what should we have her doing? Maybe we could have her cooking, maybe sushi in a high end restaurant. How's that sound? With a chef's hat on. Okay.
Jordan:Cool. So, you know, it does seem that the generation time is somewhere between, like, three to five minutes. Meaning that like, you know, to run this model is probably incredibly expensive for them. You know, I've generated at least like 10 or 15 plus videos so far at a cost of zero. And if you recall like when we were using Veo three, you know, the costs were like 6 to $8 a video and then driving downward.
Jordan:So, you know, I don't know how much it's costing ChatGPT to produce these videos. I presume it's incredibly expensive. You know, right now access to Sora is limited through invitation only to plus members. And so, you know, they've done some some mechanisms to restrict the amount of volume. I even was trying yesterday and kinda got denied because there wasn't enough available servers.
Jordan:So you kinda get an idea that this is just the early days, but what I think is so interesting about the Sora two is this is kinda the first time we've seen video continuity. They call it I think they call it Cameo or Avatar, but you know, essentially you you film your face, you you move your head left, right, and all around, and then you can simply put yourself inside your own videos. And, you know, I don't know if any of these will come up. Let's let's try. Hopefully, can see some of my some of my previous videos.
Jordan:But, you know, here's me tanning leather.
Jordan:Work them. The liquor keeps them supple while the sun does its job. After a good scrape Yeah. You get this even grain. And that's when they start to feel like leather.
Jordan:Hang them up. Let the breeze move through.
Sam:And the oil refinery one, I think captured you quite well, if if that's available.
Jordan:Here's me working on the oil rig. It's eating everything out here. Come on. That's it. Bite down.
Jordan:One more turn. There she goes. Let's get another
Sam:I mean, it's really you know, there's some, you know, small, I would say,
Jordan:obvious Inconsistencies.
Sam:Yeah. Imperfections. Yeah. But it's really, really good in terms of, I think, the facial fidelity of of what you've upload. I mean, that looks just like you.
Jordan:Yeah. I've never lived in a bowl in my lunch.
Sam:Well, you're doing a pretty good job there. Gates coming open. What do you think compared to v
Jordan:o three? Well, I think that's the big difference here is the ability to use your face, you know, inside this content. And I don't know if you've seen these, but, you know, here I am on an elephant. Here I am, you know, as a Buddhist monk, scuba diving with sharks, fighting in a hockey game, working in a Tesla factory, barbecuing, farming. Here I am on the top of Mount Everest playing football, making pizza.
Jordan:So you can look, sushi, fishing, skydiving, formula one. I mean, here's my Formula One car. That's it. We did it. The car was unreal today.
Sam:So Amazing.
Jordan:You know, it's just Amazing. It's been incredible. This is one of my favorites. You know, it's it's not so active, but it just looks so real. Just looks so good.
Sam:And, of course, you know, you get someone creative, give them these tools, and, you know, really, the sky's the limit. But talk talk me through really quick. It's it's hard it's easy once you have the app to make videos of yourself. It you can only make videos of other people if you've, a, granted access for them to use your face and your friends with them on the social app. So, like, I've tried to make a video of you and I.
Sam:It came out okay. But there are some limitations of who you can make videos of. That's why there's so many videos of Sam Altman. But who you can make videos of if you don't know them and they're not connected to you.
Jordan:Yeah. So, you know, it is a new form of social network. Think, you know, as we as you guys are aware, Facebook tried to launch something pretty similar about a week ago just to kind of seems like probably the preempt OpenAI here. But, you know, here's me skydiving. Look how realistic it is.
Jordan:I've never skydived in my life. So, you know, again, here's me making coffee in a in a home in Seattle. This is all fabricated. Welcome to Seattle. Does anyone want some coffee?
Sam:Yeah. Amazing. You know, what's interesting is I think yesterday you were you filmed your face for the app, you were wearing a black hoodie. And I think it in a lot of these videos, it captures you with this black hoodie.
Jordan:Yes. But, you know, it does as you see kind of here in this Tesla video, but, you know, or in this cooking video. But here, I'm a farmer, and it just, you know, it gave me a
Sam:Yeah.
Jordan:With a plano. New outfit. Look at my fingers there. See no good.
Sam:Okay. Let's let's move on. Any any other kind of closing thoughts on
Jordan:this sort of thing? Here's the Labrador making sushi.
Sam:Oh, it's cute. I'm sure he'd chomp it up. That's hilarious. Alright. Well, yeah.
Sam:So final thoughts, and then let's let's get into some news, which I think, you know
Jordan:Yeah. I mean, I think overall, my final thoughts on Sora two is, obviously, it's incredibly fun. Every time one of these new models comes out, whether it's been like images or video, you know, it's really fun to play with it for a few days. I think already I've noticed, you know, the videos are interesting because you're the star of it and less interesting when you're not the star of it. But, you know, the the ability to have your imagination run wild of, you know, putting you inside a shark tank with, you know, inside an ocean with sharks or, you know, you know, on top of an elephant or in a farm or in a professional hockey game or, you know, basketball game, fishing, skydiving.
Jordan:I mean, here I am here I am dunking in a basketball game. So, yeah, I just thought, you know, it's an incredible way to to really drive up creativity and and put yourself in positions that you wouldn't you know, in activities that you wouldn't normally doing. And, you know, I think this is the beginning of, you know, the ability to create con con continuous content kind of using the same character throughout and really tell a story. And, you know, while these videos are only eight, nine seconds, I think, you know, very quickly, we'll start to see twenty, thirty, forty seconds videos. And even in these videos, see they have multiple cuts.
Jordan:They're very creative. And, yeah, it just seems like the really early days. I know NVIDIA's launching a new video generation chip that doesn't even come out until about June. So I know this is just the early days. We find some areas here where physics or continuity or the image kind of drops, and I think we'll we'll see those things start to go away.
Jordan:And then, you know, these videos will become, you know, hyper realistic to the point that you you won't be able to tell what, you know, what's real and what's been generated.
Sam:Super cool. Alright. What's hot in the news, Jordan?
Jordan:Yeah. So, I mean, going off the same OpenAI news, they did a secondary, I think, yesterday at a $500,000,000,000 valuation. I think they I think they allowed employees to buy up to six points or sell $6,600,000,000 worth of shares, which is just a massive number. $500,000,000,000 valuation is obviously an incredible number. This says above SpaceX, so probably the most valuable private company on the market.
Jordan:And I think this just speaks to, you know, how impactful ChatGPT and OpenAI has been on the AI ecosystem.
Sam:Yeah. Absolutely. I mean, you know, not only the one of the the highest valuation for a private company. I mean, I you know, there's not that many public companies at at greater valuations. Yeah.
Sam:I mean, congrats to the the employees that were able to participate in the secondary. And, I mean, it just you know, I I think all of us, whether they're just gonna be a huge contender in the space. And, you know, it's hard for me now to say if that valuation makes sense, but, you know, it seems like it's only gone up rapidly since they since they've really become part of the public domain.
Jordan:Yeah. Well, let's take a step back for a second. You know? If you look at these LLMs and and these these laboratories, you've kinda got three three different verticals. You've got chatbot, you've got coding bot, and now you've got kind of content creation, kind of this image and video generation.
Jordan:And so, you know, a few weeks ago, you know, Anthropic was kind of number one in in the coding bot space. Google was number one in the, you know, image and video generation space. And OpenAI, you know, probably had the best chat bot. I would say now, you know, since GPT five has come out, they have taken the lead on the chat bot for sure. They've taken the lead on the coding bot for sure.
Jordan:Codex five is for sure then the number one coding tool in the market right now. And now they've just taken out the lead in video generation. Now, I know Anthropic came out with a new model this week. I haven't been able to play with that enough, but I still believe kind Codex five is is the market leader. And so you see OpenAI kind of taking the reign back across these three verticals.
Jordan:And, you know, we'll see more verticals as well. Obviously, as you know, there's audio and a bunch of other AI verticals in which OpenAI does not excel in. But I think in these three kind of chat, code, and content generation
Sam:The biggest ones.
Jordan:They're they're in the number one spot right now across the board.
Sam:Yeah. It really does seem like they're firing on all cylinders. Alright. What else?
Jordan:Okay. Cool. Yeah. So just a few other small things. I read today that Apple decided to deprioritize a smaller Apple Vision Pro in favor of AI glasses like Meta.
Jordan:And I think we talked about the Meta glasses just a few weeks ago. I have not tried the Vision Pro. I think, like, you know, overall people have had pretty bad reviews of it. But it does seem like kind of these these glasses are are coming to be the new the new form factor coming through with the Meta Glasses. And then, you know, if Apple 's able to launch some glasses as well as, you know, OpenAI obviously acquired Johnny Ives company, and maybe they're working on some glasses as well.
Jordan:So I just think this is kind of like the convergence of hardware and software to the point where you have kind of AI, you know, on you at all times. And I think, you know, you see all these major labs kinda competing for what does that mean to be on you at all times. You know, is that the phone? Is that glasses? Is that a pen?
Jordan:You know, what does that look like?
Sam:Yeah. I'm actually really excited about this. I know I know Vision Pro is a a bit of a flop, but I think, you know, Apple's so good at hardware, whereas Facebook historically, you know, is obviously more software oriented, even though they have the Oculus. But, yeah, I I think this is gonna be really interesting. I think it comes out in 2027, so, you know, about a year and a half away.
Sam:And let's let's see, you know, what they create. But, yeah, I was excited to see this. I I've never used the Vision Pro outside of the Apple Store. You know, it seems like a little bit too big, too much. So let's see maybe what a little refined vision feels like with, you know, obviously powered by AI and augmented reality and all the other features.
Jordan:Yeah. And I know, you know, the market's been down on Apple because they've done a poor job of adopting, you know, AI into the ecosystem. But I think, you know, as you and I can attest to, seems like the Apple 17 phones are sold out pretty much across the country for days and days. I mean, every time a store gets a new stock, they're sold out. He had a friend trying to obtain one, it was almost impossible to to get it at any store.
Jordan:So, you know, even though we're seeing kind of this, you know, Apple Apple not so great at AI stuff, it sure seems like consumers, have a insatiable demand about the devices.
Sam:Alright. Well, anything else or time to wrap up?
Jordan:One last story here. I thought this was really interesting. This guy's name is, Steve Simone, and he sold his last company to to DoorDash, and he's now working on an AI robot that shoots down drones. And I thought this was pretty cool simply just because kinda if he changed the space and, you know, this is kind of a new new type of investment using AI for for for security and defense. As you know, Israel has had its Iron Dome system for years, but we're starting to see more and more companies kind of build out US defense technology.
Jordan:And I thought this was a pretty cool one as well. So he's building these like automatic devices that shoot down drones in the case of some military issues or things like that. So, anyway, I just wanted to see this is what it looks like.
Sam:The use of drone warfare has, like, rapidly increased with, you know, the conflict in Russia and Ukraine. I'm sure it's being used in other places. But how does it shoot it down? Is it, like, bullets, or is it some sort of net or lasers? Or had I I haven't seen this article yet.
Sam:So how does it work?
Jordan:I know it's like this turret that I don't know if it's using bullets or what it uses, but it's using this turret that sits like you install it on the back of a flatbed truck, and then this thing kind of senses for them and shoots them down. But I you know, yeah, like you said, you know, we're moving into a warfare state where, you know, these wars are being fought by drones. And so, you know, shooting drones down seems like a defensive mechanism. But yeah. Anyway, I just thought it was pretty interesting how AI is, you know, entering the the defense space, and this is just a good proof of that.
Sam:Well, thanks, Jordan. Entertaining as always. And don't forget to like and subscribe to Built This Week. New episodes every Friday. And any closing thoughts?
Jordan:Yeah. Awesome episode. Again, you know, we're we're rolling into the fall season. And, you know, AI is still heating up, I think. You know, we're seeing NVIDIA continue to hit market highs.
Jordan:Stock market is still ripping. We have government shutdown this week, yet, you know, the market is still hitting highs. So it still seems like the early days for AI. Regardless of what happens in, like, the overall economy, it seems like this AI train is is gonna continue to chug along and, in fact, you know, be the the driver that probably moves moves the market forward for sure. But, yeah, this was a great episode.
Jordan:Really fun to talk about some of these toys. Really having a lot of fun with Sora. We have put together a little Sora best of clip from this episode. So hopefully, can chop that up and play that in here somewhere. But thanks everyone for listening.
Jordan:Like Sam said, like and subscribe on YouTube and all your other favorite podcast platforms. For Sam, I'm Jordan. So much and see you next week.
Sam:Thanks everyone.
Jordan:Bye everyone.