I honestly don't understand what your intention with this statement is.

Todays companies are full with average software devs but they take 10x more hours to implement something, then miss 3x more edge cases while costing way more money.

So yes... and?

Exactly. ”And?”

That the ai is better means nothing. It is still pretty bad.

Is there any reason to not enable it globally?

The PEP has some notes on this.

https://peps.python.org/pep-0810/#global-lazy-imports-contro...

https://peps.python.org/pep-0810/#making-the-new-behavior-th...

the main reason is import-time side effects, particularly intentional ones like the registry pattern.

also let me plug my work project, an analyser which will scan your code and see where it can safely enable lazy imports (tagged alpha, but we did our best to get a usable release out before 3.15 landed): https://github.com/Facebook/lifeguard

Very cool for large existing projects! I'll definitely be taking it out for a test drive on a codebase which is luckily using quite a recent py version (so should be easy to upgrade).

Also bonus points for the mascot :)

Completely nondeterministic import order affected by code execution patterns?

Its really not that, so maybe read a comment or two before posting something like this?