"My main concern is PCB assembly services, especially that you would make at least 100 boards to make it good enough right (all SMD design of course)? how about if they buy your parts (assuming you put digikey\lcsc part number)? delivery cost? customs? how about hand assembling it?"
Have done a few China runs, where the customer wanted more than 30 or 40 widgets - mostly for time, not for quality or $$$. Most of my production is 5 to 10 unit runs, which I build in my shop/lab using 'neighborhood' labor.
"Assuming a normal-sized project like a lab power supply (like uSupply or something) 30v-3A... I would estimate it to have something like 20 different SMD part, 3 through-hole parts, roughly 200-250 SMD pads... correct? how would this be profitable at 100 pieces of volume?"
You are offering a commodity product. If your stuff is good, you can have one or two profitable production runs for 100-piece volumes before it is copied in Asia and mass-produced and sold at less that 50% of your cost. The products that have a decent margin are typically more high-end stuff; that is, specialty instruments for unique industrial niches. And you will still have people trying to reverse engineer/copy this complex stuff. For example, some of my widgets have somehow ended up in Asia and Eastern Europe, where it is not uncommon for these people to demand schematics and source code (they never ask for technical support).
As for a blog - dunno, but go for it.