3/3 combination of sequential/parallel read/execution

```bash
(log_with_sleep "1"; cat <(log_with_sleep "2") <(log_with_sleep "3"); (log_with_sleep "3" & log_with_sleep "4" & wait); log_with_sleep "5") | cat
```

2/3 parallel read; parallel execution
```bash
(log_with_sleep "-1" & log_with_sleep "-2"& wait) | cat
# 10:11:43.079 starting-1
# 10:11:43.080 starting-2
# 10:11:44.094 end-1
# 10:11:44.094 end-2
```

Actually it's a cool idea. I'll try to write something about this and at the same time I hope it will help me to understand better this topic, because I always feeling that I'm missing something here

I'm thinking it's bad idea mostly because of interfacesad tests. Both are pretty bad at k8s side if we are talking about infra management.

Oh no, I'm sure there is a vast difference, primarily because of state. So managing infra is more like managing stateful applications, with only difference a much bigger impact.

Yeah, that was a main argument. I agree, reconciliation sounds pretty cool, still I believe that it didn’t bring enough value, compared to the problems it brings. The main problem is the obscured interface, which usually requires other tools (like helm) to fix

We talked more about concepts, but used crossplane as example how k8s not fitting infra management in general case