Yeah, my approach to an idea isn't "how will this fail" but "how can we make this work".
After we make it work, I decide for myself -- is this model worth it? If not, then back to the drawing board.
A dynamic tech tree sounds interesting in isolation, but is hard to execute without changing other aspects of the game. I'm a fan of the resource-tech idea.
But I'm more into pulling social techs (is this what they mean by religion and civics?) out of the tech tree. So your scientists research the scientific and engineering technologies, and you can even trade for these technologies. And anything religious or social is something you have to progress through independent of science. There social techs would have complicated requirements that are affected by your play style. If you play like a tyrant, you'll probably take a long time to discover democracy, for example. But it wouldn't necessarily impact how fast you can master a science. Iran, Korea, and India have all developed Nuclear Weapons in the past 5 years under completely different social conditions (Fundamentalism, Stalinism, Democracy).