Sora causes Adobe to lose $18B in 1 day, Sora's impact on the marketing industry, and more

OpenAI introduces Sora and it's going to shake up the marketing world like crazy. And it's already shaken up the business world like crazy. Adobe stock in one day lost $18 billion in market cap because of Sora. I don't know if you saw that.

I didn't see that. Weren't you— you were just looking at Adobe stock or you saw a headline somewhere? CNBC?

I looked at Adobe stock and it tanked right when Sora got announced. Like literally $18 billion.

What percent was that?

7-ish percent.

Wow. 7.41, I think.

7.4.

And this is not an earnings-related thing?

No.

Wow.

Because OpenAI is now going into Adobe space.

Yeah. Yeah. Nothing is safe.

Nothing is safe.

But also what we talked about yesterday at dinner, we're having dinner with one of your VC friends, and I just— I think the phrase here is AI is overhyped in the short term. It's underhyped in the long term. That's what I believe.

Yeah, I'm with you on that. Because even when you look at— so for anyone who hasn't seen it, it can create a 1-minute video for you.

Yeah. And this is OpenAI's Or I just serve a prompt.

Yes. So you type in a prompt just like you would on ChatGPT and it can create a video for you that is a minute long. The hard part with AI, no matter how good the video quality looks, and it looks really cool and amazing, there's a few things that happens. One, if anyone's tried to create AI-based videos or images and put that on their social, like through Midjourney, you know, using all their images, most people don't see crazy engagement. These images and videos aren't doing that well from a sharing, like perspective, It's cool and novel right now because AI can create it for you, but no one really cares to engage with it as much as we're seeing compared to human-based or human-created content.

You know what the problem is with generative AI? It's like when you ask DALL-E or Midjourney to make something, you actually can't edit it. So if you say, hey, they make something for you, it's like, hey, can you change this for me? Or like, I like this, but change the shape of this over. It actually can't do it. It actually makes a brand new one for you.

Yes.

So basically what's going to happen, I believe, in the future is that you're going to have a situation where it'll create the initial one. And then it might push over to like a model, like an actual engine where you can, you can edit things, right? So it might create the initial like batch for you or the initial like inspiration, which will save you a lot of money, the concept, right? And then you can build on top of it. Right now you just can't do that, which is, which is why it's not that useful at the moment.

Dude, totally. But even when you're able to edit it, the way I see it is this will take out a lot of people who are doing production, but you're going to still need people to come up with creative ideas for the content.

Yeah, it just comes down to strategy at the end of the day. Whoever's good at strategy, and basically that means whoever's good at continuing to learn and continuing to practice, those are the ones that are actually going to stand out. The rest of the people are unfortunately not going to have jobs.

Dude, funny enough, speaking of jobs, I was speaking to a group of people a few days ago and they were talking about AI and how they can make a ton of money from AI. A lot of these were college grads and I was talking to them virtually. But when I was talking to them, they were just like, yeah, I want to create something like OpenAI or ChatGPT or Midjourney. And the first thing I ended up telling them is—

Are these people smart?

They were smart.

Yeah.

I was like, you're actually looking in the wrong place. If you want to make money from AI, it's really hard because to create a ChatGPT or OpenAI or any one of these projects, it costs a ton of money and you have to have really, really, really talented engineers. It's not easy at all.

These aren't even half-million engineers. These are like million-dollar-plus engineers. Yes.

And for every one of those that is created like a ChatGPT, there's thousands that failed and billions and billions that were Dude, look at Jasper. Do you remember when Jasper came out?

Yeah, Jasper. They raised a bunch of money from David Sacks' Craft Ventures. I think they were valued at $1.25 billion. Do you want to describe what it does?

Yeah, so Jasper is an AI content writer. How many people really use Jasper right now to write AI content?

I think their churn has to be through the roof. And I like the team a lot, right? I think they got in really early, like a year or 2 before ChatGPT came out. Once ChatGPT came out, it's like, They're like, oh crap, right? What do we do? But I did talk to them last year when we went to HubSpot Inbound. I talked— the Jasper team actually sponsored there. I was like, hey, so like with all the stuff going on, like how are you guys pivoting? How are you adjusting? And they're like, well, we're just focusing more on B2B now. But the problem with that is like OpenAI also has a B2B version too. There's a Teams version that I pay for. So—

Dude, and the way I look at it is like, why the heck do I need Jasper? I need a lot of these tools out there.

Yeah.

And I believe that if people want to make money through AI, The best thing for them to do is be a prompt engineer and be good at it. And when I say prompt engineer, I'm not talking about someone who knows how to code. I'm talking about someone who knows how to modify the prompts to get the output that's desired. Because if you give ChatGPT or anything a crappy prompt, you're going to get a crappy output. Even if you give them a good prompt, you still may get a crappy output. But being a prompt engineer is a large part of the future because how can you continually get what you want out of AI? And according to Forbes, there's a lot of jobs that are paying prompt engineers $300,000+ a year.

Which company is paying $300,000 a year for a prompt engineer?

I don't know, but that's what Forbes is saying.

That's crazy. I mean, I believe it. It's crazy. It's how fast that's going already. And when you think about a lot of these SaaS tools that are AI, it's just prompt engineering in the background, right? There's a lot of these AI writers and things like that. A lot of the SEO writers like Byword or like there's a couple others that I've seen, they're all just prompt engineering with a SaaS like UX on top of it.

And Funny enough, a lot of them are for sale because I get hit up by so many of them.

Oh yeah, the churn is through the roof.

Churn is through the roof.

Like one guy told me, oh, I was like, what's your churn right now? It's like, oh, it's pretty good. I was like, what's pretty good? It's like, oh, it's 7% a month. I was like, that's not pretty good.

Yeah. It's just like your customers stay for roughly a year.

I think you lose 70 to 80% of your customers in a year if it's 7%.

Yeah, roughly a year.

Yeah.

So it's just like if you can't keep your customers for 4 or 5 years, it's really tough. And ideally these days in SaaS, investors don't want people who keep their customers for 4 or 5 years. They want a snowflake type of company. We're like, well, if you turn off sales and marketing with the upsells that you get, even with the losses, the company keeps growing.

Yeah, I think the other thing too is talking about agencies, agency churn too. Agencies like last year, what we've seen, what we talked about is like churn is like 50 to 70%. Yeah, but, but, but like you made a good argument. It's like, well, here's how you actually reduce churn. What is that argument?

Yeah, the way that you ideally reduce churn is you sell multiple services through to your customers, they stay way longer.

Yeah. So this concept is known as switching costs, right? If they have a bunch of services that are going pretty well, they're not likely to want to switch. Like, even if someone offers something better in one area, you have the whole suite, so you're likely to keep them longer. So it just comes down to having a good product and a good service at the end of the day. That's what it is. Nobody cares, like, how you're wrapping it.

No, but what's funny is, is when you look at corporations, they don't care if ChatGPT can do something for free or another SaaS tool charges $100 a month. Like, if you're a big enough company, it's all rounding errors. They'll just Pay for whatever's more convenient for them.

By the way, Neil and I have an agency owners group called the Agency Owners Association. All you have to do, just go to marketingschool.io/agency. Once again, it's marketingschool.io/agency to learn more. And now back to the show. There's an argument that YouTube is actually going to become the number one search engine in the long term. And the reason for that is because YouTube has so much new content, so much new data that previously wasn't really crawlable by search engines. But I don't know if you've seen this video from Gemini Pro. It's called Gemini 1.5. So Bard is now named Gemini, and there's a 1.5 version coming out. You probably haven't seen this video, right?

Uh-uh.

So basically what this video— this AI does now is you can show it like a 44-minute video and then say, hey, tell me what happened at 44 minutes. Like, what exactly is happening? What is being displayed? And it shows exactly, hey, this is what the note says over here. This is what happened in this moment. And then you could also then draw a picture and say, hey, tell me the moment where this happened. So someone drew like a stick figure picture of like people falling out of like a water bucket or something. And it's like this happened at minute 15, 15 minutes and 34 seconds. And so Gemini is becoming multimodal now and you're going to be able to crawl all of YouTube. And so I think in the long, long term, it makes a lot more sense to just continue investing in YouTube, right? I think it's just getting stronger and stronger and that's their data moat that they have.

Yeah, dude, I can see YouTube still being really popular. I also see Google being popular for the next 10 years. Yeah, there's a lot of use cases for why people keep using Google and they have 8.5 billion daily searches like it is a powerhouse. A lot of people take it for granted.

Yeah.

The other thing that you'll notice too is a lot of the older demographic is slower to switch. You know, they're stuck in their ways. It's a younger demographic that we're seeing that is starting to use like the ChatGPT first or the TikToks first or a lot of these tools. And a lot of the older demographic is where the money is. If you think about it, they're the ones who own the homes, they bought them for pennies on the dollar. It was easier during their time to come up because things were cheaper.

Yeah. Look, I think just, just this is really for the SEOs. It makes a lot of sense to look at YouTube. We actually get a lot of reach-outs. I'm sure you do too, a lot about people looking to optimize YouTube. And we're— dude, by the way, I don't know if you've seen our YouTube channel, our YouTube channel, like we're getting like 5,000 views this video over here, 7,000 over here, 1,000. And that is—

Your YouTube or our YouTube?

Our YouTube. We have our marketing school one, not, not the one with 8,300 subscribers, the one that has less. So not MS Archive, the other one.

Not MS Archive. Okay, what's the other one?

It's just type Marketing School and it should be like the one with 1,500 subscribers or something. But as Neil looks for it right now, my main point is this is actually— we actually, by shutting down our other channel or stopping the other channel, we're getting more views now. So I had a thesis, I had a hypothesis that our other channel was in YouTube jail. And so because each video is only getting like 100 views or whatever, right? And so we started this new one and it has like no subscribers, but, you know, some Shorts will get like, you know, a couple thousand views or so for channel and then the long-form videos will get a couple thousand views too. And so, that makes a good argument that if you feel like you're in YouTube jail, maybe just go start a new channel.

You could. Or what I've also found too is if you continually produce good content and you stop producing crap and you don't produce as much but every one is a hit, over time, you just get out of YouTube jail and you start getting tons of traffic.

Yeah, I think both are fine. I think we just switched the quality of ours and then we just made a new channel and it's performing a lot better than the old one. I mean, that's like 70 times more.

So cool.

What do you— I think going back to Sora real quick, right, we talk about how we should talk about the risk of AI companies, right? So Jasper, right? So you can't—

Yeah.

I think it's— if you're going to build an AI company, because the problem with marketers right now, here's an AI mastermind. Oh, I'm an AI influencer now. And then they're on magazines and stuff, right? It's like, it's like, no, like, and then they're like, oh, I'm building this AI software. Like, there's a lot of people you and I know, it's like they're just building this, but they're just wrappers at the end of the day. Like, where you're going to make the money is like probably building, building like a proprietary proprietary LLM or the picks and shovels, right? Or the GPUs, right?

It's the people who are building— yes, like what you're saying. Or you can sum it up with the platform plays, right? The people who are actually building the technology that you need. A lot of people need Nvidia's chips to do anything. That goes back to the picks and the shovels, like Levi's, right? They made a killing during the Gold Rush. Same goes with the platform. If you look at OpenAI and what they've built, you can build a lot of stuff off of it. And at the end of the day, the one that keeps making money from it is OpenAI.

Would you— okay, let's, let's, let's, let's do this fun exercise. So if you were to invest, you can only buy and hold in an AI stock for the next 10 years. If you picked one, what would it be and why?

Microsoft. I think they're leading the way.

And so why, why, why are they leading the way?

They're the ones who funded OpenAI with tons of server resources when no one else would. And their whole play was we want to get people onto our ecosystem.

Yeah.

And I think, you know, they have a really big head start.

Okay. So that's your number one. Do you have any others that you would throw into like your buy and hold for 10 years?

I would just do concentrated bet and go all in on Microsoft.

Okay.

I'm not giving financial advice, by the way, but that's what I would do. I think Nvidia will have a lot of competition and so will Microsoft and stuff like that. But how many times do you need to keep buying chips? Right. I'm not saying it's a dying business. These chip companies are massive. What you do need is you need to keep paying for the software each and every single day. Chip is a one-time expense. I get you may need more and more chips over time, but the software, you keep paying for that every single month.

Yeah, I think the cost of compute— I mean, we're not engineers either, right? But I think the cost of compute is continually going to come down and the resource drain is going to continue to come down. But I do think that we talked about this yesterday, that Nvidia is probably going to have a run for maybe the next 5, 10 years or so. And so, As the world shifts. But my bet too, I guess it would be boring, it would be Microsoft. And I'll give you some other reasons as to why. One, they own— nobody really uses Skype anymore, but they have Skype.

Dude, I use Skype.

You still use Skype?

A lot of people use Skype.

Neil used to call me on Skype all the time. I'm like, who calls me on Skype? And the only reason—

A lot of people use Skype still.

How many billion people use— can you check how many billion people still use it?

How many people use Skype?

2023.

Skype has 300 million active users.

That's a lot.

That's a lot.

So look, they own— Microsoft owns—

More than 40 million people use Skype daily.

Daily. Daily act— that's 40 million DAU, guys.

Yes.

Um, but to, to end it on this, so Microsoft owns LinkedIn, right? What— they own Xbox too, right? Oh, guess what? Who uses Excel? Oh, we do. We still use Excel, right? So like the Office suite, Teams, and everything, they have lock-in. And really, they're not under antitrust scrutiny because they went through all that in the '90s or so. Right. And so I think Microsoft's well positioned here and they've made a lot of good moves. The 50— they own, what, 49% of OpenAI, something like that?

Something like that. But yeah, who knows how the ownership works out? But OpenAI is very reliant on Microsoft servers.

Yeah. So look, I think that's probably a safe bet. I think, you know, the other ones would probably be Google, Meta, Amazon. But, you know, we're only going to pick one here. So that's it. This is not financial advice, everyone. Please don't forget to rate, view, subscribe. It helps us grow. Also, don't forget, if you're an agency owner, go to marketingschool.io/agency to find out more about our agency community that we do, and we'll see you tomorrow.