Agreed. But that doesn't mean it doesn't make sense to embark on big projects. Rather than a "Hey, we're going to walk somewhere new" sort of thing, I'd like to see work on one of any number of space-related megaprojects - for example, a launch loop and/or fallout-free nuclear rockets**. Something that could actually lower the cost of access to space to the point that it doesn't take a vast effort to go walk on another celestial body.
** - There's so many competing designs it's hard to know where to start. My personal concept I've mulled over is a variant of the nuclear lightbulb concept, but instead of the fused-silica bulb containig a gas or plasma core reactor which requires some unknown containment method, the concept calls for a dusty fission core (akin to a dusty fission fragment rocket), which can be electrostatically contained. The energy would be released in the infrared, not visible or ultraviolet (as in a conventional lightbulb concept), but that's fine - fused silica is also transparent to infrared, and moreover doesn't lose much IR transmission as like happens in higher frequency bands; the lower radiation rate of infrared would be compensated for by the huge surface area of the dust radiating it. The simultaneous huge amounts of electric output (from fission fragment deceleration in a grid) could be used in part to run a microwave beam, creating a plasma sheath in ducted atmospheric air surrounding the bulb (airbreathing mode) or injected gas surrounding it (rocket mode) to aid in IR absorption and keep as much of the heat away from the (reflective) walls as possible. A VASIMR-ish mode is possible if you use low gas injection rates and a magnetic nozzle. In space, gas injection could be terminated altogether and the core could be opened up to run in dusty fission fragment mode and get Isp figures in the lower hundreds of thousands. To make up for the problems with using the standard dusty fission fragment rocket proposal's (heavy) moderator in such a high power environment, my thoughts were to have it operate as a subcritical reactor with a spallation neutron source as the driver - after all, there's no shortage of electricity to run an accelerator if you're decelerating a good chunk of the fragments; you don't even have to deal with Carnot losses.