[Capricorn2481]

> They may come to the conclusion that Zig is incapable of delivering on its promises or is deficient at satisfying their requirements

Sure, but what you're suggesting is not related to the timeline I gave. They did not determine Zig was deficient in some way. They tried to get a cheap gain, and the gain breaks parts of Zig and they didn't even realize it, and it was worse than the gain already available in Zig. That seems less like they've made a pragmatic choice about speed and more like they are doing headline based development.

What you write makes it sound like there's a pragmatic process being followed that only you are privy to, and I'd like to know what it is. Zig may be inappropriate for Bun after all, but this makes it look like they don't understand what they are doing, and the agentic coding doesn't help.

[baranul]

My description was pretty much on point. The authors of Bun appeared to have grown tired of Zig's "memory leaks, segfaults, and other stability issues". The statements (check their twitter/X) are not at all shocking, considering the over 3,000 stranded issues at Zig's GitHub repo.

Unless team Bun had some type of philosophical or financial allegiance to Zig, which they appear not to, it shouldn't be surprising that they jumped to some other language (Rust in this case).

> this makes it look like they don't understand what they are doing

Think it's more a matter of Zig not being able to deliver nor meet their requirements. That they may have got tired of Zig, so felt it was time to switch to something else, is their prerogative. Especially with having near unlimited tokens to blow.