tsgo crashing work M3 MacBook Pro

[osuritz]

I would open an issue but I'm not sure where to start with yet. Here's the setup of a large repo we have at work:

• Volta for managing Node version (don't ask)
• Yarn 4
• Node.js 24 project (although I think I started testing tsgo with the codebase still on Node 20)

tsc runs “fine” on the codebase: ~80–120sec and we have to bump Node to 8GB or the process crashes with OOM.
Naturally I wanted to experiment with native TS given the anticipated performance improvement.

However, running yarn tsgo crashes the computer: after a few mins it runs out of CPU and memory and then hard reboots. This happened 3 times today so it's reproducible for us.
When I ps, I see a crap load of forever duplicating processes of yarn and node processes. We're talking dozens of them. If I Ctrl+C several times, the madness stops and all the forked processes go away correctly.

I've tried to reproduce in a simple repo with just Yarn 4 and Volta but it didn't happen there, it just correctly processed the files.

I don't know what params or w/e to setup so start sharing the info with the team: there's zero output happening so I have nothing to share at this time. Where do we start?

@jakebailey if this piques your interest, you should be able to reach me on MSFT Teams (work)