This all seems to be a very odd route to eventually saving pilot costs.
I think it's fairly clear that there's no long term future of human pilots sitting at the front of the aeroplanes that ferry people around from place to place. AI and automation will eventually get so good that they will out-perform humans on every metric, with no startle delay to compromise the handling of in-flight emergencies.
The question is, how to get there. You can halve pilot costs with current tech, at the cost of additional risk to both actual flight safety and public relations. But you can't go any further - current CAT tech needs at least one human, so going single pilot ops (except possibly for short haul cargo), is a dead end.
Whereas, we already have military pilots flying drone attack missions from a comfy seat back at base. And we already have
AI that can outperform FJ combat pilots.
Couple the two together, and it will lead to one pilot being able to manage several flights from the same seat depending on the scheduling - effectively whether the busy bits associated with takeoff and landing coincide - and as the AI gets better and ATC gets more automated, the number of pilots required will diminish - ultimately to zero - in a scalable way.
When one pilot approaches the end of a shift, the next one arrives, there's a handover. Simples.
It will take longer to achieve than 2025, but it seems more like a roadmap to a solution to me.