With the MacBook Neo, Apple Made the Perfect AI Computer

A lot of the conversation about the MacBook Neo is whether the compromises Apple made in order to sell a Mac for under $600 meant that you ended up with a computer that wasn’t actually able to do anything useful. Of course, it doesn’t take long to realize that the Neo is, in fact, more than capable of handling most of the computer things people who are inclined to buy this particular Mac might need it to do. One of the things that conversation seems to have missed is the idea that the Neo is perfectly equipped to do the only thing that tech companies seem to think anyone cares about: AI. You can argue whether that’s actually true, but there’s no question that the Neo is one of the most interesting computers in the age of AI computing. To be clear, the MacBook Neo does come with compromises. I’m not going to go through all of them now, partly because I wrote about them when I reviewed the Neo. But also because all of the Neo’s compromises are irrelevant to making it a great computer for AI. It’s not that other Macs are less capable. There is, however, something magical about the idea that a $600 entry-level Mac is as capable as a $4000 MacBook Pro, or $6000 Mac Studio, when it comes to the most intensive computing that any of us do today. That, of course, is because most AI computing happens in the cloud, not on your computer. That means that the limiting factor isn’t memory, storage, or how fast your processor is. No, the limiting factor is how well you’re able to get your AI tool of choice to understand what you want. Oh, and I guess the speed of your internet connection. That means that a MacBook Neo, with an A18 Pro, 8GB of memory, and a 256 GB or 512 GB SSD, will be just fine to run the Mac ChatGPT app or run Gemini in Safari. And that changes what your laptop actually needs to be. I don’t know that Apple had that specific thought when they made the MacBook Neo. Maybe they just wanted to make a low-cost, entry-level MacBook that would appeal to people who wouldn’t otherwise buy a Mac. Either way, they ended up making what might be the most accessible AI-first computer yet. With the MacBook Neo, a high school student, freelancer, or small business owner can now own hardware that gives them full access to the best AI tools in the world. Interestingly, this isn’t exactly the way Apple has framed the marketing. In fact, Apple isn’t shy about how it markets the MacBook Pro as the laptop for AI. The new M5 Pro and M5 Max chips, Apple says, deliver up to 4x faster LLM prompt processing than the previous generation. The MacBook Pro, in Apple’s words, is built for “AI researchers and developers to train custom models locally.” I’m not arguing that isn’t a real use case. But I think we can all agree it’s a very narrow one that most people don’t understand or care about. Training models locally or running 30-billion-parameter LLMs on-device are things that matter enormously to a specific kind of user — and are completely irrelevant to almost everyone else. The average person using AI doesn’t need to run a model. The average user just wants to talk to one. When you ask Claude to help you rewrite an email, or ask ChatGPT to explain something complicated, or use Gemini to summarize a document, none of that requires local inference. The model lives somewhere else. The compute happens in the cloud. Your laptop is basically just a keyboard and screen for a computer that does the work for you. The MacBook Pro is a remarkable machine for people who need what it does. But positioning it as the computer for the AI era implies that on-device model training is how most people will use AI. It isn’t. It’s how a small number of highly technical users will use AI — the same people who were already buying MacBook Pros anyway. For everyone else, the question was never whether their laptop could run a model. It was whether their laptop could get out of the way while someone else’s computers did. For $599, Apple may just have given us the computer that answers that question. EXPERT OPINION BY JASON ATEN, TECH COLUMNIST @JASONATEN