I have a fancy new laptop with 64 GB RAM and a semi-competent GPU, so I configured Zed's agent to use a local ollama (via Alpaca) with gpt-oss:20b and asked it to make a relatively simple change across a few files. The GPU usage monitor looked like this for 15 minutes until I gave up, and it never managed to edit a file successfully. But at least I could generate heat to keep my lap warm while offline on a train. The way of the future!

Screenshot of system monitor app showing 2 minutes' history of utilization of an AMD Radeon 860M GPU: 100% utilization; 100% use of 4 GB dedicated memory; 13 GB of shared memory; both fans running at 100%.
0

If you have a fediverse account, you can quote this note from your own instance. Search https://mastodon.me.uk/users/wjt/statuses/115922274926677337 on your instance and quote it. (Note that quoting is not supported in Mastodon.)