So: the bigger M chips waste 13% of real estate on buses, leaving less room for cores. AMD and Intel have big cores (relative to ARM) — so, nope not a good match.

But I don't get the point about M design being (over) "specialized". Also, wasn't the magic sauce that Apple added that the DRAM is integrated on the same die? (Ie. very low latency)

Anyone with better understanding care to explain?

AMD and Intel made for cheap off-the-shelf components, and still perform good. - They use ordinary RAM, ordinary clocked buses, ordinary power converters, etc.

So in reality Apple made very precise hit, and it is very low probability of similar thing repeat.

And objective experts say, Apple sell their solution cheaper than it cost.

If unfortunately, Apple will miss anything in modern SoC infrastructure, Apple will miss, but AMD and Intel will survive and will gather people from Apple.

Why ORDINARY important - because it's production scaled very well, for example. today they sell million i7s, tomorrow could easy scale to 100 millions.

Apple solution is not bad, but "one size for all" MUST tie to very certain scale, it may overproduce, but people will not accept deficit (this will lead to very huge negative information trail, which even Apple could not survive).

So, Apple solution MUST sell lower than it really cost and overproduced.

What this could say me about future, if M1, etc, offer better performance for your tasks, buy it, I think this will not last long.

I think you this missed the point, point being power efficency.