Ask HN: MacBook vs. Dedicated GPU for LLM

A simple Ask HN question about running local AI on a MacBook versus a computer with a dedicated graphics card instantly turned into a classic internet showdown: convenience vs. raw power, with a side of buyer’s remorse. The original poster just wanted to know what a MacBook can handle. The community, naturally, responded like they were settling a family feud at Thanksgiving.

The cleanest explanation came from one commenter who said MacBooks act like a slow but roomy option: they can hold bigger AI models in memory, but they won’t be speedy. Dedicated graphics cards, meanwhile, were described as the opposite: faster, but with less room, so they run smaller models better. That was the sensible answer. Then the drama arrived.

One camp basically yelled, don’t buy anything yet. The doom-and-gloom take was that both options are "super super slow" and have "low payback," which is the kind of phrase that can vaporize shopping enthusiasm on contact. Another commenter made it even spicier by declaring, without hesitation, "Dual 3090 >>> Any Apple product." Subtle! Meanwhile, MacBook defenders countered with a very practical vibe: if you already want a big MacBook, it’s a solid all-rounder, especially because it’s still, well, a MacBook, and AppleCare+ has your back.

And of course, because this is the internet, someone said the real winning move is to wait 6–12 months for better AI models and maybe an OLED MacBook Pro. Translation: the hottest gadget debate of the week may end with everybody... not buying anything.