Last few posts yield some interesting data, if you use a text sampling technique. Outlined like this:
1.). Grab a random sample. In this case grab the last 5 posts, maybe with some minor filtering.
2.). If a text sentence mentions the OP saying, in this case, "...they don't actually extinguish a battery fire..." ...Then that's a good 'hit'.
Also maybe mentions "...burning at 2500° F. enough to melt steel...rather than 750° F for
diesel fueled fire...". That also is counted as a 'hit' because it's direct from the OP.
Now then, what would be counted if the small text sample quotes a seemingly related point, just not directly from OP, AND strongly stays tightly linked. Like for example if unrelated fires in an industrial battery waste dump ALSO achieve the steel-melting high temperatures. That's a good hit.
What wouldn't count, in your little informal tally, is quoting data from a crash...who died, or who didn't die, in a crash. That's A MISS, and if you do this survey correctly, you can start to count the little tests up, yay or nay.
Reading the last few recent posts, there is very little mention, of any 'steel melting' hazards. Well gee whizz, the OP sure did mention that dynamic, as pretty darn significant.
You can build up a small, informal table, and then, whattayaknow: The OP posts discussed 'steel melting' while the responses, you note, seem to divert away from that failure mode, instead substituting a seemingly same concept.
But melting structural steel isn't the same as a Tesla driver dying in a high speed crash...even if the driver suffers the 2500° F burns.
See what I mean ? There are ways to sample text (especially in political discourse) to obtain a crude measure of subtle diversions, vs staying directly on-topic. It's harder, here in this topic, to maintain discussion on that 'steel melting' structural risk, when a responder wants to shift to 'high speed crashes', out on open road somewheres.
And, since we are in the subject, what high speed 100 mph + crashes are there, down inside a 3 level building parking basement?
Text sampling filters can quantify that sort of B.S. feeding. And I'm just a mere amateur.