You're right, but business folks don't care.

Their view: "If Go is good enough for Google, it's good enough for us... and I hear hiring is easier and it's more Ai friendly."

I'm not saying any of that is good or true, only that the difference is lost on the people spending the money.

I don't mean to sound defeatist, quite the opposite, I'm saying there is hope, but it will require someone to be brave and find some new space to occupy or an existing one to disrupt.

Being better at HTTP isn't a good enough selling point.

I hope to be inspired and impressed again. 😊

Fluke or not, all the fancy meta-programming in the world won't help if we don't have a purpose, a niche, a reason for investment.

You can build better backend services in Scala than in Go. But so what? We could be better at data than Python (again?), but they have the ecosystem / manpower / hype.

I suggest there is no commercial point of difference anymore. Yes yes type safety, correctness and effect types. But I mean products. Things businesses care about: We have no Akka, we've lost our Spark.

We have to offer something new or better, as Python did. An abundance of Json libs won't cut it.

I've been experimenting. I have instructions set up that say "don't be sycophantic, don't assume I'm right, check my assumptions, ask clarifying questions, if I'm wrong tell me, etc."

It's still a bit of a kiss-ass, but the work/results seem better.

The upshot of this, to the original effects system question, is that ZIO (and IO) will remain as supported as they are now via the 'classic' APIs.

The new experience will not expose either directly, but uses IO as an implementation detail. You can still use IO based Cmds/Subs if the need arises.