Thomas James writes that NASA is (with little fanfare) disposing of the tooling to build Shuttle Orbiters.
This doesn't make it impossible to build new ones--the blueprints probably remain available, and new tooling could be built in theory, but it dramatically raises the (already ridiculously high) costs of building any replacement vehicles. Even if we were to continue to fly the Shuttle, we will do so with a three-vehicle fleet, so we would never get a flight rate higher than the current one (which is the highest it's been this year since we lost Columbia). Until we lost another one, anyway.
This really is a point of no return.
ET tooling is critical too. It especially features heavily in the Direct proposal. Direct being 8 m ET + 4 seg SRB:s and two launches vs Ares 1+ Ares V, the latter with 10 m ET and 5 seg SRB:s.
"we were to continue to fly the Shuttle, we will do so with a three-vehicle fleet"
not true, since China (remember the spy story?) seems have its blueprints, so, it will build cheaper Space Shuttle and sell them to all countries of the world that want one :)
That "it dramatically raises the (already ridiculously high) costs of building any replacement vehicles" doesn't, of course, mean we won't be hearing in 2025 or 2035 about how we threw away a valuable capability that coulda been a contender -- viz. the cult of the mighty Saturn over the last 20 years.
The cults surrounding spacecraft that were never built can be even more enduring. Witness the (nuclear pulse) Orion, or the original fully reusable shuttle concepts.
The most reliable rocket we have is the Atlas, which uses pressurized tanks rather than unreliable turbopumps.
Rather than the unreliable Space Shuttle, a new rocket would use the fuel oil-liquid oxygen Atlas as its propulsion, rather than liquid oxygen and hydrogen and solid rocket boosters.
The capsule should be small, then attached in modular fashion to other modules as needed to get the necessary space.