043 The AI Hype Phase Is Over

[00:00:01] Speaker A: Awesome. [00:00:02] Speaker B: We are streaming life. [00:00:05] Speaker A: All right, well, welcome back to the Gregory and Paul show. I'm Gregory. [00:00:12] Speaker B: I'm Paul. [00:00:14] Speaker A: And we break down latest in SaaS, startups, AI. We always touch on a few memes. We've got a great show for you today. On our list of items to discuss, we want to kick it off with the Lovable versus Vercel hacks and responses. They both had very different response every single week to the. To the hacks that they were involved. They were involved in. Yeah, I know you're passionate about that one. [00:00:46] Speaker B: I. I'm very passionate about this one. I gotta give Versa credit. [00:00:49] Speaker A: Like, I. I think they did a good job. [00:00:51] Speaker B: They did a great job. Yeah. [00:00:53] Speaker A: All right, all right. Other topics. We'll talk about GitHub, pausing, copilot signups. Interesting. We'll cover on a Morgan Stanley report that I dug up about AI's impact on business. [00:01:06] Speaker B: That's a good one. [00:01:07] Speaker A: We want to. Yeah, it's really interesting. I want to discuss ChatGPT launching performance ads. We'll talk about SpaceX cursor, Mercer cursor. Cursor merger. Cursor merger. It's like a tongue twister, which you probably all heard something about. We'll discuss some things happening over at Amazon when it comes to vibe coding and tool consolidation. Vibe coding's impact on the Apple App Store. And we'll finish off with a segment on live TV wars with the launch of Monitoring the situation. Yeah, Paul, you want to kick us off? [00:01:49] Speaker B: Yeah. [00:01:50] Speaker A: Lovable versus Vercel. Yeah. [00:01:53] Speaker B: It seems like every single week there's. There's a hack and not a small one. So this one involved a Versal. They're one of their engineers installed. [00:02:04] Speaker A: Is it Versal or Vercel? [00:02:06] Speaker B: Versal. Vercel. I don't know. [00:02:08] Speaker A: Is it Versal? [00:02:08] Speaker B: It's French for something. [00:02:09] Speaker A: I mean, I'm into it. I'll go with Versal. [00:02:11] Speaker B: Versal. [00:02:12] Speaker A: I was saying Vercel. I'll go with Vercel. I don't know. I've never heard them say it. [00:02:16] Speaker B: Yeah, I don't know. It's a French word, I think. Okay. So they basically the context is that their engineer, one of their engineers, installed this one off tool called Context AI and apparently Contacts AI got mixed up with some hacking situation and opened up a attack vector to rest averse system. Seems like their public disclosure is that it's limited. The hack is Surface is limited. However, the attacker did access a ton of their environment variables, which are secrets these companies keep. Their compare their hack to Lovable. This is, this is going to be a good one. So exactly. Same time Lovable, somebody came out and essentially showed and demoed how some of the Lovable's older projects can be accessed publicly. All of their source code, all of the chat logs that people use to generate their lovable projects were completely publicly accessible. And essentially their public communication that came out was, you did it wrong. Essentially they said yes. We were doing tests, testing how the concept of public projects meant. As in everything about a public project means that the source code were public, all of your chats were public. They've since deprecated that feature, however. They just, you know, left it. [00:04:05] Speaker A: Yeah, so this was like one of those were like, what? Oh dude, it's crazy. [00:04:11] Speaker B: In what world is this good communication? [00:04:15] Speaker A: So the loveable one. So Vercel did a good job, right? [00:04:20] Speaker B: Vercel did a great job. [00:04:21] Speaker A: Got up there, he explained it very simple. It didn't even like, it didn't even, no one bat an eyelash. Look, things happen, they get fixed. The lovable one was like, well like it wasn't, it wasn't a hack, right. It was more like a bug. They rolled out some code that changed permissions on how certain things. [00:04:45] Speaker B: Yeah, it was a, it was an old feature. It was a legacy feature that allowed people to get a peek behind your. Your look. [00:04:53] Speaker A: If you want to know the history of every product release they ever made. Like, they did a great job at detailing that out before getting to the fifth paragraph where they said, hey, this changed the permissions. And that's what happened. I was like, oh my God, just start with that. Just tell everybody like, hey, sorry, mistake, right? Like it. Oh my. I'm reading this thing and I'm just like, like what? Like it was not a great way to handle something like that. They seemed really intent on proving to everyone that like there was like a long history of reasons why they approached it the way that they did. Which like people don't care. People just want to understand like why this happened and what they're going to do to fix it. That's it. Simple. Yeah, just acknowledge it. [00:05:36] Speaker B: So now the question is, you know, who, who's responsible for these kind of screw up? Is it the developer? Is it the organization? Is it, oh, it doesn't matter. [00:05:48] Speaker A: It doesn't matter. Like everyone's too worried about like blaming people. Just like fix it, move on. I wouldn't worry about it. They're new companies, they have new pro, new apps that are like they're changing the products all the time. Like that's just the way it is, I think. [00:06:00] Speaker B: But it does kind of like destroy the confidence a little bit on all of these Vibe coded products. Right. Love a Boy is kind of the, the center of this, you know, vibe coding platform. And even if they can't. [00:06:18] Speaker A: Great question. [00:06:20] Speaker B: If they can't secure or if they can't communicate how they're securing their users products, it really doesn't add any confidence into this whole, whole platform. [00:06:33] Speaker A: Like we're not gonna, we're not, we're not gonna like migrate over air traffic control to lovable. [00:06:40] Speaker B: Right. [00:06:40] Speaker A: Like it's. Right. It's just, it's really obvious like from my world. Like you have to have the right expectations about the tools that you're using. Like it's not an enterprise grade tool. So like of course there's going to be issues with it. Like what are people doing like Vibe coding internal applications to manage dashboards for reporting? No, it's fine. [00:07:03] Speaker B: Yeah, maybe. But if that's actually a little bit even more dangerous, right. Because you're potentially revealing proprietary data points [00:07:14] Speaker A: on your, the performance of your Facebook ads like for your, you know, agency, you know. Yeah. For your, your protein bar or whatever you're selling. Like, give me a break. I don't think it's a big deal. [00:07:27] Speaker B: Yeah. So we'll, we'll see. Well, I think the, there's an opportunity, like we said quite a few times for I think, I think the security [00:07:36] Speaker A: dude, the point here is the how they handled it and Versal Vercel. Someday I'll figure out the right way to say this. They did a very good job seeing it out there. He said, he's like, look, there's some issues, it's fixed, end of story, just move on. Do you have any tips back and forth? [00:07:55] Speaker B: Yeah, well, tips give like these companies who may or may not go for it running to a similar situation. What should they do? Top three things. [00:08:06] Speaker A: Yeah, it's okay. In any of these situations you need to do a couple things. The first thing when you hear about any, any crisis, actually not just like a hack, any crisis happens, first thing is to be very public that you know, that's it. Like right, I know about this. CEO posts on Twitter, posts on LinkedIn, sends out a press release, smoke signals, however you handle it, letter to your customers, email, whatever needs to happen, depending on the severity of what's happening. We know, we're looking into it. That's it. Right. Then you do an investigation and you make a determination about what has to happen next. Don't just like jump out there and start saying things and accusing people or like replying to stuff. Like, that's how you like, really get yourself into hot water. Right. So acknowledge it. You understand what's going on. You're looking into it. Look into it, and then formulate the right approach to respond as quickly as possible. One hour, a few hours. If it takes a day, that might be appropriate. If it takes more time, because it's like really complicated to look into, post an update. Like, we're looking into it. So we don't fully understand this before we like, make accusations or accuse anybody or like make change. Like, yeah, you want to be very careful with the low one. It's a good example where they should have just like internally, what was written on Twitter was probably the conversation that should happen. The product team's like, we made this change, made that change, blah, blah. And if I was like comms person or CEO, be like, okay, great. Like, now I understand what happened, but that's not what you say publicly. [00:09:42] Speaker B: That's not what you said. Right? [00:09:43] Speaker A: Correct. That should just be the internal conversation. I've been very quickly on. On Slack is what I think is how that should have happened. And then the CEO and the comms person should just been like, look, thank you for helping us understand what happened. We'll take it from here. And then what they should have just said is like, hey, we understand that there was like a change. I think it just should be super clear about what happened. Like, there was some code that was released that changed some permissions that exposed certain types of data. People were confused by it, didn't understand what's happened. We understand that this is an issue. We have fixed it. We've already fixed it. Right. I wouldn't even like, just go there and fix it and then just be like, it's been fixed. If you have more questions, like, ask. That's it. That's all that needs to be said. [00:10:32] Speaker B: We've reached out to affected customers. We've offered some kind of compensation. [00:10:38] Speaker A: Right. I think in this case there was a whole bunch of back and forth with like, like, let's call him security researcher or whatever, who he claims that like. Or she. I don't know. They. They notified them. They didn't hear back. There was some kind of history of that. So if that's true, if I was lovable, I would have just said, look, there's been someone very helpful in the community who pointed all this out. [00:11:01] Speaker B: Yeah. [00:11:01] Speaker A: And we'll be working with them to just make sure Everything is safe and secure. That's exactly what I would have done. And I literally would have called that person and said, tell me exactly what happened and explain to me like what you think should be done. And I'd listen to everything they had to say and then I would go to the team and we'd make a determination whether we implemented that or not. But, like, we would definitely get feedback from somebody like that and make them into an advocate, maybe, I don't know, give them a free account, maybe give them some discount codes, they can go and post them. I mean, there's that. You could totally turn that person into like a true advocate for the platform because they obviously care and they're just trying to like make it better. So I think that's another big thing that was missed in this one. [00:11:46] Speaker B: I completely agree. [00:11:48] Speaker A: Yeah. [00:11:49] Speaker B: Shall we move on to the Next? [00:11:52] Speaker A: Yeah, yeah. GitHub, this one's interesting. You're the one that put this in the doc, so tell us what's going on. [00:11:56] Speaker B: Yeah. Okay, so GitHub paused copilot signup, which is, I think a huge deal because that's signaling Microsoft is running out of computer, which is even more. [00:12:10] Speaker A: Do we know for sure? Do we know for sure it's because they routed compute. [00:12:14] Speaker B: What AI company is going to come out and say that's the case? I don't think anybody will ever admit this is case. We can feel it. Right. Like Anthropic is throttling, removing cloud code access from the pro plan. You have to upgrade to max now. They've been throttling a lot of access. GROK has been throttling access unless you upgrade to super Grok. So I definitely do think all of these LLM companies are running out of compute and that's. That could potentially signal the top. Top of. [00:12:51] Speaker A: They didn't. They didn't. They didn't say specifically. They just said, hey, we're pausing this. [00:12:57] Speaker B: Correct, Correct. [00:12:58] Speaker A: Okay. And we assume it's because the rang out Q. Which is quite possibly true. The reason. The reason that like. Well, the one data point I have is that that was the reason given for why SORA was sunsetted, of course, by OpenAI, was that they needed the compute. They literally didn't have the capacity to keep it going to do all the other things that they wanted to do. [00:13:22] Speaker B: Correct, Correct. So SORA was absolutely the first that got canned. One of many a. And now it seems like it's propagating throughout the industry. And so, yeah, AI companies are focusing more on enterprise usage and not the Individuals. Right. So like I said last week, I think vibe coding is coming to very, very near end. Individual. [00:13:52] Speaker A: We're at peak. Peak vibes. [00:13:53] Speaker B: Yeah. Because individual access will go away. [00:13:58] Speaker A: Peak vibes. [00:13:58] Speaker B: Peak vibes. [00:14:00] Speaker A: I don't know. I don't know. I don't know. I think, I mean, they'll get more compute and more capacities to come online. I guess that would be my answer. [00:14:07] Speaker B: It's physical, it's limited by supply chain. We're literally running out of manufacturing abilities, capabilities like, like Nvidia TPUs back ordered for two years essentially. So we have, we're heading out. [00:14:27] Speaker A: Yeah. I suppose like there, there are actual limits, like how much we can make. [00:14:31] Speaker B: Yeah. So I, I know you're pretty bullish on intel, which is a potential company that can pump up the capacity if they get their stuff. [00:14:42] Speaker A: There's a lot of, I mean it's, it's like this is the real world stuff where you're moving atoms around. Like they have to actually build facilities, get those facilities working. I know they've been building them, but. [00:14:51] Speaker B: Yep. [00:14:52] Speaker A: I don't know what the status is actually. [00:14:53] Speaker B: Yeah, we need, we essentially, we need to build the actual chips and then we need to create the data centers. We need to buy the land, build the data centers, supply it with power. That's a lot of moving pieces. Building power plants is slow. So what's going to be interesting to see how much the pricing per token goes up and who gets allocated access. That is going to be interesting. [00:15:24] Speaker A: I mean, I think it's pretty obvious that tokens are expensive. [00:15:29] Speaker B: Yeah. [00:15:30] Speaker A: And they won't necessarily be cheaper than people. That's what I think at high level, I mean, I think there's more efficiency. So I had this personal experience, I was like, okay, let me set up cloudbot. So look into this. And I'm like, I'm talking to people and they're telling me about like how much money they're spending on tokens. And I was like, I'll just go and hire like an upworker. And that is what I did. And it, from what I can tell, way more affordable. And I was like, why would I like do this at claudebot and then be responsible for like all of the support and dealing with issues and upgrades and all this and then paying for tokens. [00:16:06] Speaker B: Yeah. [00:16:07] Speaker A: When I could just get someone to do the job. And I was like, okay. So I, and I imagine this happening all over. Didn't the CTO of Uber acknowledge this? [00:16:16] Speaker B: Yeah, yeah. Lots of. So, okay, so that you basically raised a really interesting question, right? Like, I think this kind of flows into our next topic, which is. So do you think AI is making everything more effective? And the next logical question is, how do you even measure the outcome? How do you even measure. What does all of this efficiency really mean? [00:16:46] Speaker A: Yeah, it's a really good question, right? [00:16:48] Speaker B: So. [00:16:48] Speaker A: Because. So. So the next topic that I had in the list was a report from Morgan Stanley on AI efficiency and on the impact on jobs. [00:16:57] Speaker B: Right. [00:16:58] Speaker A: And so in this report, they actually led up the jobs piece. And it's also important to understand the report was done. It's a survey. So I, I guess they just asked executives, like, it's not quantitative data or some. It's quantitative, I suppose, but like, they. It's a survey. And so they did a survey of. It's like 900. It's a big number of executives around the world, primarily in English, you know, post colonial Anglo capital markets. Right. Kind of United States, Australia, uk. I think they had the EU as well. But, like, some of the findings were astounding. So in the jobs part, they found that, like, there was some, like, net job loss, but they were doing it based on, like, number of jobs created, number of jobs lost, like, in aggregate in, like, a country like Australia. So it doesn't necessarily prove that's AI, it just proves that, like, jobs are being lost. Which I thought was interesting because, like, I think we're in a time right now where things are just getting more efficient. I don't think all of it's because of AI, but this analysis and the way that they had done it shows that the US is actually net job growth, which was fascinating. They attributed, like, 4% decline in jobs, like, globally. But the U.S. was net 2%, which is fascinating. Which, like I said, I don't know. I don't know if, like, AI is really killing jobs. I think there are some areas where AI is definitely making that approach obsolete, but it's definitely creating jobs. And that was what the report showed in the United States, which is. I attribute it ultimately to, like, a more dynamic labor market, I think, like Japan, and is a good example where they don't fire people, people don't switch jobs. Like, it's just not a dynamic labor market. We're pretty good at retraining people relative to other countries. I know there might be some people that quibble with this, but. But relative to other markets, it's pretty dynamic here. Like, we're good at retraining, rebuilding, finding new ways for people to be employed. We're actually quite Good at it in the United States. [00:19:04] Speaker B: So do you think. [00:19:05] Speaker A: So there's a job, job loss piece. [00:19:06] Speaker B: Okay, so do you think the wider measurement of success of AI implementation will be job gains? Is it like employment rate or what is that? Or could it just be like earnings? [00:19:22] Speaker A: Yeah, yeah. So, so, okay, so the report also covered efficiency, covered job losses. [00:19:27] Speaker B: Okay. [00:19:27] Speaker A: Which I think is interesting. I don't think it's going to lead to massive. I, I don't think it'll lead to the massive job losses that people like Dario from anthropic claim is going to have. Should have. [00:19:39] Speaker B: Yeah. That guy is two years ago staying random at this point. [00:19:43] Speaker A: At this point you could pick any point in time that he identified this and it. We've passed that moment. [00:19:47] Speaker B: Moment. [00:19:48] Speaker A: So I don't think this is coming. I think it's pretty obvious. [00:19:50] Speaker B: I would take a discount anything that guy says about job loss, AI causing job loss. I think he's just, yeah, he has his agenda, he's just pushing it. [00:20:00] Speaker A: Of course, of course. Okay. Okay. On the efficiency side, the findings were, I thought, astounding. It found an average 11 1/2% in efficiency and productivity gains. Some companies reported over 20% productivity gains. [00:20:17] Speaker B: But this is self reported, correct? [00:20:19] Speaker A: I mean, yes. [00:20:20] Speaker B: Okay. [00:20:21] Speaker A: Correct. Like, like I don't know if it's perfect data but on the other hand you and I believe we're experiencing in our own personal businesses and jobs. [00:20:32] Speaker B: For sure. [00:20:32] Speaker A: You can see from AI for sure. I've always said that like the efficiency I see directly in the startup world is that 15 years ago it took 10 to 20 people to create what now I meet a team of three people can do or team with one person can do. Yeah, you needed like backend services and so, so I would say like I don't know if it's all AI. I think AI has become synonymous with automation. I think there's a lot of ways to automate things. There's a lot of really robust automation tools in a lot of areas that people are able to take advantage of it and that's, that is what's driving the efficiency. Sometimes maybe it's AI, but I think I would even put a number on it. Like 70% of it is automation. It's not AI. I think 30% of it is AI, but 70% is some kind of automation which isn't AI. [00:21:19] Speaker B: I see where you're going with. [00:21:20] Speaker A: Right. [00:21:20] Speaker B: So we're just automating more things. Things that used to take a person eight hours a day. It's now completely automated. Let's say it's just like a simple workflow with the right. Or at least like the upper level management is now exploring instead of hiring something new. [00:21:37] Speaker A: Yeah, yeah. It opened the door to like people being open to automating things that I agree with. All kinds of. Yeah, there's all kinds of automation happening. Some of it's AI, some of it's not AI. [00:21:49] Speaker B: No. [00:21:50] Speaker A: I don't know how big of an impact the AI portion is having. I think people attribute all of it to AI, which it's definitely. [00:21:56] Speaker B: Which is not true. Correct? Yeah, it's like. [00:22:00] Speaker A: Yeah, yeah, hold on. The numbers were big. 20 is a huge number. [00:22:06] Speaker B: 20 is a huge number. Do you think that's going to affect stock market valuation in Is it currently affecting the stock market? Is it also going to project forward in time? Okay, okay, okay. [00:22:23] Speaker A: I got a very complicated answer. The answer is that first of all the report needs to be true. Is the report true? I believe some aspect of is true. I think there are definitely productivity gains happening. How much a 20 or 30 or 10, like I don't really know. But there are people experiencing people. There are companies experiencing efficiency gains from a slew of automation tools. Some of them are AI powered, some aren't. But it's become very popular, very acceptable to automate. That's having a real impact. So if we agree on that point, then the obvious like way that this impacts businesses and impacts stocks is what some pundits and stock pickers that I follow have been saying is that it only needs to impact a small series of stocks to actually make a huge impact in business. And impact, let's say the s and P500. So if a couple of medical device companies or a couple of healthcare companies figure out how to use AI to just become a little bit more efficient, it could have a giant impact. Like a company like United Healthcare is like Fortune 2 or 3. Like they're one of the largest companies in the world. So if they can get like a 2%, 3% increase from AI, massive impact on earnings. So if this report is close to being true, like I expect the stock market to go the moon, I think that earnings could just go bonkers and that the Fortune, the s and P500, so the 500 best companies in the world all should experience some efficiency gain. Some will experience more than others. Right? Like I don't think a mining company is going to get a lot of efficiency from AI. But healthcare, obviously, telecom, tech, there are some areas that are really obvious and will experience massive gains in efficiency from all the stuff that's happening. [00:24:19] Speaker B: Yeah. Do you think it's already priced in? Stock marketer? [00:24:24] Speaker A: That's a great question. That's a great question. No one knows if it's priced in or not. I think you can make an argument in either direction. There's some people that think like, say that like, hey, the next 18 to 24 months actually be like really good for stocks. I'm in that camp. I, I, I find that to be really compelling. I think all the efficiency gains people are getting from AI are perhaps just starting to show up in business. May maybe that's the way to like try this all together. The narrative that I would spin is that this report shows that like we're starting to see the gains now in actual businesses. So for the last two and a half years, three years, there was that famous back of the envelope math I think Sequoia did that's like we're spending all this money on data centers, like where's the roi? The ROI is going to be like really difficult to achieve spending so much now. Now it looks like that actually might all come true to the upside that the massive investment, all its infrastructure was valuable. If we can grow earnings across the SB100 by a few percentage points, 10% giant, like it could be like the biggest boom ever. And I don't like to hyperbolize like this in general but like that's what I saw in this report. I'm shocked that like not a lot of people have promoted or push this report out there. I think the media and just everyone right now is really latched onto this super negative narrative around with a job loss. [00:25:53] Speaker B: AI Right, right. [00:25:56] Speaker A: It's just everyone just, it's always the sky is falling. Right. It's climate change. When I was a kid it was like aids. There's always something out there that everyone focuses on and it's the end of the world. And so far the world's still here, hasn't ended yet. Still cold and rains in Seattle too. I know there's some climate change happening. I would love it to get a little bit warmer. So I just think people latch on things right now. It's the doomer narrative that like AI is going to take all the jobs and there's people out there telling the story and like benefiting from this narrative. But actually the reality points to a very different narrative altogether. [00:26:32] Speaker B: Actually yeah, I'm a little bit more skeptical. I just want to see concrete numbers, I think. So to your point, these things do take a while to show opient earning reports. But let's say for example Google, right, they invested heavily into data centers like you said. Now they're super well positioned to write this AI wave because they are one of the companies that actually has the compute versus having to collaborate with other compute providers. I think if anyone can benefit from all of this it will be Google and I'll be looking at their any reporting. 8 month. [00:27:12] Speaker A: Yeah, I look I still, I'm still bullish on Google. I've said Google is going to I think win for a while now. I think when the, yeah, when the tide had turned and people were, people were super negative on Google is when I was like look, Google's going to ultimately run the tables on everybody. I think they've, they've proven, they've proven that which is like actually this is a great segue into our next topic. [00:27:36] Speaker B: What's. Let's talk about this. [00:27:39] Speaker A: Performance advertising launches on ChatGPT. So speaking of the battle between Google and OpenAI which I think it's really down to those two when it comes to consumer AI, Anthropic is obviously the emerging winner in the enterprise side. ChatGPT still could, could do it. OpenAI could still do it. So I wouldn't count them out. And Google isn't really powerful player in that space as well but Anthropics really run the table. But when it comes to consumer, I think Google and OpenAI are the two. It's a two horse race. [00:28:15] Speaker B: Yeah, I think you're right. Gemini, Gemini versus ChatGPT for sure. [00:28:19] Speaker A: Yeah. And so the launch of performance advertising. So they launched, they launched a, I guess a brand advertising product. It was cpm. It was sold just in volume. You couldn't really buy it on how many times people clicked. Some people had some success with it, some people were critical of it. They rolled it out pretty quickly. I guess I was critical of it to some degree but I showed the promise of what they could do and I was just I guess of the camp ultimate that needed more features, needed more technology, they needed to do a lot more with it. I think they also needed to be on the performance side which they have finally done. I think my. Now that I remember, like so what I said when it first launched was like when you look at the success that Facebook had, the mobile side, the performance piece is what really brought it home for them and that where they really won. I believe they tried some brand stuff. They weren't just, it just, it just doesn't work as well. I think it's the best way to frame it like in digital, the way it works and like television, ctv, those things brand work. Brand advertising works really great. And so now that ChatGPT is launched, CPC advertising, I think it's a, I think it's a game changer. I think it's going to be really big for marketing in general. [00:29:32] Speaker B: What do you, how do they define a click? Is it linking out to websites or is it like spinning off of a conversation? Did you look at. [00:29:43] Speaker A: It's a great question, my honor. My understanding of it is like it's the click that goes to, to take the action that the ad buyer wants the user to take. So it'd be like a click to somewhere else. [00:30:01] Speaker B: Okay. So it will take people off of the conversation. [00:30:06] Speaker A: Yeah, I think they have products that, that have a experience that are sometimes embedded. But like it's like Google where you click and you go to like your landing, your product page, whatever it is that you want them to do. [00:30:17] Speaker B: Interesting. [00:30:17] Speaker A: Charging for that. Yeah. The intent's really different. It's got a lot of differences from how Google search works. Like I don't know if it'll prove more effective than Google search. Like the original Google search when it first launched, you gotta remember it was a long time ago. Users were really different. The Internet was really different. It was much simpler. So, so it's interesting how this is going to play out, but I think it'll ultimately be very successful. I think performance will work really well. [00:30:43] Speaker B: Okay. [00:30:44] Speaker A: And I think people need another channel like and it's very expensive now. There's been a lot of Twitter debates about advertising. You had Bill Gurley and a lot of A16 guys, A16Z guys come out against advertising. That was crazy. A couple weekends ago. So. [00:31:00] Speaker B: Right, yeah. [00:31:02] Speaker A: Remember that. [00:31:02] Speaker B: Yeah, do. So do you think their best customers right now are probably E commerce? I think so. Right. Like it's going to be targeting like low level offers, bulk pricing offers versus B2B. [00:31:14] Speaker A: I don't know. So on Google the largest category is travel. [00:31:21] Speaker B: Right. Right. [00:31:23] Speaker A: But have you like spending? [00:31:26] Speaker B: Okay, so okay, so that raises a really good point. Okay. Have, have you used chat, GPT or any of these AI to like look for hotels or book. Book flights? I, I've tried multiple times. I really wanted to be successful. It just, just doesn't do a good job. [00:31:47] Speaker A: Never use it for it, does it? It doesn't even occur to me actually. [00:31:52] Speaker B: Right. And it like at the very start of all of these things, like I think maybe Perplexity was the one that did a demo. Say hey, now you can use this AI to manage all of your travel. You can book hotels, you can get it to research restaurants to. I've tried it multiple times because it doesn't work well. It does not work at all. It doesn't work. The data is old. It misses information. It can't get location correct. It thinks two cities, that's like four hours drive away are the same city. [00:32:29] Speaker A: Yeah, that's funny. [00:32:31] Speaker B: So, like, what if you, if that's the case, what does this thing do? Like, how. What advantage does ChatGPT have over Google? [00:32:43] Speaker A: All right, all right. So if I was, if I worked at chat too, if I, if I was the product manager, I'd be like, well, Paul, it's gonna get better. [00:32:49] Speaker B: It's gonna get better. [00:32:51] Speaker A: It's gonna get better. Like, I would acknowledge his problems. It's gonna get better. [00:32:54] Speaker B: I don't. [00:32:56] Speaker A: Correctly. [00:32:57] Speaker B: I don't know. [00:33:00] Speaker A: I like, I'm not, I'm not the normal user. I just like book stuff. Like, I fly the same airlines all the time. I'm so boring. I stay in the same places all the time. I'm not the typical, like. [00:33:11] Speaker B: But that's, I actually think that's a very typical traveler shopping behavior where you just go to the same hotel, right? Like, at least people will go back to the same hotel chain, right? Like, you wouldn't switch between Marriott. [00:33:25] Speaker A: It's a broader, it's a broader question. Like, I've often thought this, that like the travel booth boom was a specific millennial, like, trend that is over now that many of them have homes and families. And this idea that, like, you're gonna explore and go stay in a unique Airbnb and need to do a bunch of research and plan this amazing trip. I feel like that was a trend and that and that like what I just described. Like, I just go to San Francisco, I stay in the same places, I go to wherever and I just stay there. New York, I don't know. Like, I don't, I don't, don't like explore a lot of places. And like, there's like a one or two airlines I use. I have hotel points somewhere, so I just stay in that hotel. So I don't spend a lot of time like doing like research for my trips, which is. [00:34:09] Speaker B: Shouldn't. Wouldn't that also mean the agents would be perfect for you if they got it correct? [00:34:17] Speaker A: Yeah, that's, that's an interesting point. So, I mean, we didn't really bring agents in the discussion, right? I mean. [00:34:22] Speaker B: Or, I mean, Chat GPT. [00:34:25] Speaker A: But then like, it theoretically so that's theoretically an agent would be great at that. Right. Like, hey, Gregory, this is what you typically do. Boom, I set it up for you. I've yet to have that experience yet. I'd love to have that experience. Right. [00:34:38] Speaker B: Yeah, I just. Correct. So like just tying back to the ads thing. Like, I just don't see. Maybe I'm just being too negative, but I just don't see it being better than, let's say Google. Maybe I'll be wrong. It's a different platform. [00:34:57] Speaker A: Google has. We just, we just, we just, we just finished saying we thought Google was going to win. So like Google has a lot of. Yeah, yeah, right. They have a lot of advantages. They've been in advertising and they have, they have just more data, more teams, more knowledge. They have a lot. Right. [00:35:10] Speaker B: Of a lot. [00:35:11] Speaker A: It's hard to identify something in this space that chat GPT has as an advantage over, over Google, particularly when it comes to this part. They're just like, I guess first to market. I don't really know what they could do that would like, I don't instantly have an answer for like how they definitively beat Google at this. [00:35:33] Speaker B: And it feels. [00:35:34] Speaker A: Where is the, where's the advantage? [00:35:36] Speaker B: I don't know. I do feel like it's rushed, this ad product because Google Ads came out what, 2000, 2005, like that, let's say. [00:35:47] Speaker A: And Google had been around. They had been around a while. I guess that's. I'm ch. You know, open Ey has been around a while. [00:35:54] Speaker B: Yeah. Their product is five years old, let's say. [00:35:59] Speaker A: Yeah, I think that, I think the original like CPM push where they just sold like these just generic impressions. [00:36:06] Speaker B: Yeah, I don't think. [00:36:06] Speaker A: Yeah, I think that was rushed. I don't think that was a great way to launch this. I mean, and like what was it? Walmart? Someone was very public and said like they did not have a good. [00:36:14] Speaker B: Yeah, they did not get a good lift. All right. Okay, I'm ready to move. [00:36:19] Speaker A: Okay, we've done this enough. [00:36:20] Speaker B: Yeah. [00:36:22] Speaker A: SpaceX and cursor. [00:36:24] Speaker B: So this I did not see coming. Absolutely not on my big old card for, for 2026 where SpaceX potential moocher with cursor. So there. This whole entire story started with cursor coming out saying that they are now using Grok Grox data center to train their Composer 2A data model, LLM model and of course, you know, X, SpaceX, Tesla is essentially all Elon's baby. So they're all related to each other. But the fact that I think this deal was structured to be A. Either 10 billion now or a potential to buy the company for 60 billion in what, a year? [00:37:23] Speaker A: Yeah, there's a couple of different. They have an option to buy it. [00:37:27] Speaker B: What are your thoughts around this? [00:37:30] Speaker A: Okay, what I, when I heard this I just thought that like the Tesla SpaceX merger seems inevitable at this point. [00:37:39] Speaker B: You're saying all of Elon's companies will merge together? [00:37:42] Speaker A: That's the first thing I thought of is like SpaceX is going to merge. Like what is the difference between Tesla and SpaceX at this point? Like they both are like a mixed model of like hardware and software and forward thinking technology like I like. So that I don't see any way that these stocks survive as independence. [00:38:01] Speaker B: Yes. We had a little bit of analysis previously a few episodes back. [00:38:06] Speaker A: Yeah, I think your thing I thought [00:38:09] Speaker B: of, you're just to like refresh the mind. You're saying the same people that buys Tesla stocks are the same people that would buy SpaceX stocks and if they, if they go IPO. If SpaceX goes IPO, it will cannibalize each other. [00:38:31] Speaker A: Yeah, I can make it. I'm not really sure. I just think it's not great. Like it's not great for one of them. It might not be great for both of them. I'm just not convinced there's enough people [00:38:38] Speaker B: out there in demand that will buy either. Both stocks, let's say at the same time. Correct. [00:38:44] Speaker A: And then the SpaceX thing is so enormous. No, no company's ever gone public at that. [00:38:48] Speaker B: Like at a trillion dollar ipo. [00:38:51] Speaker A: It's just crazy. Like there's not that much money out there to buy it. So I don't know, I don't really know how it's going to work. Yeah, I do think, yeah, it's, I think it's pretty cool. I think it's really smart. I think Cursor is a great company. I think it makes sense for like the, let's call it the Elon company portfolio. All the things that they have. SpaceX, Tesla X Xai like great. I think it's a great idea, an excellent contribution to all of that stuff. [00:39:19] Speaker B: Do you think so? I, I correct. I, I do agree that with the fact that Cursor is a great company. I think they got the distribution of AI perfect. Do you? I still don't quite understand why SpaceX would be interested. Do you think it's Elon buying distribution again? So it's like the same deal with X Twitter? [00:39:44] Speaker A: Yeah. Great. This is a Good question. I. I think, I think Elon's determined to build a third alternative to like anthropic open AI. And maybe it's a fourth alternative if you were throwing Google. So Google, yeah, Anthropic OpenAI. He wants to be a fourth player in the AI in the like, broad AI space. We know that vibe coding, automated coding, whatever you want to call it, is perhaps the primary use case at this point for AI when it comes to [00:40:14] Speaker B: like, it's the most mature one. Right. [00:40:16] Speaker A: That people will pay money for. It's a giant vertical, enormous. And so that they need to have that as part of like the XAI AI initiative. I think it's as simple as that, that they need to have a play in that space. [00:40:30] Speaker B: I see. That would be really interesting though, if they come out with a fourth choice. I don't think the market will have Appetite for a fourth choice. I think 3.3player seems to be like the sweet spot. Somebody has to go away. [00:40:49] Speaker A: I don't, I mean, and the person [00:40:51] Speaker B: that goes, yeah, four players. Really? I don't, I don't, I don't think the market is ready for four players. [00:40:59] Speaker A: I mean, who do you think is [00:41:02] Speaker B: going to lose Anthropic. It's between Anthropic or open AI. It's definitely not Google. [00:41:11] Speaker A: Okay, I agree. It's not Google either. [00:41:13] Speaker B: It's definitely not Google. [00:41:15] Speaker A: It's not X. [00:41:20] Speaker B: X has distribution that the other two companies don't have. So if they, if they buy Cursor, that's even more distribution. Right. So I think it's a distribution play. They're directly competing with Google's distribution. Google has chrome, Google has YouTube. You know, Google's work. [00:41:41] Speaker A: I don't know. Like, so, okay, so, so then Ulta, if you believe that, which I think I agree with, then it means it has to be Anthropic or OpenAI. [00:41:49] Speaker B: For sure. Yeah, it's between those two. [00:41:51] Speaker A: One or the other is just not gonna, it's not gonna work long term. [00:41:54] Speaker B: Yeah, yeah, yeah. Because neither of the company has distribution. [00:41:59] Speaker A: They like, you'd think like the first thing I thought of was like, like the both of those companies, like the recent equity raises they have done. Buying Cursor is cheaper, like to get all that distribution, get all that access, I guess, like they couldn't buy it. Maybe that was the issue. Like, why didn't one of them buy it? [00:42:22] Speaker B: It's because they are stuck in a war to build the next best model with each other. That's going to be their downfall. Only One of them is going to survive because they're, they're basically competing on what the next best model is. Right. Like Google does. Google has zero need to compete on the best model front. They don't care. They could just take their time. They have zero need for that. [00:42:51] Speaker A: I agree. [00:42:52] Speaker B: And Grok is the same. Grock is just chucking along. They're improving it cycle versus cycle. But Open AI Anthropic is stuck in this war with each other. [00:43:07] Speaker A: Fair enough. I don't, I, I don't feel conf. I don't feel confident I have enough information to make a pretty prediction in either. Either way. [00:43:14] Speaker B: Either way. As of right now, I think Anthropic will likely not survive. That's my prediction. The CEO doesn't. Doesn't have it. Sam Olman, I, I think Sam Almond, he's. He's got it. The CEO Anthropic. I don't, I don't think the guy's got it. [00:43:40] Speaker A: I agree with you for a different reason. [00:43:43] Speaker B: What's your, what's your take? [00:43:45] Speaker A: I think Chat GPT is the consumer brand of preference for AI or they have the awareness with consumers. [00:43:53] Speaker B: I think the average first to market. Right. [00:43:56] Speaker A: Anthropic. Claude, with that. No one uses xai. You have to be a crazy Twitter person. People have heard of Google. I don't know if they heard of Gemini. They probably understand that there's this AI mode thing. The fact that like. Yeah, yeah. The fact that OpenAI has. Has like the consumer brand for what the average person thinks AI, I think is massive. [00:44:19] Speaker B: I think so too. [00:44:20] Speaker A: And that's why I would. If I had to pick, that would be my logic. [00:44:23] Speaker B: They got the brand recognition. Google's got the distribution. They bake Gemini into Chrome. So they're going to put that in front of your face and then they'll have a market. X is this weird little thing they're buying cursors. So [00:44:38] Speaker A: they use it at like, like Andrew and stuff. I like. I. It's. It's strange. Yeah. I don't know enough about it. [00:44:45] Speaker B: So. All right. [00:44:48] Speaker A: Amazon's quest for tool consolidation. [00:44:51] Speaker B: Okay, so this is, this is yours. This is a, this is a fun story to tell at least. So it kind of. [00:44:57] Speaker A: I'm. I'm ready. [00:44:58] Speaker B: Directly presents the problem with AI slop and companies creating hundreds of tools. That's essentially a duplication of each other. So, okay, the core story is that Amazon is running into a massive problem where they don't know what people are building inside of their own company. People are building Essentially the exact same tools over and over again. So they created this new role, this person whose entire department is to go out and catalog these tools so that they can send list of whatever these AI tools or vibe coded tools are so people stop building the same thing over and over again. What he didn't realize he publicly said on X is that he didn't realize his own department was a duplicate. He didn't realize that somebody else was already doing his job of cataloging duplicate AI tools. So that created a huge issue within, I mean the entire company. And it's not just Amazon that I'm hearing this problem. This is happening across all big enterprise companies. I think it's, I think it's going to be a huge problem in, let's say a month. What are your thoughts? [00:46:17] Speaker A: Okay, I love this. I thought a lot. So we talked about this yesterday in like our like meeting. I thought a lot about it and what it reminded me of was like some of the management principles I use when I had like a rather large team, right. And I used. I love spectrum. I always think of everything on spectrums. I don't like black and white on or off types of things. I like things that are on like graded scale because that's perfect. Better. Better how? I think it's more accurate to the world. There's always things that are like gray or you need to just think about them in that way. So what I used to tell people was like, when it came to decision making, use a sliding scale to make decisions on your own versus like what you need to get approval for. And what I said was like, if it's very easy to reverse, just do it. Don't ask me, don't ask anybody. [00:47:06] Speaker B: Right. [00:47:06] Speaker A: If it's very hard to reverse, then you should get feedback. And this is actually the same principle as like Jeff Bezos's. Yeah, one way, one way, two way door. [00:47:16] Speaker B: Right, right. [00:47:17] Speaker A: But I, I have a spectrum. He has a binary one way, two way, like analog. Right. Where I talk about on a spectrum because I say that like, there's probably some things that like maybe you should get some feedback on. There's probably some things you don't need to get any feedback on. There's probably some things we need to get a lot of feedback on. There's even things that like we should get an enormous amount of feedback on. That's why I like the, I like the spectrum. So I would apply this to this idea about individual teams and groups and organizations building tools. I would apply like a sliding scale to determine if you need approval or if you need to do your own initiative. So I think there's some things that's very obvious that there should be a company wide decision on this thing. And so I would put them in like, let's call it a platform category. [00:48:10] Speaker B: Yeah. [00:48:11] Speaker A: So we all consolidate around this, we all do this. Like I would. It's very different, I think, in different companies. Like I would say something like maybe you pick a couple of cloud providers, maybe there's one preferred cloud provider. Right. Like everything goes on Amazon, everything goes on Google Cloud. Whatever is like there's some way to manage that. It won't be perfect, but there's some way to manage that. And then on top of that, like you, you pick some languages, you pick some like services, you pick some things you like to like, kind of like platform level stuff. And you require everyone to kind of build and work within those parameters. But on top of that, maybe there's some shared tools you want everyone to use, but you let people just build whatever they want. And I think that people duplicate initiatives sometimes. It's fine. I actually think it might be more efficient, it might be faster, the tool might be customized to what they need to do. I wouldn't worry too much about forcing everybody to document everything and have every tool listed. [00:49:02] Speaker B: Of course. Yeah. That's not realistic. No. Nobody's going to. [00:49:05] Speaker A: Giant waste of resources. It's a giant waste of time. [00:49:08] Speaker B: Yeah. [00:49:09] Speaker A: So I think that like there's certain things that we need to be aligned on and I would kind of, I get, I would onboard and train people that's like, hey, these are the, this is the set of things that we all agree that we do this way. And there's a bunch of reasons why we did this way. And that's why we say, like, we all agree on this. And I, and I basically would, I guess every team I've ever managed, there's some form of this that exists. Like these are our standards, this is how we do things. Right. On top of that, you have a lot of leeway with other things you want to do. And then I would say something like just involve people that you think make sense in the initiative. If you want to build a shared version of it, go for it. You don't want to build a shared version of it. Don't waste the time. Like make it really simple. Just make. Because I think vibe coded throwaway apps are a really good idea. People should do more of them. So the fact that you can have a lot of tools that are like duplicate duplicative or customized in a way that aren't appropriate, I think is fine. [00:50:01] Speaker B: So where do you think the problem is? Do you, you think the problem is when you start introducing like third party duplicate licenses because like you're not going to be able to build everything just completely without dependencies. Right. So like you, you start subscribing. [00:50:18] Speaker A: I have a very, I have a very clear perspective on this. I think the problem is when you build teams that are a sole mission and purpose is to enforce standards. I guess at a high level I'm against that. [00:50:29] Speaker B: I see, right. [00:50:31] Speaker A: If you have a team and like it sounds like this with this guy's on, he's on a team that's job is a document and governance or some type of governance across. And that's where I would say the problem is because like if that's your mission, your team, obviously the mission is to like go as far as you can in like documenting and building standards and norms and policing everything. [00:50:53] Speaker B: Yeah. [00:50:54] Speaker A: Which I think is actually a disagree with for the most part. [00:50:57] Speaker B: You disagree? [00:50:57] Speaker A: Yeah, yeah. That's why when you we talked about this, I thought about it from like an organizational principles perspective and enough almost like a philosophy about how you manage things. And in general I'm against like yeah, large entities that have these enforcement groups that try to like manage everybody the same way. I'm much more open to like and just believe in like teams just having a lot of leeway to do what they need to do. Teams and standards. I guess that would be my, my simple way of like talking about, don't [00:51:26] Speaker B: you think like all of these big tech companies especially I would say Amazon got to their size because of these standards. Right. So like the way that I think about it is like front office versus back office standards. On the front office you could have a little bit more leeway. You could do it your own way, different from department to department, team to team. But at the back office there's a lot of things that you just have to standardize. Right. So I think Amazon as a company, they created lots of interfaces that are standardized. For example, you have to create APIs that's usable and plugged in that anyone outside of your department or team can use. That is a force of standardization. You have to create this API. [00:52:17] Speaker A: This is, this is a really big, there are like, there's like entire disciplines around this concept. So this is a big question, like right, how, what is the philosophy of the organization, how thing and how people, how teams, how things work together. [00:52:32] Speaker B: Right. [00:52:33] Speaker A: And there's lots of Books written on this. So. So at a highest level I think that like the most successful organizations are good at a certain approach and they do that really well and they're all different. I don't think there's a universal way to do it. I think some organizations are really good at building standards and making everyone adhere to that and enforcing it and other organizations that are really good at like letting individuals and teams run with things and work in a really independent fashion. I think it's like really different. I think that like. [00:53:04] Speaker B: So do you think like AI blows all. It's essentially like this AI introduced sprawl, let's say call it. Do you think that just blows up the old way of management? [00:53:19] Speaker A: Well, it's one of the, it's one of the value props of Vibe coding. Right? Just spin up an app and so far I think it's a good idea [00:53:26] Speaker B: and that screws up a lot of the governance that's introduced or the legacy governance at these. [00:53:31] Speaker A: I mean it's an excellent segue to our next one which is like that the Apple App Store is completely overrun. VI Coded apps I think it's exactly, that's exactly it. I think it's exactly what it all heads is that the nature of like creation has changed. Barrier to entry has been vastly reduced and we don't necessarily have systems and things in place to manage a world like this. I think I, I think it's so fundamental. The world's changing. Software is now going to become very easy to make. We lived in a world where software is scarce. We don't have the systems in place, we don't have the methodologies in place. No one knows how to build to govern all of this Vibe coded software company. [00:54:13] Speaker B: Yeah, yeah, yeah. [00:54:16] Speaker A: I think it's, I think it's like, it's that fundamental, right? Like what is the large scale Vibe coded software company Operating model is my model. Nobody knows Spectrums and no one knows. [00:54:26] Speaker B: No one knows. Nobody knows. And also like how do you manage the distribution of these Vibe coded? [00:54:34] Speaker A: I don't know. [00:54:35] Speaker B: Right. App Store. App Store is literally getting destroyed. So like. [00:54:39] Speaker A: Yeah, well just, just to cut, just to cover off on this specifically like what it was that there's a post that went out saying that the App Store at Apple is completely overwhelmed. There's just way too many apps that's [00:54:54] Speaker B: being submitted for review. [00:54:57] Speaker A: They don't even know how to like how to manage it. Like we have like a hundred in this category in 2000. This kind of like how do you [00:55:03] Speaker B: even think they have the exact. [00:55:06] Speaker A: Right. [00:55:06] Speaker B: They have the exact same problem Amazon has internally, which is categorizing, cataloging, governing. Yes, it's the exact same problem publicly. Essentially. [00:55:17] Speaker A: This, like, we're, we're. There's certain areas of like the world, the economy or scarcity is going away. We've. We've never lived in a world that was abundant in a. This way. I think, I think it's that fundamental. [00:55:33] Speaker B: That's. That's a good take. Yeah, that's a unique take. So back to the point about, you know, copilot going away or sign up going away. I think the only way to solve that as of right now is limiting who gets access to tokens. I think companies are. [00:55:51] Speaker A: That's the market. That's a market solution. Right. To. Tokens are expensive. That's the way you manage this. [00:55:55] Speaker B: That's right, yeah. So you reintroduce scarcity back into the token economy. So who gets access? [00:56:02] Speaker A: Like some things are not going to take that many tokens to produce. There's a whole level of things that were like not, not don't require that many tokens, but they used to require a lot of knowledge. Correct. Right. Yeah. It's really interesting. [00:56:20] Speaker B: So, okay, so there's this is, this is cool. So the two ways I think it will unfold is either token companies will raise the price that will naturally make vibe coding more expensive and more scarce. Just raise the price. Or these companies will introduce some kind of gating to token access. So you have to have a company email to have access to these plants. So. Right. [00:56:57] Speaker A: Like, so actually I think it's really simple. Yeah. Like tokens are expensive. It's like paying for gas. Right. Like there's going to be a limiting factor of like the consumables. [00:57:08] Speaker B: Right. So the token is consumable. You don't think it's going to be limitless? Right. Like, it can't, it can't possibly be limitless. [00:57:15] Speaker A: Maybe someday. I mean, is bandwidth limitless now? [00:57:19] Speaker B: Yes, in some perspective. Yeah, you can see that. [00:57:21] Speaker A: Right. [00:57:22] Speaker B: Internet's limitless. [00:57:23] Speaker A: I think it's like, I think bandwidth, Internet bandwidth is the best example of how this progressed in the past. Like, bandwidth was scarce at one point. [00:57:33] Speaker B: Yeah. [00:57:33] Speaker A: Then we built it out and then there's plenty of capacity and like there's. I don't run into a lot of things I want to do. I don't have like the digital Internet bandwidth to do it. It's pretty rare. [00:57:46] Speaker B: It's very rare. Right. [00:57:48] Speaker A: Yeah. So I think it'll be like that. I Think token scarcity will only go on for a couple of years. [00:57:52] Speaker B: So two years. Okay. So another reason Google will win. Google became Google because they categorized the Internet. They fundamentally solved how people interface with the Internet by completely categorizing it, indexing and categorizing it to make it searchable. True. [00:58:15] Speaker A: They gave access to the average person. [00:58:17] Speaker B: Right. [00:58:18] Speaker A: A hundred percent. [00:58:18] Speaker B: So like you know, as tens and tens of millions of people got onto the Internet they were there to help them search, categorize, index the Internet. Somebody needs to do the same with, with apps. Potentially Google. Potentially Google. [00:58:35] Speaker A: Well that Android. [00:58:36] Speaker B: I don't know. [00:58:37] Speaker A: Yeah, I I I like it makes sense. [00:58:40] Speaker B: So very interesting. Shall we move on to the last, [00:58:45] Speaker A: last one this year? You want, what do you want to say on this one? I so what is, what is monitoring a situation? [00:58:50] Speaker B: It's a 24. 7 live stream on X. I think currently it's a replacement for the technology broadcast. I what do you think? Do you think they will survive on their opening day they got 250000 concurrent watch streams. [00:59:11] Speaker A: I haven't watched it. I don't have the time. I'm really busy some I'm busy doing our live stream. Maybe they'll feature us. How's that? I'd love that. We can go on the shows on there. I hadn't even thought about it. Maybe we go on the shows on it. Yeah, that'd be cool. [00:59:25] Speaker B: Do you think the format will work? 24 7, 247 live stream. That was the original their pub. Their launch post says that it was the original format that CNN wanted to be. [00:59:38] Speaker A: So I think it's cool. I think they're trying things. I don't know if it'll work. They're in the arena. [00:59:43] Speaker B: They're in the arena. [00:59:44] Speaker A: That's cool. I like to do it stuff I don't know like I, I I I I Discovery. [00:59:52] Speaker B: Okay. [00:59:53] Speaker A: The best way to just frame this is that like discovery has become broken across the Internet. I'm not really sure how distribution works anymore. It's very difficult it seems to aggregate and get distribution and I think that's the problem and I think that that will be the issue. Like it doesn't pop up my feed. I don't hear a lot about it so don't really know how they crack that. If they, if they do solve that they think it could be really popular but like I think it comes down to like a distribution challenge. [01:00:24] Speaker B: It is. Yeah. Yeah. They need clips, you know lots and lots of short 50 second clips. [01:00:30] Speaker A: Short form, short forms. [01:00:32] Speaker B: Yeah. [01:00:33] Speaker A: Well, did you have fun today? [01:00:35] Speaker B: Perfect. Yeah. Should we edit here? Anything. Anything you want to wrap for. For event in May? [01:00:44] Speaker A: No, just. Yeah. We'll be in Entrepreneur first on May 12th if you're interested. You have your founder, your startup enthusiast. You want to come and hang out with us, listen to some pitches? You're more than welcome to apply on Luma, and we'd love to see you at the event. We got a couple hundred people crazy signed up now. Yeah, I'm going through all the pitch decks. Great selection of pitch decks. So I'm excited about the pitch competition for this time. And, yeah, we do once a quarter, so. Next one will be at Snowflake. People are already signing up for that one. [01:01:21] Speaker B: Wow. And that one's in September. [01:01:24] Speaker A: September. Yeah. I'm trying to do a Q4 one, but, like, would that be, like, November? I haven't figured out exactly when the Q4 one happens, but we'll figure that out. [01:01:37] Speaker B: Awesome. All right. All right, dude, we're done. [01:01:40] Speaker A: One happens, but we'll figure that out. [01:01:42] Speaker B: Awesome. All right. All right, dude, we're done. Okay.