It's partly that I think it is wasteful to remove (and scrap) one lot of diesel engines while manufacturing another lot of diesel engines for a new fleet. Remembering that manufacturing anything will involve carbon emissions.
Rail engines don't exist in a vacuum though. The MTU engines fitted to the 80x are used in a whole variety of other sectors including oil & gas drilling, and static power generation. Engines removed from the trains can be reconditioned by MTU and reused elsewhere. I wouldn't even be overly surprised if some of the engines used in the trains are reconditioned units from elsewhere, a google image search for the engine model reveals several sites offering second hand ones
If taking the engines off the IEP fleets and putting those second-hand engines under new trains is easier than swapping the carriages (complete with engines in the case of the old carriages) then I'd be happy with that too to a degree. My proposal above for swapping carriages has the added effect of enabling the GWML, ECML and MML to move (back) towards longer fixed-formation sets rather than multiple working of non-gangwayed 5-car units.
I would like to think that swapping the engines out is far easier than reforming the trains. The IEP TTS specifically requires bi-mode units to be "capable of being readily modifiable to an Electric IEP unit at a future date by the removal of Self Power Sources from one or more IEP vehicles" and "capable of being readily modified (so far as reasonably practicable) to utilise a different type of fuel and/or Self Power Source without the replacement or addition of any IEP Vehicles." - the other 80x units, utilising the same architecture will be similarly convertible. Granted it is also a requirement for the units to also be able to accommodate formation changes, but given the simplicity of simply pulling the genset and fuel tank out from underneath with a forklift and blanking off any holes, I fail to see the attraction of shuffling vehicles around and the complications that follow from discrepant ages, interior fit, etc.
As for the point about moving back to longer fixed formation sets, if it was advantageous to do that, do you not think we'd already be seeing it? Of course, in idealist-land we'd run full length units the whole time, but there isn't a need to, and the operational disadvantages from running 2 sets coupled aren't outweighed by the advantages. Not to mention of course that we should be right-sizing, there's no point running a train twice as long as it needs to be if you don't have to - sure if they're all electric and everything is powered from renewables (and hence zero carbon) you've still got the wear on the infrastructure/unit to consider as well as plain old entropy.
The key thing to me is not to manufacture any more diesel engines, except perhaps for regional bi-mode trains capable of no more than 100mph in either mode.
Don't hang up on diesel engines too much, the key thing is to reduce carbon emissions, preferably sooner than later, and in the most cost efficient way. If the best way to do that is quickly introduce a whole bunch of bimodes with fresh engines, then so be it. If for example a decision had to be made tomorrow, Hitachi could get a full fleet of bi-modes out and onto the XC network long before wires reach Nottingham/Sheffield/Bristol/Swansea/Plymouth etc. The extra carbon from producing those MTU engines would be more than offset by electric running that the bi-modes would be doing, and long before the bonkers cascade-everything-to-XC-and-build-new plan came to fruition.