Flyte enables production-grade orchestration for machine learning workflows and data processing created to accelerate local workflows to production.

Flyte

Since it is early stages of development, the change might be easier

We used Asyncio because it is native to python 

Anyio is a layer on top, any reason you think we should prefer anyio in the framework 

Does it cause problems in context or interoperable 

Shamelessly coping from Chat-GPT

```
```Key Advantages of anyio
	1.	Structured concurrency:
	•	Prevents task leaks.
	•	Tasks are cancelled when their parent scope exits.
	•	Helps write safe, maintainable async code.
	2.	Interoperability:
	•	Works with both asyncio and trio.
	•	Lets you write backend-agnostic async code.
	3.	Unified API:
	•	Same API regardless of whether you're using asyncio or trio.
	4.	Better cancellation handling:
	•	Uses cancel scopes like trio, making cleanup easier and more predictable.```
I have just started looking at the code, so might be right now asyncio does the work well. But i have been using anyio recently for my operation and i think it give all functionality of asyncio + added benefits as above. Moreover it can plug asyncio and trio, so the user is not bound by the backend.

But with asyncio, they are tied

There are a few gotchas with asyncio, that trio and anyio resolves: <https://sailor.li/asyncio|https://sailor.li/asyncio>

Great points - few things are changing soon <@U085FDMUZFD> 
The core is actually moving to rust - for reasons of concurrency management 
But as I read it the challenge is for end users to write correct code providing them an option to use trio would be beneficial 