Meta's AI Experiment Failed Against Reddit's Real Advantage

got a lot to get to in your in your realm here, but let's let's start on Apple and and where you think we are in terms of the the hope and hype cycle around September or anywhere else. >> Yeah. So, um really for this stock to work, they need a hardware replacement cycle. So, I think people are um hoping that the foldable phone creates a hardware replacement cycle. My guess it will be delayed. It won't actually launch on time in September. Um but and I am really worried about China and I'm really worried about component increases like memory and so that means prices are about to go up on new iPhones. So and and you know they've leased their AI life to Google which is their biggest competitor on Android by using Gemini and AI models created by they're really their only competitor in their core business all of which and we're hitting all-time high valuations. So I think the riskreward here on Apple is very tenuous. I would I would look elsewhere for investments here instead. >> It's a good reminder that Apple certainly has held its valuation better than a lot of the other multi-trillion dollar names uh that we look at. Now, one of those, Meta, we talk about its strength recently. It has been a real underperformer before that, but it did at least on the surface look cheap, but you don't uh believe that this is a time to to really uh get in there in an aggressive way. Why is that? This morning we published a note saying we prefer Reddit which is a $ 35 billion market cap to um Meta which is 1.7 trillion. And basically what we were looking at is Meta's revenue growth has been 23% total over the last two years versus three times faster for Reddit. Its margins are expanding much faster at Reddit than at at Meta. And the strategy diffusion at Meta is a value destroyer. every quarter Mark Zuckerberg comes out and he's laying off some huge percent of his workforce and pivoting to a new strategy. Latest last week was we're going to do Iris um with chips. He's going into competition now with Nvidia. He's already competing with Amazon in his announcement a month ago having to do with shopping agents on WhatsApp. He's competing with all the LLMs and he's just opened his first proprietary LLM um versus Llama which was all open. He's competing with Google search of course and he's competing with Apple right with his Quest and Rayban glasses trying to displace that smartphone platform. So um not to mention Reels and Threads which compete with you know X and they compete with YouTube and Tik Tok. So he's like in competition with five other huge companies core business and now he's decided he's going to go do chips and compete with Nvidia too. So the strategy diffusion here is very troubling, especially since it's the smallest hyperscaler among all of these competitors that he's picking a fight with. >> Do you read that uh whatever you call it a shotgun approach? Maybe they're just laying a lot of bets trying to to sort of get a piece of the next thing. Do you read that as implicitly Zuckerberg having deep concern about the sustainability of the the moat around his core business? >> Yes, I do. I think there's, you know, I think if if generative AI is as disruptive as Wall Street is funding, like we're going to fund 700 billion in um infrastructure buildout, it means that the entire web and the entire open web might be obsolete. That's possible. But the notion that what you do is every quarter look around, see what other people are doing, and go do that, which is a reactive instead of a sort of thoughtful leading singlepurpose reaction, um I don't think we'll be successful. And just a quick word on why you believe Reddit is going to uh actually be able to find its way in this new world. >> Right. So what's differentiated about Reddit is it 100% human content and what we've seen is they have deals with both Gemini which is Google and Open AAI which is Sam Alman and we expect them to renew their deals with Reddit at higher price points than the three year ago deal they originally went into. Because what we see with humans is after a LLM gives a human a recommendation, they then go to Reddit, which is a top three-sided firm by all LLMs, to see what humans are saying about that same purchase or that same advice. So that there's an authentication layer by humans that algorithms prefer actually and humans prefer. So I think there's going to be I think every LLM because there's going to be commoditization of algorithms and LLMs, lots of people competing in that space. I think human content becomes the