Why Labs lists the things that did not work
We archived a peer-to-peer sync system that worked perfectly and solved the wrong problem. Publishing that is more useful to you than another post about our values.
Every company's "Labs" page is a highlight reel. Ours lists the archived experiments too, with the reason attached, because the reason is the part with any information in it.
The clearest example is branch-to-branch sync. We spent five weeks building peer-to-peer replication so shops could keep trading through an internet outage. It worked. Two branches could go offline, keep selling, and reconcile when the line came back.
The problem is what "reconcile" means when both branches sold the last unit in stock. That is not a merge conflict. That is a business decision about which customer gets a phone call, and no replication strategy makes it correctly. We had built an elegant answer to a question nobody had asked.
It is archived, and the reason is on the page. If you are building something similar, that saves you five weeks — which is worth more than us looking clever.