I'd swap the Intel 660p's immediately. Not that they're bad drives per se (I have one myself) but at over $200/TB they're about 2x what I paid for mine. For that price point (or close to it) you could get a PCIe 4.0 NVMe with a fantastic controller that'll blow the 660p out of the water in both performance and durability. Samsung 980 Pro, or it's OEM equivalent, the PM9A1, are great choices.
While the 3955WX seemingly must be a good CPU because it's expensive, it's really not meant for general computing or gaming. It's a workstation chip. A 5950X on the AM4 platform will cost about the same (if you go with one of the ~$1K mobo's, like MSI's Godlike or Gigabyte's Auros Extreme) and massively outperform the 3955WX in gaming and general computing. If you absolutely still need to piss away money, you could go with regular Threadripper which will be decent for gaming (not as good as the 5950X) and still be a big CPU that you can gawk at. But in that case, wait for the Zen 3 Threadrippers since they shouldn't be too far away and should be a large improvement on the Zen 2 TRs in the market right now.
Can I ask why there's two different sets of RAM? One 8x8GB and one 4x32GB? You can't use both at the same time. I'd recommend going with Samsung B-die if money is no object, so try to find a 3200-CL14-14-14 kit or 3600-CL16-16-16.
1300W is overkill, but hey, go big or go home. At least there's room for expansion
(and lots of it)
Though I'd recommend a different brand, CM is decent in the cooling space but I have no experience with their PSUs. Corsair and EVGA are trusted brands.
You're also missing a case, and cooling.
Honestly, I think you're always much better off going with a high-end ($2-3K) build today and another refresh 5 years from now. Look at anything from 10 years ago, it doesn't matter if you got an i7-2700K or an i5-2500K, both would be pretty terrible these days, the i7 might have lasted a year or two longer. And that's with about 3 years of stagnation taken into account. There really is no such thing as futureproofing, no amount of money can adequately compensate for technological progress. So your friend is better off going with what is admittedly about 98% of the way there with top-end parts, then putting the rest of the cash in a fund and withdrawing it for another build in another 5ish years.