Nuclear thermal rockets (NTRs) are notable as the most mature technology (the second-most-mature is Project Orion) to provide both high thrust and higher specific impulse than chemical rockets.
But "high thrust" doesn't mean the same as "high enough thrust for planetary takeoff", and most NTR designs have been intended for use as upper stages, orbit-only engines, or sometimes for planetary takeoff and landing on low-gravity bodies.
For those of us who are perpetually frustrated by the failure of the world's space programs to take advantage of nuclear power, there is a tantalizing promise of much higher thrust-to-weight ratio nuclear thermal rockets such as the DUMBO design that was developed alongside NERVA, and the military-funded TIMBERWIND project during the "Star Wars" era. While these still may not be enough to make the dream of a nuclear SSTO launcher or shuttle practical, they seem worthy of attention... if it is actually probable for them to develop the levels of performance claimed.
Various optimists have suggested that it should not be technologically difficult to build nuclear thermal rockets with T/W well over 20, which would make a variety of planetary takeoff options viable, and perhaps even make the nuclear SSTO feasible. This optimism has sometimes even been extended to closed-cycle gas-core NTRs. Other trustworthy figures have looked with intense skepticism on these hopes and suggested that in the absence of an actual demonstrator, things like the claimed T/W of over 30 for DUMBO should be considered mere fantasy.
Is it technologically realistic that, if there was a real will to produce high-thrust NTRs, that they could be built? Are there fundamental problems limiting how fast one can transfer heat out of a nuclear reactor?