1. Electrical fire starts, takes out ADS-B transponder, pilots unaware and continue the flight as normal 2. They radio a goodbye as they leave ATC, still unaware of anything which means that a fire has taken out a transponder and they are unaware of that fact 3. The fire is discovered. The crew put on full-face oxygen masks and the pilot makes for that airport either by dialling in a bearing into the autopilot or disengaging it and flying for it. 4. They attempt communication but fail because the systems are out due to the fire
At that point, we then have a lot of things to consider: was it pilot input or autopilot that took them to FL450? When did the pilots lose the ability to land, either because of control surface damage or because they were incapacitated? In the event of being alive but unable to fly the aircraft because control surfaces were no longer responsive to cockpit or autopilot input, is there anything else they could/would do other than do some maths and work out when the inevitable crash was coming? If they could control some surfaces but not others would they use that to buy themselves more time, e.g. climbing to ceiling (FL450), or would they get low quick?
This theory raises quite a few questions then, but it's a perfectly valid hypothesis: the issue for me around it is understanding why the pilots didn't go on to make the landing suggesting loss of control or their own lives by that point. The only way we'll get to understand that is by finding the black box which of course would explain the whole thing anyway.
This theory also suggests a likely ditching or crash into the water, and if that were the case then I would expect sonar operators over a large part of the region to hear the black box locator, pretty much as they did with the Air France crash.