Local LLMs are overrated, says dev who regrets 128GB M3 Max over a 5090 PC
AIandDesign · x · 2026-11-22
AIandDesign argues "local LLM stuff is overrated": he bought a cutting-edge MacBook M3 Max with 128GB RAM three years ago, but local models still "don't hold a candle" to major cloud providers.
- Locally hosted models can't compete with frontier cloud models
- The same money on a 5090 PC wouldn't run 70B LLMs, but would render images and video at acceptable speeds — Macs can't
- He'd still prefer Macs overall: ideally a MacBook Air sending jobs to a Linux AI server
The thread sparked discussion on whether local LLM hardware purchases are worth it in 2025.
Related event: M3 Max Owner Says Local LLMs Are Overrated vs Cloud(2 posts)→
More from Infra
- Inkling 2 Mini hits 200 tok/s on RTX 5090 via vLLM with native NVFP4 — VoidAsuka · 2026-11-22
- Redditor predicts cheap GPUs will return, 'only a matter of time' — ridhwanss · 2026-11-22
- Designer argues local LLMs are overrated; would buy a 5090 PC instead — AIandDesign · 2026-11-22
- An AI bubble fable: from free cash flow to private credit to annuity blowup — Svwang1 · 2026-11-22
- One RX 7900 XTX runs 524k context at 767 tok/s, 3x faster than llama.cpp — W61k3r · 2026-11-22
- How GPUs organize work: kernel launches, thread blocks and scheduling explained — blaizedsouza · 2026-11-22