You get what you pay for. Robots hands down win at minimizing risk to human life, but people risk their lives just to have fun let alone gain fame or to make great discoveries. Orbital robots are also clearly superior as their tasks are easily automated. This is one reason I totally don't get go to the moons of Mars first plans. Doing 90% of the work for 1% of the return makes no engineering sense to me. However, humans are clearly superior when it comes to looking for evidence of life past or present on the ground due to their ability to adapt and change course instantly based on new data. I firmly believe humans are dollar per dollar far more effective on the ground. Even Steven Squires agrees on this. Michael Golumbeck stated Sojourner did 10 minutes of work for a human geologist. I've done automation for a living and have intentionally programmed out operator decisions as much as possible, but when it comes to a totally new task, I'd give it to a human every time.

Regards,
Mark