Welcome to the hobby. I'm having a hard time seeing the "downfall" (maybe I'm just blinded by jealousy ). You're so far ahead right now that you could buy a crap load of dies and still be light years from "downside".
I started my reloading journey on the 650 and even paying retail I couldn't be happier with the decision,
Find the published load data for both IMR and H4064 and work from there. They may be close in performance, but won't be identical.
As twodownzero states, the two 4895s are different enough that many manuals show separate load data for each. Always work up toward max charge because even if it's published data, you can still get pressure issues. I have experienced that with 180g Nosler BTs and H4895 in .308 where I was getting pressure signs a full 2 grains below Hodgdon's listed max.
Yes, I did that and had good results with Varget, even in .223
The trick is to slow down and give it time to clear the throat, something I learned here. It's a good thing too, because Varget works so well with both .223 and .308
Nothing meters worse through my 650 than IMR4064. +/- 1.0g inconsistency
H4895 meters well for me, as does RL15. I can run those in progressive mode and get SD in the mid teens
I was going to start with the 550, but there were so many comments from guys like thheard above that I decided to just step up and get the 650.
Why not?
Well, this noob started his reloading journey with a 650. I load for accuracy in .223 and .308. I can single load when I want to measure every round and I can run in progressive when I need more volume. The nice feature for a beginner with the 650 is the powder check station. No way to miss a charge or over change. That's peace of mind for a new guy who's worried about not getting himself "blowed up real good"
I figured out something loading .308 on my 650. I've been focused a lot lately on precision with emphasis on runout. I've managed to get it down to .002-.001
My seating technique when I'm running in manual-single (not progressive) was to hold my finger against the case pushing it tight against the shell holder as I'm raising the ram. My epiphany came when I measured runout on a batch in progressive mode and run out was .0005 on all 50 rounds. My theory: there needs to be a bit of slop so the the case and bullet center in the seating die (in my case a Redding Comp die) and I was adding to runout by inserting myself too much into the process.
Also, I added an O ring under the Dillon sizing die that also reduced run out by ~.001