The main problem is that it’s just not battle tested like GNU coreutils are. And Canonical has only tested this in one cycle, 25.10, before introducing it in an LTS. Would’ve made more sense to wait until 26.10.

Other find problem with it being MIT licensed.

https://www.youtube.com/watch?v=tG2ZMvBT8W4

4:25

(sorry my third party youtube frontend can’t share timestamp links)

tldw:

• more CVEs than the old core utils that have been tested and in prod for over 30years
• no feature parity, so existing stuff that uses them will suddenly misbehave, when certain flags are missing
• different license, MIT instead of copyleft, so it’s more friendly for companies to use it for profit, while abusing the work of volunteer contributors

Its two fold for many (not for me): Rust and MIT