Unless you plan to discover new ideas by mistake, you absolutely need a vision, and then find the way to get there.
If you start from the bottom (bottom-up), you'll constantly compromise technically, as lots of things are not yet possible to do.
If you start from the top (top-down), you'll "know" that it's possible to accomplish, and you'll only have to find out how to do it.
You have far more chance to solve an enigma if you know there is an answer than if you don't. Visionary ideas make you believe the answer exists, which makes it much more easier to accomplish.
Clearly, you have to "know" that it's possible to accomplish and then find out how to do it. I don't think even PG was disagreeing with that.
When he says "don't have a blueprint", he is saying don't presuppose you know how to get from where you are to replacing Google as the de facto search tool. Instead, just make progress. It seems you agree with that.
It's also the Lisp philosophy of building up from small functional pieces, rather than starting with an overdesigned top-down blueprint.
I think what Paul is saying in the first part I quoted is different though. He is saying don't even say you are going to build the "grand vision" because it sets unreasonable expectations and defocuses you from making incremental progress.
The second part I quoted is about avoiding top-down design / blueprints and you are right on the money there.
This is critical. I have tried it the other way, and struggled for these very reasons.
I think the way to use these big ideas is not to try to identify a precise point in the future and then ask yourself how to get from here to there, like the popular image of a visionary. You'll be better off if you operate like Columbus and just head in a general westerly direction. Don't try to construct the future like a building, because your current blueprint is almost certainly mistaken. Start with something you know works, and when you expand, expand westward.
Eat small morsels, chew well!