I'm probably talking out of turn, but there have been so many of these threads recently that I thought I'd share my opinion as someone who's had a lot of the same questions.
There seems to be a consensus that older used gear is actually of better design and manufacture than the current, inexpensive new models, especially in terms of the quality of components used in construction. The implication is that those modern PSU's are manufactured or even "built down" to provide a similar [competitive] price point with the older stuff that's still available. Failures can happen, whether old or new, but the older (often linear) designs seem to be simpler (or at least more intuitive) and therefore easier to repair. The tradeoff between older and newer models at the same price point is often modern digital meters (as opposed to analog) for current and voltage, coarse and fine controls, and switching over linear (more efficient, but also often noisier). The newer switching supplies generally involve more parts, so that there are more things to fail, if they do fail. For the OP who seems to want something only for digital purposes, and is content with only 3.3V, 5V, and 12V, a truly variable power supply might actually be overkill for the present need. The modifications to an ATX power supply could provide all these, if it's the right design. Conversely, there is also a tradeoff in using a cheap old ATX garbage-rescue also; they aren't designed to be repurposed in this way and so there are a number of hurdles (including safety) that should be taken into account before going that route (limiting current, et al.).
I did a lot of looking into this while considering what I needed, and I'm still on the fence about the options (redundancy of functions might be the best tradeoff for me). As yet, there doesn't seem to be a "perfect" choice that will suit every entry-level hobbyist. To me (and take this with a grain of salt, as I'm no engineer, just a consumer), it seems to really depend on limiting your choice based on what you're going to do with it, which applications you'll potentially need down the road and whether you'll still be able to find a use for that same model then, or at least that it might hold some of its value when trading up. The other option is to buy something expensive and new from a "big" manufacturer that "does it all", but then you are paying for a great number of features that you well might never use in any application (and there doesn't seem to be a medium ground, they are a very expensive step up if you're used to using 9 volt batteries to power your projects).
The constant advice as alternative seems to be to "build your own", but I've come to the conclusion that I'm not personally competent enough to choose a 'good' design, especially considering that the designs that are often suggested often seem to have hidden design problems (overshoot, etc.) despite their vaunted simplicity.
There have been a lot of threads lamenting the absence of any decent DIY PSU kits left on the market. There are a few, but they all seem to require a lot of modding and extra work to bring the features up to date with modern expectations for performance, and the BOM for components and considerations of additional time involved basically bring them up to the same price point as buying a completed model. There is a gap there, one that I wish would be filled by an enterprising designer; if I could buy a kit that would allow me to learn more about design while also constructing a model with competitive features, I would go that route instead of buying something I don't know how to fix without having access to service manual/schematic and a bunch of repair videos helpfully posted by others with a (hopefully) similar dilemma.