Ask HN: People with new Macs / computers with GPU's, do you run LLM's locally?

3 vishalontheline 2 5/9/2025, 4:42:25 AM

I am considering finally upgrading to a new computer - most likely an M4 Mac with the primary goal of running coding assistants and training my own models locally.

Is this a good idea? Have you tried it? How's the performance?

Thank you!

Comments (2)

fiiv · 16h ago

I do this. Ollama makes it very easy - just pull the model you want. The great thing is the ability to test them in the same tasks, there's a huge difference in comparable models.

You can set it up in your editor of choice - I use Zed, and it's just listed as one of the providers you can choose.

In terms of performance - it works decently well.

bookworm123 · 14h ago

Cool idea, however when I have to create an account for a service just to test it out, I will naturally decline to do so. Maybe you could upload a sample document for people to play with?

I'd rather read the prompt (claytonwramsey.com)

The curse of knowing how, or; fixing everything (notashelf.dev)

From: Steve Jobs. "Great idea, thank you." (blog.hayman.net)

Show HN: Clippy – 90s UI for local LLMs (felixrieseberg.github.io)

Ty: A fast Python type checker and language server (github.com)

Void: Open-source Cursor alternative (github.com)

Design for 3D-Printing (blog.rahix.de)

Technical analysis of the Signal clone used by Trump officials (micahflee.com)

Show HN: Free, in-browser PDF editor (breezepdf.com)

Zed: High-performance AI Code Editor (zed.dev)

Gemini 2.5 Pro Preview (developers.googleblog.com)

The Death of Daydreaming (afterbabel.com)

OpenAI reaches agreement to buy Windsurf for $3B (bloomberg.com)

My new deadline: 20 years to give away virtually all my wealth (gatesnotes.com)

Accountability Sinks (250bpm.substack.com)

Claude's system prompt is over 24k tokens with tools (github.com)

CLion Is Now Free for Non-Commercial Use (blog.jetbrains.com)

Matt Godbolt sold me on Rust by showing me C++ (collabora.com)

Evolving OpenAI's Structure (openai.com)

First American pope elected and will be known as Pope Leo XIV (cnn.com)

I decided to pay off a school’s lunch debt (huffpost.com)

Waiting for Postgres 18: Accelerating Disk Reads with Asynchronous I/O (pganalyze.com)

LegoGPT: Generating Physically Stable and Buildable Lego (avalovelace1.github.io)

Unity’s Open-Source Double Standard: the ban of VLC (mfkl.github.io)

We fell out of love with Next.js and back in love with Ruby on Rails (hardcover.app)

Show HN: Real-time AI Voice Chat at ~500ms Latency (github.com)

Show HN: I built a synthesizer based on 3D physics (anukari.com)

Why Archers Didn't Volley Fire (acoup.blog)

Mistral ships Le Chat – enterprise AI assistant that can run on prem (mistral.ai)

Reservoir Sampling (samwho.dev)

DuckDB is probably the most important geospatial software of the last decade (dbreunig.com)

India launches attack on 9 sites in Pakistan and Pakistani Jammu and Kashmir (reuters.com)

Ask HN: What are good high-information density UIs (screenshots, apps, sites)?

Buffett to step down following six-decade run atop Berkshire (bloomberg.com)

Gorgeous-GRUB: collection of decent community-made GRUB themes (github.com)

High tariffs become 'real' with our first $36K bill (blog.adafruit.com)

Time saved by AI offset by new work created, study suggests (arstechnica.com)

Curl: We still have not seen a valid security report done with AI help (linkedin.com)

VVVVVV Source Code (github.com)

So Much Blood (dynomight.net)

Replacing Kubernetes with systemd (2024) (blog.yaakov.online)

The vocal effects of Daft Punk (bjango.com)

ALICE detects the conversion of lead into gold at the LHC (home.cern)

Sneakers (1992) – 4K makeover sourced from the original camera negative (blu-ray.com)

Old Soviet Venus descent craft nearing Earth reentry (leonarddavid.com)

Launch HN: Exa (YC S21) – The web as a database

An appeal to Apple from Anukari (anukari.com)

Judge said Meta illegally used books to build its AI (wired.com)

When flat rate movers won't answer your calls (aphyr.com)

Why can't HTML alone do includes? (frontendmasters.com)

Ask HN: People with new Macs / computers with GPU's, do you run LLM's locally?

Comments (2)