Indeed, much depends on the human element, a single person on board to make decisions should be more than enough in that regard. Looking at your second paragraph, you clearly missed my point so I reiterate: the more complex, more capable, more modern aircraft, the one with the most sophisticated sensor suite on the market, with the theoretically higher workload, retained the single seat layout as well.
The ability to reduce that theoretically higher workload to something more palatable and increasing overall effectiveness has been around in service aircraft since the early 2010s now. What benefit would there be to go back to the objectively inferior option of having a second person in the aircraft when even a strike fighter from the 2010s can effectively break down the modern battlefield for it's pilot into the most important pieces to pay attention to? Especially when the point is reached where the aircraft can essentially fly itself and leaves more room for other duties with the human operator. I would have assumed that the benefits of AI should be obvious to anyone really, and that their integration into the military ecosystem is progressing in large steps. But it seems like some are still rather backwards and stuck in the 80s.
The second person for endurance reasons is a compelling argument however.
Your reasoning misses one specific question, which is -- what is the actual requirements that a single human is expected to do, for a next generation aircraft?
Let's assume that we use the term "A" to describe the amount of workload that a single pilot in a standard single seat J-20 can perform (as a baseline), with workload defined by a combination of -- flying the aircraft, controlling sensors and weapons and EW, developing tactics and coordinating with friendlies, controlling friendly UCAVs/CCAs etc -- in context of a standard J-20's existing degree of automation, AI and sophistication.
The question for J-36 (as well as other aircraft, from J-20A, J-20S, J-35/A, J-XDS etc) is comprised of:
- what are their respective requirements relative to "A" in terms of the needs for them to conduct their missions in the future, for a given aircraft type?
- what is the degree of automation, AI and sophistication that will be developed in future
- is the degree of automation and sophistication projected in the future sufficient for a single pilot to be able to manage the workload of whatever multiple of "A" will be?
For example, if the workload of J-36 is expected to be "5 x A" then perhaps a the future projected automation, AI and sophistication may be enough for a single pilot to manage.
But what if the workload J-36 is expected to manage is "10 x A" or "30 x A" -- would that be enough for a single pilot to manage in context of projected future automation and AI? Or do they need two crew members even with future automation and AI, due to the sheer extent of the requirements they have for the aircraft?
Because all of this talk about how "even a strike fighter from the 2010s can effectively break down the modern battlefield for it's pilot into the most important pieces to pay attention to" -- is only partly correct, because it is describing a battlefield of the 2010s, rather than the 2020s, 2030s, 2040s... and even for the 2010s, the extent to which 4.5th and 5th gen aircraft of leading air power nations were able to recognize the extent of a mess that a true high end air war would be, was likely dubious.
Fighting an air war with a couple dozen friendly tactical aircraft, with AEW&C control, and maybe a couple of friendly EW aircraft, against a moderately capable opfor air force that is projected to be technologically near peer at best, lacking sophisticated and distributed EW, against late cold war era SAMs, is different to an air war where there may be hundreds of friendly and hostile manned tactical aircraft in the same airspace that are stealthy, with hundreds more of unmanned (friendly and hostile) aircraft and payloads -- all of which are in permutations of actively emitting and jamming, or controlling, or launching, while trying to ingress, egress, engage, or evade, with likely hundreds of missiles simultaneously in the air during periods of engagement, alongside large scale theater wide EW and perhaps cyberwarfare efforts, and you not only have to make sense of what is actually happening but also develop tactics on the fly in context of the larger operational scale picture.
And the purpose of manned tactical air in that situation is not just to know what is going on, but to be as practically useful as possible in context of what their individual airframe can handle.
In that context, the number of human beings will likely be the rate limiting step for combat effectiveness (rather than automation/AI).
All of which is to say -- it is correct that automation is going to further improve and increase the overall effectiveness of the human crew member in an aircraft relative to the past.
However, I would not bet that improvements in automation/AI will improve human effectiveness to a degree that is able to keep up with the modern warfighting environment in a manner that it would not be beneficial for certain parts of one's air fleet to have two crew members versus one (which is where J-36, and even for today perhaps J-20S, would come into play).
