My point is less about the AGPL license itself and more about the people who choose it.
If the OSI came out and made a statement on the ambiguities in the AGPL, and cleared the FUD in such a way that all companies agreed and could reference it, I'd wager that the AGPL would over time become much, much less popular for commercial open source. But I'd wager that they won't do that, because they win when COSS wins.
But if they did, we'd likely even see a move towards non-OSS licenses, ones that are clear as to their intent and rules -- rather than relying on ambiguity -- because there would no longer be an OSI-approved license that businesses could use to have their cake and eat it too.
Very few COSS business/startup/w/e right now are choosing AGPL for altruistic reasons. This thread and every other COSS licensing thread here are evidence of that.
Few of them care about software freedoms, or know why they chose the AGPL. They have a playbook that says AGPL lets them take advantage of the open source distribution flywheel, while largely protecting them from the risks associated with commercial open source, i.e. competition. They choose AGPL for this, not because it's the best license for users.
I honestly don't get how people don't see the deception under the AGPL right now.
But why wouldn’t they change their minds if those big corporations actually changed their ways? Is there a similar sentiment against red hat?
I now understand where you’re coming from but I am not sold on your prediction that the agpl would crater if Google started complying with it. It would mean that Google open sources everything which is derivative work. That sounds like it would buy a lot of good will amongst precisely those people who are mad about how Amazon screwed redis (to put it bluntly).
This point of view is new to me to be honest.
Let's say I don't care about the intent of people choosing the AGPL (I do, I wish people did stuff for altruistic reasons, but the economical system in which we live makes it so we can't rely on this).
You say people are choosing the AGPL because they think it lets them do effectively source-available while benefiting from open source washing. Fine. I don't like this. But the effect of this for me is that we actually get actual free software.
What's so bad with this?
I've reread your second text and didn't find what's actually bad with the AGPL.
Now, I wish all the FUD around AGPL was cleared; the FUD is what's bad, but I don't wan't to wait for this to happen before picking the AGPL for my software.