[jychang]

You can try running LLMs on your own computer!

They have extremely distinct sounds coming from the GPUs. You can hear the difference between GPT-OSS-20b and Qwen3-30b pretty easily just based on the sounds that the gpu is making.

The sound is being produced by the VRMs and power supply to the GPU being switched on and off hundreds of times per second. Each token being produced consumes power, and each attention and MLP layer consumes a different amount of power. No other GPU stress test consumes power in the same way, so you rarely hear that sound otherwise.

[vrighter]

this. I was running a reinforcement learning training run. I could very clearly hear from the coil whine whether it was simulating or backpropagating

[LargoLasskhyfv]

Hrrm. Maybe up to 25 years ago, but certainly 30 years ago you had similar phenomena via FM-Radio. Depending on what you did, there were different interferences in the radio. Unzipping something made different sounds than compiling, running a raytracer, or zooming into fractals.

One could use that while half asleep in the bedroom, whith a radio tuned into the right frequency, almost muted, and then know if Portage on Gentoo, or build.sh/pkgsrc on NetBSD was ready, or interrupted.

Because no buzzing or humming anymore :-)

[andai]

That is so cool. My computer isn't loud enough though. I think I'll have to install a guitar pickup. TEMPEST@HOME!

(I've also gotten great use out of a $5 AM/FM radio.)