Of course we can still learn valuable lessons from these surviving structures.
Why not both? How do you think you discover that certain things work better than others?
Humans are clever in any time.
For example, the Pantheon has a very specific profile (thick at bottom to much thinner at center of rise) to the dome in order to allow unreinforced concrete to be used. That's a case of very deliberate design.
Pompeii is in a highly volcanic area. So, digging is difficult and expensive. This put a wrench in the standard formula for Roman building. So, instead, they elevated everything slightly. That's on the spot ingenuity and convenience.
Finally, there is a survivorship bias--structures specifically built with longevity and durability in mind and maintained by people are the ones that survived.
Taking the rebar out of a building tends to make it a touch less durable.
The Pantheon was turned into a church, and thus was off limits. In Rome they call it "recycling".
An interesting article: https://engineeringrome.org/ancient-structures-in-rome-the-c...
In comparison, iron could be made basically anywhere.
All I have to say is that on Windows the Blue Screen of Death was a prominent problem for many years. On the original MacOS that was non-preemptive multitasking and no memory hardware protection, any program could cause a denial of service or take down the whole computer. These are 30 year old technologies that were clearly worse in many ways than modern equivalents.
Has software improved recently? No, I would say it has gotten somewhat worse recently but if you look at where we are today versus say 1998, things are clearly better. It isn't really possible to know if we are in a major decline or a minor local minimum.
It reminds me of Shepard Tones which sound like they are infinitely increasing in pitch which is impossible because the human ear can only hear certain frequencies. (https://www.youtube.com/watch?v=BzNzgsAE4F0)
Their business model was (and to a degree still is) predicated on getting people used to everything having bugs, and not complaining or, more importantly, returning it for a refund, angrily. They succeeded beyond their wildest dreams, and now everything else works almost as badly.
Plinys “Natural History” contained detailed geographic information, but was considered a work of history (surprise, surprise).
https://en.wikipedia.org/wiki/De_architectura
Ptolemy (100-170AD) wrote treatises on many subjects, and his works on astronomy and geography "never ceased to be copied or commented upon, both in Late Antiquity and in the Middle Ages" https://en.wikipedia.org/wiki/Ptolemy
Vegetius wrote a military manual On military matters, extremely influential until about 1500.
https://en.wikipedia.org/wiki/Vegetius
Apicius wrote On the Subject of Cooking, a cook book.
https://en.wikipedia.org/wiki/Apicius
(So much ancient writing hasn't survived, e.g. none of Aristotle's many writings survived - what we know as his writings are just his students' lecture notes.)
So many being lost is an inevitable consequence of very very few copies ever existing.
Arguably Roman concrete was discovered by accident when someone mixed a batch of lime plaster using the volcanic ash around Pozzuoli instead of regular sand.