Transcription
I've been saying it on this channel now for months. Apple is about to burst the AI bubble. I know many of you disagree with me on that, but if you stick around, I'll explain why, and I'll show you a medium article from a guy named Michael Malawitz that makes almost exactly the same argument as I've been making, just from a different angle.
For 3 years, social media has framed Apple as an AI failure. Meanwhile, they've lapped up the hype from open AI and anthropic fear, doom and gloom, mythical magical models, and technology that's supposedly too dangerous to release. But what have we got to show for it? There's been hundreds of billions of dollars spent on training and data centers. So, where are these breakthroughs in education, healthcare, scientific discovery that we were promised? All we see is low value AI slop polluting our social media streams.
More significantly though, almost nobody is paying for it. Consumers are very happy to use chat GPT and Claude while it's free, but very few are putting their hands in their pockets to pay for it. And it's no different in corporate. Many businesses are scrapping or deferring their internal AI projects because the economics just don't work. It's a bubble and somewhere there's a pin.
Up until recently, I felt like a lone voice on this. But a few days ago, I came across Michael Malawitz's piece on Medium and his argument is almost identical to mine. He says that Apple has already won and they didn't win by building the biggest model. They won by giving people the hardware that makes AI cheap, fast, and secure.
90% of the AI workload inside a normal business is not frontier reasoning. It's classification, summarization, extraction, and just looking things up. And for that, we don't need a model trained on hundreds of languages, the complete works of Shakespeare, and a million Reddit posts. We just need something good enough to do the job. And for that kind of work, a 7 billion parameter model running locally on an M series chip isn't just good enough, it's actually better. It's faster, it's cheaper, and safer than sending off your data to some data center in Virginia.
Now, the industry talks about the Magnificent 7, but only six of them signed up to the AI cloud model. Apple didn't. And this is important because the AI cloud cult has bet everything on one strategy. massive data centers or warehouses of NVIDIA GPUs which suck electricity guzzle water and serve LLMs by the token. And to commercialize that model, they use the same SAS playbook that's worked for 20 years. There's just one problem. It didn't work. Most AI users just don't receive value to warrant paying money for it, even when it's heavily subsidized. So, the only thing that's keeping this alive is hype and fear.
Meanwhile, Apple stuck to their core principle. They want people to own the technology, not rent it. And when your device stays your device, it's not harvested or sold to advertisers. So, as time goes on and open- source models get smaller and Apple silicon gets more powerful, the collapse of the cloud AI model is inevitable. Stocks are already down 20% in the last few months. Investors are getting nervous and people are starting to wonder what is the pin that will burst the bubble. Will it be Nvidia missing their numbers, OpenAI failing to raise, or Anthropics financials being leaked ahead of an IPO?
Malawitz argues, and I agree, that it's going to be something much more boring than that. It will be the launch of a 128 GB Mac Mini with the M5 chip. If the rumors hold, the M5 will be one of the best onremise infant chips that money can buy for mid-market workloads. That's the power of a server room sitting on your desk drawing less power than a kettle. And I know this because I've been building large models on that kind of technology for years. When I worked in corporate, we'd pay IBM hundreds of grand a year to rent a computer. But I worked out that with the right algorithm, I could match that and beat IBM's flagship technology using a Mac in my office.
When the Mac Mini drops, the cloud AI business model will break. And after three years of fear-mongering hype, and outrageous claims, we're going to finally see that the emperor has no clothes. And then it's going to feel like Y2K all over again. Because 25 years ago, we were terrified that planes would fall out of the sky. And it turned out to be a great big nothing burger. And we'll look back at this AI bubble and see it exactly the same way. An Apple sitting quietly and patient from the side will have been right all along.
I'm Dr. Errol Brandt, the founder of Kira, and I'd love to hear your thoughts and comments, even if you disagree with me. Thanks again for watching, liking, and subscribing, and I'll see you in the next one.