Very comprehensive (didn't read it all yet) guide about self-supervised learning. It'll likely become good reference material.

We got a problem with research around software estimates. This won't help us get better at it as an industry...

This is indeed very much true... there's a clear crisis in research. It turned into a hamster wheel of publishing articles at a constantly faster pace. The incentives are misguided which pushes that behavior to even have a career. Meanwhile, knowledge building suffers.

The rebellion against the academic publishers is still going on. Hopefully this will really change soon. That cartel of publishers needs to go back to its rightful place.