The site should be useful before it is impressive.
That feels like an important constraint.
It can be beautiful later.
It can become more expressive later.
It can develop more craft over time.
But the first job is simpler.
Can someone understand what I am thinking about?
Can they find the live questions?
Can they see the source trails?
Can they tell what is current and what is historical?
Can their agent get enough context to be useful?
If the answer is no, then visual ambition is probably hiding the problem.
The useful version is quieter.
It gives the thinking somewhere to land.
It gives the reader somewhere to start.
It gives the system somewhere to point.
That is enough to be real.