Probes are good for exploration, but colonies are useful too, and it's a shame that we stopped putting in nearly any effort along those lines.

but the problem was that we reversed the order. We should have gone to moon in the late 80s rather than build the shuttle, after having built a ton robotic probes and having much better automation. Imagine where we would be if we'd had a sustained 10-50 Billion /year increase in AI and robotics over the past 50 years. But because we pushed manned systems first, we both became disillusioned at the cost/benefit ratio and made that ratio worse by becoming very concerned about human life risk. And thus we lost momentum - budgetary, career-wise in engineering, public sentiment, political support ... On the other hand - I am a big supporter of the Goldin era "Faster, Better, Cheaper" approach. We got more done even with a 50% failure rate, and at half the cost - just by flat out doing more.