tbreschi

This is fun!

Great design and artwork. How did you generate the portraits?

Have you thought about extending this into some sort of pipeline for AIO?

This is an interesting yet simple approach. To the OP’s original question, how might you abstract this into a “design system” that can be applied to their other projects?

Brilliant marketing here in the title

1.. I’d have to think through about how to wire it up to listen to the other terminal where the job is running, but yeah I suppose this could work.

2. Yeah this is a smart feature would be cool.

Thanks for the feedback!

Yeah, I can do that. I actually bundled 8 different scenes into this release… was also considering including a skill for agents so others could write (and contribute) their own animations…

It suppresses macOS sleep and the display going off while it’s running, but it does not “survive” the lock screen in the sense of drawing over or replacing the system lock/saver.

The terminal app (and javalamp) will still be there and running when you unlock, but the lock screen or any password prompt remains fully controlled by macOS.

Right? Such a useful feature. I struggle to correctly tune my energy saving mode and screensaver settings - especially on a corporate laptop. Hence the motivation to seek out the setting.

Hey HN I'm Terrence and I built Javalamp, a terminal "lava lamp" screensaver that keeps your mac awake while your agents/tools are running.

I built it because I kept having longer running coding/AI jobs pause when my Mac went to sleep, and I wanted something lightweight and fun that lived in the terminal.

I’m curious whether others have this problem too and what features would make this actually useful (metrics, different visualizations, non-Mac support, etc.).

Feedback on both the idea and the implementation is very welcome.

Thanks!

Fun! Repo?

This is neat. I wonder… Is there a more comprehensive analysis of how well the Apple Vision framework compares to other multi-modal AI? Would there be any benefit to pre-processing images via Auge before handing them to Claude, GPT?

Yeah that typing effect at the top (expanding the composer) seems to be the isssue

Very cool. How’d you build?

Love this. Is the music AI Generated?