Episode Transcript
[00:00:01] Speaker A: We're back.
[00:00:02] Speaker B: All right. Is it working? I think it's working.
I can't tell. I can't tell. Oh, that's okay. Hey, welcome back to the Gregory and Paul Show. I'm Gregory.
[00:00:13] Speaker A: I'm Paul.
[00:00:14] Speaker B: And we always break down the latest SaaS, startups, AI.
Our Elon clips are always popular, so we'll have Elon corner every week.
Paul, how, how are we doing on the podcast platform?
[00:00:29] Speaker A: Yeah, we're uploading every single week. People are listening. Number one podcast soon.
Can't claim it right now. Soon.
[00:00:39] Speaker B: We're almost at a hundred subscribers on YouTube. So that's cool.
[00:00:43] Speaker A: That's huge.
[00:00:44] Speaker B: I think we, I think we figured out, I think we figured out YouTube, but I, I've never listened to us on Spotify, have you?
[00:00:52] Speaker A: I do. That's actually how I consume.
[00:00:54] Speaker B: Oh, really?
[00:00:55] Speaker A: Yeah, yeah, for sure. For all podcasts. I listen to you on Spotify, dude.
[00:01:00] Speaker B: We have a lot to talk about today.
[00:01:03] Speaker A: What's, what's number one? What's the most exciting AI technology?
[00:01:07] Speaker B: Yeah, I, I'm, I, I, I can't believe that OpenAI has finally announced their I product.
[00:01:16] Speaker A: I product.
[00:01:17] Speaker B: Yeah.
[00:01:17] Speaker A: So they, they rumor something came out, what like two, three months ago and then Sam backtracked on it and then now they're run.
[00:01:27] Speaker B: Let's share, let's, Let me share quickly this screen because like I got like, you know, Adweek already ran something on this announcement. I actually don't think there's like a date in here and then there's a blog post this, that this quotes.
But this has been long rumored and I'm sure on this show, but in other places I've been saying that like they better get this out the door. I think that they should have had it out last year.
Yeah. Now it's coming out this year and the article even mentions that perhaps that was part of what was in the code red or the red alert, whatever Sam did, you know. But I think it's, I think it's good, obviously.
Or there's some speculation that they're trying to get this out in time for the Super Bowl.
[00:02:25] Speaker A: Hey, right, Right.
[00:02:27] Speaker B: Makes sense, right? Like digital tie ins role have always been a way to introduce ad products. Twitter did that at one point and so I think it makes, I think it makes a lot of sense and they might get some traction. I have no idea though. I haven't heard of anyone's actually buying anything around some rule or how this is going to get rolled out. But it's interesting, the reaction, you know, Like I'm on the marketing side. And I was like, look, marketers are like excited. They, they're like, I need a new channel. Like Google search is going away, clicks are dying. Just everything is expensive and difficult.
So a new channel that they can take advantage of I think is something that Marker is looking for. But of course for users, people are like, oh, it's the end of ChatGPT now it's going to be influencing all of the content.
Yeah. So what are your thoughts on it?
[00:03:21] Speaker A: Yeah, they're coming out with a cheaper access subscription.
[00:03:28] Speaker B: It looks like that's the case. Right. Is that there will be a.
So what it says here is that they today we're bringing Go so Chat to me to go, which is the like low cost version to the US and everywhere chat to be available, giving people expanded access, messaging, image creation, file uploads, memory for $8 a month. The coming weeks we're also planning to start testing ads. So it looks like whatever they go version, that's the one that I guess you'll pay and you'll get ads. So they need to monetize.
[00:04:03] Speaker A: So that makes sense. Right. So it's going to be more than the free access.
So $8 plus ads gets you all access.
I think that's smart.
[00:04:14] Speaker B: Does it mean that, that the free tier has ads? It has to. Right?
[00:04:19] Speaker A: Has to for sure.
[00:04:22] Speaker B: It doesn't actually specifically say it's the low cost subscription tier. Interesting. You know, it's funny, I've had the paid version for so long now, I'm not even aware of like I wasn't actually aware that There was an $8 version.
[00:04:38] Speaker A: Right. I didn't know either.
[00:04:40] Speaker B: Yeah, I've been using the $20 a month one.
I even canceled a bunch of other stuff and I still use, I still use. Because there's a lot of stuff in there.
[00:04:48] Speaker A: Yeah.
So what I'm wondering is will the ads affect the actual conversation?
Have they said, have they said no, like it will not affect.
[00:05:02] Speaker B: Yeah. There's a long blog post that they wrote and of course like they, they have to publicly state that the content integrity won't be influenced by the advertising.
But as we know with like this is actually quite interesting question. Right. Like, because I guess with Google, like there are, let's say influences on the content.
Are they advertising driven? Are they business decisions? There's a lot of ways to kind of spin it.
What happened with Google, which I don't know if this will happen with ChatGPT but is that they would have a big category of Advertising that was popular and they would actually release a competing product and then they would take a lot of that traffic and send it to their own application. It's kind of a. It's technically, I think, different than like influencing the editorial or the content.
Right. Like, it's not actually influencing the content, but it is influencing the ecosystem.
Definitely influence like business and advertisers.
[00:06:10] Speaker A: Right.
Okay. Okay. So, right, this is going to happen. And then now do they like all other social media, do they optimize time spent?
[00:06:25] Speaker B: What do you mean?
[00:06:26] Speaker A: More people spending time in ChatGPT, more ads are served, they make more money.
Right. All social media optimized to keep people on the platform.
So ChatGPT will probably do something similar, right?
[00:06:41] Speaker B: Yeah, that's a really interesting question. I mean, we had, you know, Andrea from thread on. Right. And so thread is an early LLM ad network that puts ads in other LLMs non chatgpt.
And so there was a lot of questions about like measurement and how the ads perform and stuff. So it's still really early. I don't think it's clear what is the right metric to optimize to like how you would increase ad load. They're still trying to test this stuff. Like one theory that a lot of people, let's say are betting on for this year is that the ads will be really popular with retail.
So commerce.
Yeah, I would, yeah, that was the first thing I thought of was like travel and then just retail. Like, I mean, what I, what I want is the ability to like just talk to my phone and say, buy me some new drivetrain cleaner for my bicycle. And it just goes and gets it for me from Amazon or whatever. I mean, it's all like LLM driven.
And so I guess that scenario, perhaps there's an ad for like a different version of it or something. But that's what I see is like chats happening about products and then the product recommendations are sponsored.
I can't say how that's not the number one use case. Maybe restaurants.
[00:08:04] Speaker A: So do, do you think that retailers or not retailers or aggregate sites are in trouble.
Right. So Expedia, let's say people go on to Expedia for travel, to book hotels, rent cars, to buy plane tickets. Now they can all do this in ChatGPT.
[00:08:24] Speaker B: Yeah, it's really interesting question.
[00:08:26] Speaker A: They're massively in trouble. Right? Like they're, they're a public company, dude.
[00:08:29] Speaker B: I don't know. I feel like there's one part of it that's like, there's so much demand Right now because Google's declined so significantly in terms of the traffic that it drives.
[00:08:38] Speaker A: Right.
[00:08:39] Speaker B: That everyone's looking for an option. And if you have an option where it's a lot more traffic to be able to people to buy, it could be good for the whole ecosystem. Like some of this is a supply and demand, right? Like there's a lot more. So right now the supply, the ability to buy traffic and reach users is being constrained.
And so now there's just more places to buy ads.
In some regard the price might go down for everybody and the performance might just be better. It's that big that it could impact us.
I remember when Facebook launched their first mobile ad product and initially like to drive business, it was like dirt cheap.
It was like super cheap. So if I was OpenAI that's what I would do. I would make the pricing so ridiculously cheap to just win a lot of business.
So there's theoretically like a huge arbitrage opportunity coming for people to take advantage of this and buy a bunch of low cost advertising and traffic and kind of scoop up a bunch of users or traffic, whatever you're trying to aim for.
[00:09:45] Speaker A: Who's the biggest loser in this? Google Ads? Would you say search ads? Providers?
[00:09:52] Speaker B: Users.
[00:09:54] Speaker A: Users, of course.
[00:09:56] Speaker B: Of course, yeah. I mean like you knew is inevitable right that like they're going to stuff ads in here and like you know, like look, I take them at face value some degree that they, they, they aspire to. You want the content to be editorially neutral and very hard to do so. So I believe that they will at least aim for that. But it will influence the content and it'll change the nature of how the product works. It's, it's just how it's gonna, how it's gonna play out. I don't think there's anything sinister. I think people get really up in arms about this stuff.
But I'm not that negative. But I do think that like it just will change, it'll change the experience,
[00:10:32] Speaker A: it'll change the platform. Right. For sure.
[00:10:34] Speaker B: And then there's, and then, well this dude, like what if they open up the ability to buy content on like controversial topics? I mean like ads have influence. Yeah, like political content. Right. Like they could let people buy that and then, so, so by the nature of how advertising works, someone who's a big ad buyer could just influence the content because they bought a lot of ads.
I mean that's how advertising works.
[00:10:59] Speaker A: This opens up a huge thing because then advertisement brand safety is going to be a problem, right? You don't want somebody to be searching.
[00:11:09] Speaker B: You know, there's a lot of questions that like I assume that they're not going to allow like adult content.
I mean
[00:11:17] Speaker A: they came out saying that adult content or adult writing is okay
[00:11:22] Speaker B: on ChatGPT, some, some forms of it, right?
[00:11:25] Speaker A: In some forms, yeah. You can use chat GPT to create exotic writing, let's say exotic writing.
[00:11:35] Speaker B: And I don't want to know how you know this, Paul, but I'll take, I believe you.
[00:11:40] Speaker A: And you know what's going to be hilarious is like Disney like that. They're not gonna want to run ads side by side with that.
So yeah, This raises a question of real time observability into your conversation.
[00:12:02] Speaker B: Right.
[00:12:03] Speaker A: Giving, I mean you're giving access to advertisers to say that I don't want to be running ads inside of these kind of conversations.
[00:12:12] Speaker B: So yeah, yeah. What I've seen with the examples that are running now and other LLMs is that the ads are very effective from like a marketing perspective.
Yeah. I think most users like nerds who are not obviously watching a live stream like this, they don't care.
They're not even all that aware that they're ads. They're just happy that their problem got solved. And like I needed like a ticket to wherever and I, I figured out how to get it and it was easy. I needed to buy drivetrain cleaner for my bicycle. I got that. Right. That's what the end users like actually care about. And LLMs, at least in some contexts that I've seen, are actually really good because you can ask questions like this, like if you have this conversation, let's use the bike drive clean drivetrain cleaner example because I think it's actually a very good one.
[00:13:05] Speaker A: Right.
[00:13:06] Speaker B: Is this non toxic? Is it like you could just ask a bunch of questions that you have about the product, right? Like when will it be able to be delivered? Right. Like if you could have that conversation. LLM that's actually really a good user experience and trying to like click around on Amazon and you know, avoid like all the, all the noise. Yeah, like, and all the reviews are full of just like spam.
They're not very believable anymore. I don't believe them.
And so having a conversation with LLM about a product and you're just trying to get like basic information, I think it could be a really good experience.
[00:13:39] Speaker A: Right. There's a level of trust. You're trusting the LLM to be factual, let's say, you know.
[00:13:45] Speaker B: Yeah, factual is an interesting term. It has to. It has to. I think ultimately I would say something like satisfy the user. I don't know if they're facts. I think they are like measured on like the user satisfaction. Because if the user asks questions and then they get the product and they're happy with it, that's the measure, the facts.
I think this. And this isn't like this. I think it's actually a very important point that like, because I've been in marketing a long time, like, facts is an interesting term. Like users don't want facts, they want solutions.
It doesn't matter. The facts are not important. What's important is like he got the drive train cleaner that he wanted on time and it cleaned his bike correctly. That's what they want.
[00:14:33] Speaker A: Right.
[00:14:34] Speaker B: The facts might be a different.
It's kind of a side quest to some degree anyway.
[00:14:40] Speaker A: Got it.
So the measurement's going to change satisfactory to AI Ads or AI.
[00:14:49] Speaker B: I hope so. I mean, yeah, I think it has to some degree play out like that.
Right.
But yeah, big news.
[00:15:01] Speaker A: What's.
[00:15:01] Speaker B: What we Y.
[00:15:02] Speaker A: Predictions. Predictions on when they're going to roll this out.
Paul, buy a.
[00:15:07] Speaker B: Like, I mean, like I said, dude, if I was them, I would be racing to get this out for super bowl.
[00:15:14] Speaker A: Right.
[00:15:14] Speaker B: And I would do some tests and stuff like.
And like they announced it January 16th. Like, do you know what day the super bowl is?
It's in a few weeks.
Yeah. I don't follow football anymore.
I wish I've had more free time to follow sports. I do like sports a lot, but I don't have the free time. Like I go to a baseball game. It takes all day. I don't have all day to go watch baseball, but it's, it's coming up. And that, that was why I thought they made the announcement now was that they want to have some type of like ad execution available to test during Super Bowl. But that's just my, my assumption.
[00:15:50] Speaker A: So people can search up football, faxing, chatgpt, have arguments with their friends.
[00:15:54] Speaker B: No, I think it's gonna be like Doritos and Bud Light, you know?
Yeah, it's always Doritos. There's always like 10 Budweiser commercials. There's always like some wild card SaaS company that decides they want to do one and then they never do it again.
[00:16:14] Speaker A: Yeah, right. The Coinbase, World
[00:16:18] Speaker B: Coinbase actually did a really great job with that QR code. I thought that was fantastic, personally.
[00:16:24] Speaker A: That's right, that's right, that's right. Yeah. Much better than the chatgpt Ad I
[00:16:29] Speaker B: was not a fan of the Chat GPT ad they did last year.
I, I got pretty heated about that on, on X. I was like, this is, this is garbage. Like, I'm not a fan.
[00:16:41] Speaker A: What other news are.
[00:16:43] Speaker B: All right, so we move on. Yeah, We've talked about OpenAI. What do we got in our document?
[00:16:50] Speaker A: Okay. Apple unlikely. I did not have this on my bingo card, but kind of predicted that it was coming.
Google is partnershiping with Apple into a multi year collaboration.
[00:17:06] Speaker B: Yeah, let me share.
[00:17:06] Speaker A: It sounds like.
Yeah. So is this it?
[00:17:12] Speaker B: There we go. Yeah.
[00:17:14] Speaker A: So it sounds like Instead of using ChatGPT, Apple will leverage Gemini as to drive Siri to drive a personalized AI on Apple devices. This is huge. This is, this is wild, actually.
Yeah.
[00:17:34] Speaker B: Dude, what happened? Like, there was an announcement where OpenAI said they were like gonna be integrated into Apple and I guess that that's not happening.
[00:17:45] Speaker A: That's not happening.
What do you think?
Any hypotheses on what the fallout was?
[00:17:56] Speaker B: Dude, there's actually a lot of angles on this. Like, like, well, I throw it back to you.
Is Apple winning in AI or losing? Like by like kind of sitting this like round out.
Is it smart for them? Like, what do you think?
[00:18:13] Speaker A: Sorry, smart for who?
[00:18:15] Speaker B: For Apple. Right. They never invested in an LLM. They don't really have an AI strategy or initiative or product. They kind of pretended with Apple intelligence at one point and then, you know, like Siri still kind of is just terrible. Right. Like, I found that all really interesting and I'm curious if you think that that was like a smart move. That's kind of like there's two interpretations. One is like, it was really smart. They didn't make that investment and they're just going to own whatever, you know, the hardware layer that they own and the, and the connection to the consumer which is really valuable. Right. And that why even bother? Just give the keys to the kingdom to Google or AirPoint AI, whoever. Google.
The other interpretation is like the keys of the kingdom have gone to Google.
Is that good for Apple?
[00:19:05] Speaker A: So I think Apple's really good at figuring out what the consumer want.
AI so far has been really focused on B2B. So business side, nobody has figured out AI for consumer yet. Not, not really. Not really.
So Apple, I think with this partnership, they are actually going to roll out personalized AI and I think the market is so huge that they're willing to partner up with Google.
[00:19:36] Speaker B: All right, you're totally wrong. This is what I think, actually.
I think I know the answer.
[00:19:42] Speaker A: Go for it.
[00:19:43] Speaker B: Johnny eyes works at OpenAI.
[00:19:46] Speaker A: Works at OpenAI.
[00:19:47] Speaker B: Yeah. I just remember this. Johnny AI. Johnny. Johnny AI.
[00:19:51] Speaker A: Johnny,
[00:19:53] Speaker B: Johnny AI Ives is at ChatGPT and OpenAI creating a competitor to the iPhone. And that is why Apple does not want to partner with OpenAI.
It also makes me think that whatever they release might actually be good. A lot of people are saying like, oh, it's garbage. I don't want it.
I am going to get it. I'm going to try it out. I think it's going to be some form of the humane AI pin and the friend puck. It's going to be some kind of like voice activated device. You're around recording everything.
I actually think that Jony Ives is a genius. I think the device is going to be cool and I'm going to try it out. It might not, which is okay too. I don't think there's anything wrong with that. It might flop. But I think that they're really onto something. They've got a lot of money and Apple in one interpretation is scared.
[00:20:47] Speaker A: So I'll push back on this. Okay, so Johnny, I've no doubt he's going to create something really beautiful. And I read the specs or the rumors of what's coming. So they're going to use, first of all, it's a small device, right? Personalized device. Yeah. The rumor is that it's going to use two nanometer chips inside of this device. Okay. There's only literally two, three companies in the world that can make this device. So it's obviously based in China.
I think that what Apple is really good at is they can figure out the device like the iPhone and manufacture a billion of them.
Like that's what Tim Cook is amazing at. He is the supply chain guy. He figured out how to manufacture billions of iPhones over people like Jony Ives who are designers and Product people for ChatGPT or for OpenAI and Johnny Ives to challenge Apple on this. They need to figure out the supply chain. They need to basically figure out how to get this device into a billion people's hands. I don't think they know how to do that.
[00:21:58] Speaker B: Well, it's super hard. It takes time.
[00:22:00] Speaker A: Yeah, I, I think the relationships are not there. The trust is not there. The manufacturing know how is not there. Only companies like maybe Samsung, Apple, two of the only consumer companies in the world that can distribute.
[00:22:20] Speaker B: Dude, this is a really good question.
I don't think the distribution matters at this point. I think having a breakthrough device that people like is so like what held back the humane. Yeah, humane AI pin was not manufacturing.
[00:22:33] Speaker A: Yeah.
[00:22:34] Speaker B: What held back, like, I don't know. Meta Ray Bans is not manufacturing like what's been holding.
[00:22:38] Speaker A: There's.
[00:22:39] Speaker B: I mean, I remember as a kid, but I haven't, I don't remember the device or the physical product today where there's more demand than they can meet. Like when I was young, I remember that there'd be some toy, some computer device or external peripheral or hard drive feels so old that like they were hard to get because they just really couldn't manufacture enough of them. And there was so much demand, people were buying them.
[00:23:02] Speaker A: Right.
[00:23:03] Speaker B: I can't remember the last time. Well, I guess during the pandemic perhaps there were like shortages of like physical things, but I can't remember the last time that that was the barrier.
[00:23:13] Speaker A: Well, the, There's a huge demand of hard drives and memory sticks, so memory RAM instead of computers right now.
[00:23:22] Speaker B: Yeah.
[00:23:22] Speaker A: Price hard drives and RAM doubled.
[00:23:25] Speaker B: But they, but, but the people are able to like deliver them. It's not like a huge backlog.
[00:23:29] Speaker A: No, no. There is literally a shortage. You cannot get your hands on these components instead of like hobby computer stores.
[00:23:37] Speaker B: But like what, what device is it?
Is it.
No one's lining up to get it though.
[00:23:44] Speaker A: So PC gamers, right. So PC Gamer has been this like consistent consumer base that buys computer parts like RAM and hard drives. And because of the AI stuff, they can't get their hands on Memory City or it's. The price has doubled. So like for.
It is actually so with the, with the device that ChatGPT Johnny Ives talked about, it's based on 2 nanometer chip technology.
We have not figured that technology out yet. It's.
You're literally putting your entire phone into a smaller device.
[00:24:21] Speaker B: Yeah. So, I mean, look, I think that's interesting. I did hear about that. I mean, I haven't, I don't know, I haven't experienced it personally and I don't know anyone has.
But I think the reality is that if they can sell 100,000, 200,000 of these things, it'd be a huge hit. And I think there's enough. Actually don't know for sure, but my guess is that there's capacity for that. Yeah, I think a new device to drive. How many humane AI pins did they sell? We can actually look it up.
It wasn't a lot, but.
[00:24:53] Speaker A: Okay. Remember when AirPods came out?
[00:24:56] Speaker B: Huh?
[00:24:57] Speaker A: When like I, I remember very early on I got, I got the first generation AirPods and people literally stopped me in the Streets and asked me how it is. And there were lots of chatter online that says like, this is the dumbest thing Apple has ever, ever came out.
[00:25:14] Speaker B: AirPods.
[00:25:14] Speaker A: AirPods for sure.
And then suddenly millions of people started wearing AirPods and it became like this cool thing. Right.
I do still think you need millions of people to wear a device to make it.
[00:25:30] Speaker B: Yeah, it takes time. Okay, I got the numbers. Humane AI pin was targeting a hundred thousand.
They sold ten thousand.
It was a flop.
[00:25:41] Speaker A: Yeah, that's a flop.
[00:25:43] Speaker B: That's what I think. Like if, if this new device comes out and they sell 100, 200, 300,000, like that's a hit in the, in the hardware world. And then it takes time to get adoption. Up to like half a million. A million. That takes years to have to happen. It's just, it doesn't happen fast.
[00:25:59] Speaker A: I see, I see. That's what I see.
Well, let's see.
[00:26:04] Speaker B: Okay.
What else you want to say about the Gemini stuff?
Is there anything else we should talk about? I mean, will be powering, right? It will be powering Apple intelligence.
So I don't think they've been specific about this, but my hope is that it'll make Siri work.
Like, Siri is a disaster.
Do you ever, do you ever use it?
[00:26:31] Speaker A: No.
[00:26:31] Speaker B: I know some people that actually are like, that are good with it, but like every time I've ever used it, it just is a total fail. And I'm just so frustrated, I just give up. So if like Siri was like relatively easy to use, like the ChatGPT voice activated feature works great. Like I could talk to it, have conversations. It does what I need to do. Like if Siri achieves that level of like confidence would be fantastic. Right. And so I hope that this partnership with Gemini and Google does that.
[00:27:06] Speaker A: And also Google rolled out. We'll talk about this quickly with the personal intelligence stuff. So they're rolling this out to.
[00:27:13] Speaker B: Oh yeah, yeah. This is actually. This is important. Let me share this. Yeah, keep talking.
[00:27:18] Speaker A: So Google announced personal intelligence a couple days ago. They're rolling this out for personal Gmails, personal Google Photos, YouTube histories, and they're going to roll this all out into Gemini app so that you can chat against your personal emails, photos, YouTube, everything that Google already knows about you.
Super cool. I think it will work really well if they incorporate this into Apple devices as well. Right. So maybe they get access to Apple Health, they can get access to all of your device information, metadata.
I would love to have a conversation with my friends.
How to get more active Be more healthy, et cetera.
[00:28:00] Speaker B: Yeah. I thought Google did a good job with this thread.
[00:28:04] Speaker A: Very good job.
[00:28:06] Speaker B: Yeah. I think this is big. The other thing I wanted to mention about this is like, I think it's changed the game in terms of marketing. So if, like, your inbox now is going to be algorithmic and summarized and very few people are going to, like, use email in the capacity they used to use it.
[00:28:25] Speaker A: Yeah, I think it's.
[00:28:27] Speaker B: Yeah, it's big for email. I think it's bad for email marketing.
[00:28:34] Speaker A: Yeah.
[00:28:34] Speaker B: Yeah. I don't know what happens with, like, beehive newsletters and stuff. Do they get summarized? Does that continue to be something that people want to read? I actually don't know how that's going to work with this new Google tool Summarized.
[00:28:47] Speaker A: People are gonna stop visiting aggregator sites because you can ask questions. Show me the top 10 travel destinations, book me a flight, book hotels.
It's gonna change how we.
[00:29:01] Speaker B: Yeah, I. I think it's a. I think it's a big deal.
Have you tried it? Is it available yet?
[00:29:07] Speaker A: No, it's not available yet. I'm on the wait list.
[00:29:10] Speaker B: I'm interested. Okay, let's talk quickly about this UCP stuff.
[00:29:16] Speaker A: Let's do it.
So actually, this kind of like, plugs into what we literally just talked about. Right? So if people are.
[00:29:25] Speaker B: Another Google announcement.
[00:29:26] Speaker A: Right, Another Google announcement. They've announced something called Universal Commerce Protocol, which is a wrapper around things like MCP and a couple other protocols. What it enables is it puts a commerce layer, so checkout payments, on top of all of these AI native protocols.
All it means is that now you can interact with AI inside of Gemini, let's say. And also the Gemini will talk to things like Shopify and basically will let you purchase things directly inside of AI instead of going off the platform.
Right?
[00:30:06] Speaker B: Yeah.
[00:30:06] Speaker A: So this is huge because AI is becoming this huge aggregator of places where you can buy things. So the biggest thing that Shopify has not really ever figured out is how to create a marketplace similar to Amazon. They let people create E commerce stores. None of the E commerce stores technically are grouped together so that you can't really shop across multiple E commerce stores. Right.
[00:30:35] Speaker B: Shopify stores, they're all disconnected. They're all fragmented. Yeah, correct, correct.
[00:30:40] Speaker A: And there has been attempts to create things like the Shop app, where you can kind of search, but it doesn't have the same penetration as Amazon.
But Google, with this new API, essentially says, now, regardless of what platform you're on, you could be on Shopify, you could be on TikTok shop.
All of that gets aggregated into AI and the user can shop across all of them. That's huge.
[00:31:07] Speaker B: Yeah, it's interesting, like all the building blocks being put into place for like AI to crush and converge a lot of things into one interface.
What I don't know is like, will consumers accept this and will consumer adoption go in this direction? Google's pretty powerful and they can force it, but I don't know if people want everything into one kind of super chat. Apple maybe.
It'd be interesting to see how this plays out. I also, this UCP thing is interesting because, you know, MCP came out, we talked about MCP at one point and then it kind of like, it kind of, it kind of didn't really go anywhere. I think it was missing some elements. And like what people want to do with MCP is create a protocol so they can do agentic shopping. And now that's available. And so that that use case I described before, like, hey, get me this drivetrain cleaner from my bicycle. Like now you actually have the building blocks in place to do that. And like you said too, the implications are pretty big that it conceivably could go across every single Shopify store in one shot. And that was not something that was like easy to do.
[00:32:17] Speaker A: Yeah, exactly. So like the problem with MCP was a few folds, right? So with all of these protocols, you kind of have to convince the developers to adopt it. And the developers tried the MCP and they quickly realized that it's just not that good. So there's can't go into the details. It just uses too much token.
It's very verbose. So the cost incentives of using MCP will just weren't there. So like now basically what Google has done is, okay, let's focus on the consumer side of things. Let's not kind of talk or put too much attention on mcp. Let's just combine all of these AI native protocols into one and focus on what people actually care about, which is just buying things. Right.
[00:33:02] Speaker B: And MCP had no commerce, no commerce to it. And other people release things and there was all these like hacks and workarounds. So this is just everything you need to do shopping in one one like protocol, which is what is, I think, what people want. I'm still skeptical that this will get adoption. These things come out all the time. When there's some new technology that emerges, they don't always get adoption. So the idea that I can see why this is good for Google to some degree, I can see why this is good for Shopify. But do consumers adopt this? Do they say, yeah, this is how I want to do it.
This is great for me. It's convenient, simple and easy. That one I like, I'm not sure about at all.
[00:33:48] Speaker A: So I think the precursor to this is Google Shop. So this is where on Google if you search for a product name, there's a specific tab for shopping.
[00:33:59] Speaker B: Yeah, yeah, yeah.
[00:34:00] Speaker A: Probably had enough stats or data around the number of people who would just go to that they're visiting the actual website.
I think they have enough knowledge to know if this will be successful. Personally, I will use this.
[00:34:15] Speaker B: Yeah, I could see advantages to it if.
If I could talk to my phone and it was easy and it hurt and it like heard what I said. It wasn't just totally confused by it. And like, so you got. So you have to have a lot of elements, right? Voice chat needs to work relatively well. It needs to be able to connect to some of my like Amazon account or Shopify account or something like that.
[00:34:36] Speaker A: Right.
[00:34:37] Speaker B: But I could just speak to it, say like, order me my coffee that I love or order me like these things that I need and I could just make that happen in like a flash like that. I didn't have to like go and text and type and open up things like that would be really convenient. And so I don't know if this unlocks that, but I could see a future where it is. It is valuable.
[00:34:58] Speaker A: Right. So. So that's going to be super convenient. The second thing I'm just kind of going to the weeds now is, you know, you still kind of have to like sign up to things on every single e commerce website. You have to.
[00:35:11] Speaker B: Yeah. You got an account everywhere.
[00:35:12] Speaker A: Absolutely everywhere. I have like probably a thousand accounts on different e commerce sites, like coffee shops that I've been to once. And then like somehow I sign up to the royalty program, so I occasionally get emails from them.
Hopefully that all goes away.
[00:35:30] Speaker B: That is like, that's an old one, man. That's like. That's like back at the like Live Journal days when I was at LiveJournal.
I mean, did that people that problem because like the founder of Live Journal, Brad Fitzpatrick, he's a bit of a legend in the nerd world, but that was something that he was trying to solve and I think what was this protocol called Open id.
He had a universal login concept. Yeah, exactly described.
[00:36:01] Speaker A: Right.
[00:36:02] Speaker B: Don't you remember Facebook Connect?
[00:36:03] Speaker A: Totally, totally. Yeah. Yeah.
[00:36:05] Speaker B: It's not around anymore at all. Right. There's no Facebook Connect now.
[00:36:08] Speaker A: Social logging was kind of like the thing that we settled on.
WordPress had that like one WordPress.
[00:36:17] Speaker B: I mean, so they're there now you have the login with Google, login with Apple.
[00:36:21] Speaker A: Right.
[00:36:21] Speaker B: Is there, is there a third one? I only use those two sometimes.
[00:36:26] Speaker A: What happened?
[00:36:26] Speaker B: I actually don't even know what happened. The Facebook connect one, it used to be everywhere. That was like a long time ago. And then it just kind of like people I remember having customers were like super frustrated. They're like, no one, no one uses. Everyone doesn't want to long on Facebook. And then I guess they just got rid of it.
[00:36:42] Speaker A: Yeah, totally. Because people stop using Facebook. So like WordPress has one.
[00:36:49] Speaker B: There's a login with WordPress. Yeah, I would never do that.
[00:36:51] Speaker A: Yeah, that was the biggest one.
[00:36:54] Speaker B: My brother, my brother sent me some joke. It's like a screen with like 300 logins and it's like login with only fans log in with like. Yeah, I was like, I'm just gonna go all only fans now. Just trust all my, all my personal data that like login with OnlyFans. All right, all right, let's, let's move on here. I got something I'm very excited to talk about.
[00:37:13] Speaker A: Let's do it.
[00:37:14] Speaker B: So this is the five year stock chart of Tesla. Is there a way I can make this bigger? I don't know. They change this interface and I'm not excited about it, but there we go. Okay, perfect. There we go. Now we're talking. So this is the, this is the five year Tesla chart and it's close to like, not exactly, but it's close to an all time high, which I was surprised. I look at this the other day and I was like, Tesla, you know, they, they had this peak back in 2019 and it just has been like a long decline. That 2023, like that real crash right there, that's around like the.
Was that the Twitter purchase, right? Or the Twitter like. Or the. I mean there. This stock has been through some tough times and I sold it like a long time ago and really was just kind of, it was just not on my radar as Tesla sales declined. So like they're not selling as many cars as they were and they still sell cars, but like it's not a growth stock or growth business from like a sales perspective, which was when I kind of like stopped paying attention at it as like an investment or stock. But obviously shareholders like someone's buying this and they believe that the Optimus product will work. The robo, the Robo Taxi. Right. There's a whole bunch of stuff that Elon's been promising that obviously someone believes in. But what I thought was really important about this, and then I'll let you get a word in. But what I thought was really important about this is that the stock is so high that it possibly means there will be no SpaceX IPO, and that Tesla and SpaceX will just merge. If the stock was in the toilet, not doing well, I just don't think it'd be possible for that to happen. But given the strength of the Tesla stock, I think that that is a possible scenario. And I've heard other people say this too, that SpaceX actually will not go public. It'll all get crushed into Tesla stocks. And Tesla will represent Xai AI company, it'll represent Twitter, it'll represent Tesla the car company, it'll represent SpaceX, the space company, and it'll also represent Starlink.
Right. So you've got like this industrial conglomerate like no other, which in spite of all its, like, flaws and challenges, which there, there are many, it's still an incredible company. If all of that comes together, I guess is kind of like how, how I would position this. What, what do you think about this?
[00:40:18] Speaker A: I think that's a really smart way of putting it. If they just aggregate everything, kind of follow what Alphabet did with Google products.
Now the question is, do you think it's still a discount? Would you buy today?
Would you buy Tesla?
[00:40:34] Speaker B: Oh, my God. That's a great question.
Because I looked at it today, I was like, should I buy this?
And I couldn't pull the trigger, at least not today. I was like, I just, it's, it's expensive, it's purely speculative that it'll merge with SpaceX. I think that might be partially what's driving some of this, is that there's definitely a group of people who think that that's going to happen.
So I haven't bought it yet. It's on my radar. I think it's interesting.
I like the idea of like this massive conglomerate of modern industrial companies. I also think it's like, perhaps important to point out that, like, you know, Elon really saw his brand.
[00:41:20] Speaker A: Let's.
[00:41:21] Speaker B: I'm gonna use word like explode during the election.
Like, it's so secret that I'm a big fan of Elon, big supporter. I'm on X. I think all this stuff is for the most part great. But, like, that was, like, that was not smart. Like, like it was not smart and it really hurt him really bad. And I'm glad to see that he's, you know, he's still involved, and that's his right to be involved. But, like, he's taking a step back. He doesn't have to, like, personally appear on stage with Trump. Like, that's just not something I think I would not, not recommend it. So I think it's good that he's no longer doing those things. And I would like Elon to just continue to focus on, you know, building rockets, catching them out of the sky and sending things to Mars. Like, that's what I think is really cool. Oh, and launching satellites in Starlink.
[00:42:13] Speaker A: He's got too many things to be distracted by Paul.
[00:42:19] Speaker B: I think it's even important to point out because we always get lots of comments when we talk about Elon. Some of them are positive, some of them are negative. I think it's important I understand why he got involved in politics too. They really went after him in a big way. And I would say I would they force his hand. But still, like, his, just the way he put himself personally, like, like with Doge and all that, I just think that was a huge mistake.
[00:42:45] Speaker A: Yeah.
[00:42:46] Speaker B: Distraction. Just focus on doing cool stuff, being an innovator, inventing things, like just pushing the technology. It's so impressive. Just, just do that.
[00:42:57] Speaker A: Yeah. He, he just wants to fix things. And then I think the biggest lesson out of this is that he realized the government's not fixable.
[00:43:07] Speaker B: So.
[00:43:07] Speaker A: Awesome.
[00:43:09] Speaker B: I've made this point many times on X.
I think it shows the limits to some degree of, like, this very right brain, very, like, engineering, very quantitative approach. I like to joke that, like, he couldn't really fix Twitter because it turns out that, like, advertising and marketing is harder than rocket science.
[00:43:32] Speaker A: That's right. That's right.
Or somebody like, that's extremely engineering focused.
[00:43:39] Speaker B: It's just, it's not the problem that can be solved in an engineering way. Right. Like, and, and just the damage he's done to, like, the brand and the reputation.
[00:43:47] Speaker A: Right.
[00:43:47] Speaker B: Is just terrible.
[00:43:50] Speaker A: You can't put a handful of really, really smart people into a room and, like, fix all you need hundreds and thousands of people.
I, I just, like, chip away at.
[00:44:04] Speaker B: I, I know people have found opinions on this. And so that's fine. And that's fine. We'll get, and we'll get comments. I, I, I, I don't, I don't.
[00:44:14] Speaker A: We'll give YouTube a sound bite. Elon Musk cannot fix the government.
[00:44:21] Speaker B: Yeah, well, I was thinking about X. Right. Like, they've made all These changes on X, and I don't know if it improved the experience at all. Like, I, I find it to be really painful to be on there still. And I can't give up. Been on a long time. I met a lot of people on there. I met you on there. This, this show started on X. Like, I've got a lot of, like, value and a lot of, like, friends, and a lot of things for me have happened on X, but, like, it can be like, really painful to be on that, that app.
And I struggle. I struggle with this all the time.
[00:44:52] Speaker A: Totally, totally.
[00:44:53] Speaker B: I.
[00:44:55] Speaker A: What I've noticed is really just X cannot make up its mind to help small accounts grow or retain big accounts.
So, like, people are just saying, like, why am I not seeing my followers or my following's content?
And you look at some of the big accounts, they follow literally tens of thousands of people. So it doesn't make sense for them to see the people that they follow.
Right. And then the algorithm just shifts constantly back and forth between helping small accounts grow versus helping big accounts get retained exposure.
It's hard problem to solve.
[00:45:38] Speaker B: Yeah.
I don't actually. I love to pretend like I know the answer to fixing X, but I actually don't think I know.
I.
I always feel like going back to follower count mattering is better. I don't. Like, it's 100% algorithmic. It doesn't matter who you follow or I follow you. And it just like, tweets pop for no reason.
That part I just find, like, it's. It's like kind of like gambling. I don't like that at all. I wish there's more predictability to, like, how things work. And if you had more people following you, you would get more reach. It's very logical.
That's how social media work. In the beginning. You could attract more followers. That's why they're called followers.
They follow you to read your content. So, like, in my mind, like, some form of, like, that being influential on, like, how the content and how the fee gets doled out is, I think, how, I guess all social media should work. Like, I think it should go back to some form of that the idea, like, follower counts don't necessarily matter.
I, I think, I think sucks.
That's my, that's, that's my statement. How's that?
[00:46:49] Speaker A: Yeah, totally. And then I don't have the answer.
[00:46:53] Speaker B: They'll figure, I mean, like, I don't know. One knows. I do think Nikita is doing a good job. That's a hard job. I think he's really bold for taking that on too, to be had a product at, at Twitter. He's got the right attitude. So it, I, I don't know. I, I, he's got the right approach. Like, he, he's doing a good job. It's just a really hard job. It's a really hard product to.
[00:47:18] Speaker A: Yeah, Thankless too.
You're always going to piss off somebody. You know, like, even if you get it right, you're gonna, somebody's not gonna be happy.
[00:47:27] Speaker B: Oh, dude.
I mean, the, the, the, like online discourse between him and Elon is hilarious. Like, that's like your boss, like you're like debating out in the open on Twitter. Yeah. It's, that guy's like, all right, let's move on, let's move on. Let's talk about, like, just quickly. I don't want to go too far into this stuff, but.
[00:47:44] Speaker A: Yeah, what's going on?
[00:47:45] Speaker B: Yeah, we should talk about the meta layoffs a little bit.
There's a lot of layoffs happening in general, actually, not just in high tech. There was even like a bunch of layoffs at Citibank and stuff. So companies are still like. Right. Sizing. I think they overhire during the pandemic and so they're still trying to sort this out.
But I think what's ironic about this one is like, did, like, did Zuckerberg rename the company for like a completely failed product?
[00:48:14] Speaker A: Yes.
[00:48:15] Speaker B: Like, who uses this stuff?
[00:48:17] Speaker A: Metaverse.
[00:48:19] Speaker B: No one goes in the metaverse.
I have no interest in the metaverse at all. I do. Like I've joked about before, I do think augmented reality glasses are a cool idea. I read about them and I get excited and then I stop reading about it and I go do something else and I completely forget about them. But I do think there's a part of me that would get value out of like wearing glasses that had some kind of Internet access. But the 3D, like fake metaverse, like, I have zero interest.
[00:48:54] Speaker A: Yeah, I think they just completely dropped the ball on the go to market strategy with like, they couldn't ship enough units.
It's like, it's like the new Apple device as well. Right. They need to get a hundred thousand shipped, then half a million, then maybe a million.
[00:49:11] Speaker B: They can't ship them because no one wants it. No, no. To go in this like cornball 3D universe that like looks like a video game from like 1995. No one wants it.
[00:49:22] Speaker A: And also Zuckerberg was like the face of the device every time you saw a picture of the metaverse it had his, like, pasty face.
Like, get somebody that's like, I don't know, get Sweeney.
Sydney Sweeney in the metaverse.
[00:49:41] Speaker B: Okay, that. That making it sound more interesting. I'm down for that.
[00:49:44] Speaker A: But.
[00:49:45] Speaker B: But that brings up a great point. I remember when the metaverse, like, first launched and there's all that, like, harassment in the metaverse and stuff. Like, dude, I just think the whole thing doesn't work. Like, look at the way people get harassed and, like, insanity of Twitter and add like a 3D layer to it.
[00:50:05] Speaker A: Add a 3D layer to it.
[00:50:09] Speaker B: It's a terrible idea. It's like, I don't think there's any way, like, what kind of guardrails would you put in there? Like, would you literally prevent people from, like.
[00:50:19] Speaker A: Yeah, like, lewd gestures?
[00:50:22] Speaker B: Like, just take all the stuff people can do. Right? It's like, how do you police any of that?
[00:50:28] Speaker A: I.
[00:50:28] Speaker B: It's like, I actually never even thought that deeply about it because the whole thing is so, in my opinion, obviously ridiculous.
But, like, you're gonna attract a bunch of, like, 14 to 18 year old boys, and in between, like, shooting each other on Call of Duty, they're gonna go into the metaphors and chat. Like, what do you think is gonna happen? Like, like, it's really obvious to me, like, what I would have done when I was 15 when I would go into, like, the metaverse or something. So I just don't see any way that this thing.
I think it's a total failure.
[00:51:03] Speaker A: I mean. Or you literally just go sell this to 15 to 18 year olds. Right, but that's how Facebook started. Right? Like, that's literally how Facebook started, which is no one.
[00:51:15] Speaker B: No one wants it. No one wants it.
No one wants it. I think. So that.
[00:51:21] Speaker A: What?
[00:51:21] Speaker B: Dude, so did I told you about this? Like, I remember, like, the first virtual world boom of, like, Second Life, and there are actually a bunch of other ones.
[00:51:29] Speaker A: Yeah, that's right.
[00:51:30] Speaker B: And we were, like, trying to sell these, like, sponsorships and we sold video game. Yeah, but the video game people bought sponsorships because it was like, it aligned really well with their audience. Whatever. The video game people were like, this is just a video game with, like, no, nothing to do. And I think they actually kind of nailed the problem with, like, metaverses. Like, we have a thing called games. And you could play these 3D 3D games. You could play Grand Theft Auto and Call of Duty and. And you could run around, like, go on missions and shoot stuff and do
[00:52:00] Speaker A: all kinds of crazy looking, wild stuff.
[00:52:03] Speaker B: You Go into the metaverse and there's nothing happening. Like, you have to like, invent your own narrative. It's. It's a complete bore. Like, it's not interesting to people.
[00:52:13] Speaker A: Steve Jobs solved this with the App Store, right? Like when iPhone first released. What the heck did you.
[00:52:21] Speaker B: All right, we're done talking about that. I'm done. The metaverse is failure. He renamed it. He should go back to calling the company like Facebook or something.
[00:52:28] Speaker A: Facebook. Go back to Facebook.
[00:52:30] Speaker B: I think so. Okay. All right, the next one really quickly. I'm very excited about this one. I'm not going to spend too much time on it, but I did want to show that, like, well, according to this data, I buy the narrative. I buy this narrative. Everybody and their grandma lined up and said, oh, it's gonna not work. And even piketty and all these famous economists were like, it's a disaster, blah, blah. And they are all wrong. At least wrong based on the data that we have today in this timeframe.
And I would just make the point that PhD economists don't know everything.
I, I buy this.
[00:53:08] Speaker A: I think just a context for potential listeners. It's. We're talking about Argentina's economy under Malay. Right? Despise what all the economists has predicted.
[00:53:23] Speaker B: I mean, this thing is pretty. Thomas Pike, I think, right? 108 economists signed a letter saying they did not support. I mean, like, it's not like they just made some comments on a, on a podcast like, hey, do you think this is going to work out? It's like, I don't know, it's not going to work. Right. They literally like got together, signed a letter saying that they thought this was a failure and.
[00:53:50] Speaker A: Right.
[00:53:50] Speaker B: They've been proven wrong. And I, for me, that part is like really important. I also think that it just shows like, where the economics world is today. I think they're totally disconnected from reality. But that's a popular criticism of the economics profession for many years. I'm not the first person to say that. All right, let's stop with that one. Do we have anything else in our doc?
[00:54:21] Speaker A: Maybe we'll show this agencraft.
[00:54:25] Speaker B: Okay, yeah, you do that one and then we'll call it a day.
[00:54:28] Speaker A: This is literally my childhood dream where RTs, real time strategy, video games and orchestrate agents and be productive.
I think we're not very far away from this. Where are we?
[00:54:45] Speaker B: You sharing it? You get it? Okay. Yeah.
[00:54:47] Speaker A: Okay. So basically this person has built something called agentcraft. It's an orchestration layer for all of your agents inside of an interface that looks like RTS games, which is a real time strategy games like Starcraft, for example. I, I'm.
[00:55:05] Speaker B: What are these agents? What are these agents doing?
[00:55:07] Speaker A: So they're, they're literally, for example, like clock hold, right? This is just providing an interface on top of the clis and essentially you can use this interface to give commands.
[00:55:23] Speaker B: So I could have them go buy me drivetrain bike cleaner.
Is that what you're telling me? Yeah, if I wanted.
[00:55:31] Speaker A: It could literally be anything. Right?
Because if you really think about agent orchestration, at the core is you're gathering inputs so like feedback loops, locks, new piece of data, and then you're transforming them into some kind of output. And what you really, really should do is you want to give them a goal to work towards. Right. So all video games has an end state. All video games has a winnable, winnable state, a winning state. So like games plus agents are actually the perfect combination of interface and how you should think about orchestrating agents. Because you're multitasking, you're thinking about different things. You're thinking about resource. Resource should be like, you know, the amount of time you have, the amount of resources you have, and then you're also managing two, three different things at the same time.
[00:56:34] Speaker B: Look, Paul, when you, when you, when you vibe code interface for an agentic marketing CRM that looks like this, I'll be the first to sign up.
All right. All right.
[00:56:44] Speaker A: I'll make it happen. We'll make it happen.
[00:56:47] Speaker B: All right, that's the show for today.
[00:56:51] Speaker A: We're good.
Any memes or any elon.
[00:56:56] Speaker B: I think we're good. We'll just end on that one.
[00:56:58] Speaker A: Okay. All right. All right, dude, next week.
[00:57:01] Speaker B: Take care.
[00:57:02] Speaker A: Bye. Bye. Bye.
