The printer arrived flat-packed, which I had been warned about, and I still managed to mount the gantry slightly out of square on the first attempt. That set the tone nicely. An Ender 3, the obvious starter machine, because if it goes wrong there is a forum thread for every conceivable way it can go wrong.
The first print was the test dog that ships on the SD card. It came out fine, which lulled me into a false sense of competence. The second print, my own model, did not.
the failures, roughly in order
The first real failure was adhesion. The print lifted at a corner about three layers in, dragged itself sideways, and produced a small plastic bird's nest. The bed was not level, or rather it was level but the nozzle was too far from it. I had skipped the paper-drag levelling step because it felt fiddly. It is fiddly, and it is also the single most important thing.
The second failure was heat creep. I printed too fast with the part cooling fan off and the hot end softened filament above the melt zone, jamming the extruder. The motor clicked, ground a flat spot into the filament, and stopped feeding. I learned what a "click of death" sounds like and learned to actually read the recommended temperatures rather than guessing.
The third was stringing, which is cosmetic but maddening. Fine wisps of plastic strung between the towers of a calibration model like a spider had got bored halfway through. Retraction settings. I dropped the temperature five degrees, bumped retraction distance, and most of it vanished.
what actually helped
Slowing down. Every problem I had was made worse by impatience, in the slicer settings and in me. I dropped the default speed, turned the first-layer speed down further still, and suddenly things that had failed three times in a row just worked.
A glass bed and a glue stick. Cheap, and it removed an entire category of adhesion problems overnight.
Writing down what I changed between prints. I started keeping a little text file: date, model, what I altered, what happened. After a week it was the most useful thing I owned, because it turned "the printer is being temperamental" into "I changed two things at once and cannot tell which fixed it, so stop doing that".
The honest summary of week one is that I produced more waste plastic than finished parts. But the failures were legible. Each one had a cause I could find and fix, and that is genuinely satisfying in a way I did not expect. I came for a tool and stayed for the debugging.