I refer you to an article (http://www.behardware.com/articles/670- ... ption.html) where they did some real power consumption measurements and recommended that 350W was more that enough for most high end systems.
Admittedly, they did create systems which exceeded 460W, but those systems were ridiculously over the top in terms of their specs. Hence, I would suggest that the Seasonic 460W PSU can power a majority of musicians PCs, including their DSPs and medium-sized RAID arrays, without pushing the PSU to the limit and, hence, leaving the PSU with a bit of breathing space.
That article is over 4 years old now through. Power requirements have risen although admittedly mostly in the gaming sector. The highest rated CPU I can see in that bench test weighs in with a 75w requirement where your looking more at 130w now with the current Intels.
It's possible to pull over 1k in a gaming rig. One of our dream p.c's a couple of years ago was pulling 1.2k from the wall socket in benching but it was hardly an average system (Tri SLI FTW ). If you take a dual SLI 5.2ghz gaming rig you can clear over 800w's with ease under full load.
Back to the point in hand through your right that your average music maker wouldn't need over 460w in an average system. I think our's pull about 250w/300w at the wall last time I checked but that was a fairly basic spec (High end chip/mid range board/fully loaded with memory/low end gfx/3 or 4 hard drives) and once you load up with UAD cards and other bits hanging off left right and center you'd be heading in the 400W mark.
So why would you want to spec higher than that? PSU's deteriate which is something they failed to mention. Each year and a loss of 5%-10% total load can occur so on a 460w you'd have the potential to be below the power requirement threshold after only a couple of years, especially if you keep adding kit to the system.
Another point is, is that the components are often rated to higher tolerances on higher models so they can withstand far more over time. You notice that on the upper models the active cooling tends to only activate after certain load thresholds are hit and a lot don't tend to need to spin up until after the 50% mark. It's in instances like this a 650w can be benifical noise level wise as it would rarely spin up.
In this case, the Seasonic 460W version of the X-Series is an excellent choice for most musicians. For my own system based on a bus-powered M-Audio FW 410 and a PC that was top level high spec only 2 years ago, my own power measurements with a current clamp show that the 400W version would run my system with plenty of PSU 'headroom' and in (near) blissful silence............
Seasonics are great PSU's and my second choice in spec's. My first choice through are the artical mentioned Enermaxs although annoyingly the best Enermax psu out there is the 1000w/1250w Revolution edition which has amazing load handling and has the lowest noise print of any active I've handled. Being rated that high and pretty much the same as everything else over 850w it's designed as a server psu solution first and foremost (it has a dual 12v board connector setup) and not as a home user bit of kit.... which would explain the price tag! Stepping down the range the Modu+ series they do is the home user one and ranges from 500ish upto around 800ish so makes sense and the shock headlines about 1K systems at home seems a bit over the top.
It's dismissal of servers needing large psu's seems a bit shortsighted through as I say the hardware market has changed since that artical was written. A basic twin hexcore cpu set up with 24GB with a stack of drives would pull 400W+ and you throw in a couple of quadros for rendering and your back over the 1K mark again although I suppose we weren't seeing those figures 4 years ago when that artical was written. You have to keep in mind that the chip firms have roadmaps and products prepared at least 2 or 3 years in advance and firms working closely with them will know the requirements of hardware coming out upto 4 or 5 years further down the line. Just because something seems out of place now, doesn't mean it won't find it's niche in 12 months time as the rest of the market matures around it.
TL;DR I agree with you to an extent at this moment in time, but the are a number of reasons in the broader picture as to why I consider Martin as being right in the long term.