Oh boy, the LM723. My eyes glaze with nostalgia......... as my very first project that included an IC and that actually worked was a power supply with a LM723 and an analog voltmeter, a 15-pound transformer and "computer grade" filter capacitors, the size of a small beer cans. Applying (and smearing all over the place) thermal compound to the TO3-case transistors and its mica insulators attached to a hefty heatsink.
As other people have mentioned, the LM723 was quite crude for today's standards and had several limitations. However, it had all the essential building blocks for a variable PSU, and its limitations required that one read carefully and understood properly the data sheets and app notes.
Also... lots of experimenting whereas perhaps a pair of LM723s and several output 2N3055 transistors would be blown up.
No simulators. Only paper, pencil, and a scientific calculator like a Texas Instrument's SR71.
Once that you had a proper working supply, you could actually claim to be a "power supply designer".
And crowbars....there was always a real possibility that one of the series pass transistors would short out, and it could take out expensive equipment connected to the PSU. Crowbar design was a black art itself, too slow a response and would'nt do any good. Too fast, and one would have nuisance tripping.