That said, I don't think that Apple was wrong to give equal attention to battery life and portability. The number of people who truly need 32 gb of ram is probably over-represented here, and not indicative of Apple's overall user base.
Sometimes, niche markets are important. The solid gold iWatch "editions" were a niche market - too, but Apple understood that they fed a strategically important market - in that case, getting loads of celebs - the opinion-formers for airheads - to pose with their Watches.
With PCs, niche markets are important too, but the target "niche" is less good looking and well-groomed. Apple need a "power user" machine in their line-up that is going to attract geeks and programmers, because its those geeks that develop software for other Mac and iOS users, provide community support, advise people on computer purchases and generally keep the Mac ecosystem going. People won't buy Macs if there's no software or support. Microsoft can choose to produce only eye-wateringly expensive "Surface" products with consumer-grade specs, because there's a whole PC industry producing everything from cheap'n'cheerful entry-level PCs to refrigerator-size gaming rigs and workstations to keep the Windows platform alive.
What Apple have done is produced a new MacBook Pro that is more solid-gold Edition watch than power-user pleaser. Even if they sell, they're not going to sell to the sort of people that keep a PC platform alive.
Conventional wisdom in the past has been that if Apple ever produced the mythical XMac - an expandable mini-tower - that it would either be so laughably overpriced c.f. a PC mini-tower that nobody would buy it, or it would be so cheap and low-margin that it would decimate sales of MacBooks and iMacs and Apple wouldn't get the income it needs (unlike other PC makers) to develop and maintain Mac OS. That's been pretty sound logic in the past, but maybe it needs revisiting?
Today, when the mass market is for tablets and ultrabooks, would a Mac mini-tower really hurt the sales of laptops? It would be dirt cheap to develop - get Foxconn to knock out an Apple-branded MicroATX or Mini ITX motherboard sans PS/2 and VGA ports, make a nice mini-cheesegrater case (heck, you can buy them) and flog it for a nice premium: not going to sell huge quantities, but the Mac power users will already be shouting "Shut up and take my money!"
Another "courageous" possibility: an official "Hackintosh" program: $50/year subscription (or pay the Apple Dev subscription) for "Mac OS Enthusiast's Edition" with "not for resale" & "must already own a Mac" clause, a maintained list of officially supported chipsets and the guarantee that Apple aren't going to deliberately disable it at the next update. Are enough consumers going to download it, write it to a stick, configure it correctly to significantly hurt Mac sales (if they were already queueing up for that option, Hackintosh installation is already pretty much point'n'click)? Anybody unscrupulous enough to ignore the license, risk Apple's wrath and distribute Hackintoshes or Mac OS install sticks already can.
Less courageous: a quad i7 Mac Mini with dGPU (they've had that in the past and the world didn't end).
OK, those two ideas might fail closer scrutiny, but any losses Apple make would be offset by much better goodwill from the Mac enthusiast community which is going to pay off long-term. If the Apple board are capable of thinking beyond the next quarter. Basically, though, Apple need to stop confusing Macs with fashion accessories, and realise that they Need the Nerds.