Ask HN: What computer are you using for AI coding tools?

Y	Hacker News new \| ask \| show \| jobs

1 points by willsmith72 5 days ago

I'm a heavy user of Claude code, had an M3 pro with 18gb ram for a few years. Until a few months ago it was always a beast.

Now I'm spinning up more Claude code sessions (5-10), each with often 1-3 subagents going, as well as using chrome with playwright of claude-in-chrome for debugging/QA/styling.

That combination has really been slowing down my computer lately. I get it, it's a lot of processes and a lot of chrome. I don't like the cloud coding agents cause I often want it to use my own cookies in my browser, or other secrets/setup I've got locally

Curious if anyone else has these issues, and what you're using. Obviously it'd be nice to upgrade to an m5 with 24 or 48gb but that's 4k AUD (or 3200 with trade-in)

2 comments

inventor7777 5 days ago

5-10 sessions?! Wow, not sure how you keep track of those.

For all my daily use + Codex Desktop and some local AIs in Ollama/LM Studio (which power my Home Assistant Voice Assistants) I have an Mac Studio M4 Max with 48GB of RAM (Which I bought early last year IIRC)

It's a complete beast and runs GPT-OSS 20B at over 100 tok/sec. The GUI never slows down, and I run Sequoia still, so it is always smooth and polished. I've also never heard the fans.

I got mine for $2250 USD because I am eligible for the student discount. I got the base 512GB SSD because I store most stuff on external SSDs/HDDs anyway.

link

willsmith72 5 days ago

Nice thanks for the info. 48gb ram would be a dream

Yeah I go back and forth using more or less sessions. Was always a fan of kanban, less WIP better, but some sessions are just slow so I end up starting new ones.

link

inventor7777 5 days ago

Sure! I know the M3 series are still super fast chips, so I wonder what is slowing down...your CPU itself, RAM getting full, OS bloat...? Seems wild to me that processes that run in the cloud would take so many resources.

link

dlcarrier 5 days ago

I have an Intel A770, because at $200 USD it was the second cheapest way to get 16 of dedicated VRAM. (The cheapest is a $100 Radeon Instinct Mi25, but it's a bit old and slow.)

I battled with Python packages a lot to get a1111 working, gave up when it broke on an update and switched to ComfyUI which has worked great. I can't get Llama.cpp to use the GPU, but I'm currently installing some of Intel's proprietary OpenVino tools in hopes of getting an LLM going, and the performance through their tools is really good.

link