It all depends on how long learning a process can be. One thing that I would suggest is a strong desire, willingness to tinker. That tends to open up the door much more in a couple of different fronts.
I encourage everyone to tinker the way doctors encourage everyone to exercise. But if the goal is to get monetary value from the skills/equipment, then the time horizon is over a year. If you have mechanical engineering knowledge, then at least you know what you want to make. But if you are trying to learn 3D modeling, 3D printing, and mechanical engineering all at once, you're probably going to take several years unless you do it full time.
The one bad thing about the "turnkey" machines is that more often then not, they require you to use consumables that they approve of. This goes even for the Dremel (last time I looked about a year ago anyway).
Sometimes this even means specific software used as well.
That's basically Markforge. Affordable machines as far as industrial printer go. Extremely reliable, high quality, and simple to use. The fiber reinforcement is unique and produces incredibly strong parts that can sometimes replace CNC metal. Expensive filament ($190/kg for Onyx), compared to $60-110 for other brands of carbon nylon. Onyx is very refined if you at look at it up close, so the unique formulation is actually what locks the printer to the material, and not some arbitrary BS like RFID chips. The settings are perfectly calibrated for Onyx, and I have never ever had to pre-dry Onyx. It's always ready to go out-of-the-box. But this is still better than Stratasys charging $260 for a kg of MG94 ABS that can be bought for $50 or so, while providing slightly better characteristics than generic $20 ABS. So you do get unique benefits with the Markforge. But if you dont need parts that precise, you can get by with a Dremel.
The Dremel is about $60-70/kg for their repackaged RFID-chipped materials. The 3D45 is their first printer (I believe) that lets you use any material you want, but you need to do a lot of tinkering with the settings and rig up an external feed as the internal spool mount doesn't fit standard spools. It's an approachable option if you are willing to spend at least a month learning how to service a 3D printer. I avoided consumer grade printers for a long time because people I know seem to spend more time fixing them than printing, but I've got over 2000 hours on my two Dremels over the last few months and its mostly reliable now. But there was a learning curve even though I thought I knew this stuff decently well going in.
This instance would be one of the few times that I would suggest one of the more open hardware options. Now, this would require that one is willing and able to tinker, but this also provides the most freedom. Usually the parts are more freely available. I don't want to say generic, because sometimes that gives the notion of being inferior, but they certainly aren't proprietary parts and one can actually improve the quality of parts (I did this with the extruders on mine).
But this also means that they usually accept input from other software compared to their own (or software from "approved" vendors). I'm able to use Blender either directly or stl files brought into Slic3r for even more control with the resulting g-code. They also tend to cooperate on more computer platforms for direct control (mine even came with a deb file (which can be loaded on a non deb system (or deb package manager I should say) easily enough)).
I got the Dremel for its build quality and warranty, and I use PrusaSlicer for slicing. Its got some weird firmware quirks, but its just 'open enough' for me without being completely DIY. Since I might be setting up a print farm with many printers and I will need to train people to use them, I don't want a mish-mash of printers with different quirks.
