From the atmosphere, humans can control surface probes in near real-time, enabling them to carry out experiments that are nearly impossible when you have to wait minutes to see the results of every action.
You look at the way the Rosetta/Philae mission went, there were a few things that went wrong with Philae that might have been addressed by a human crew if Rosetta had one.
For example, Philae had to do a fully autonomous landing, but a crew on Rosetta could potentially have piloted the lander in real-time and made course corrections to avoid the crash. More concretely, a thruster designed to keep Philae from bouncing was damaged - a human crew could have repaired it before releasing the lander.
If Rosetta had an operations crew nearby - even just one person - the cost of the mission would have been, what, 100x greater? Surely enough more expensive to pay for multiple robotic probes.
Total mission cost for Rosetta = 1.4 billion Euros. It costs something like $50,000/kg for GEO, so I'll use that as the baseline. The Orion capsule weighs 21 tons and the Deep Space Habitat at least 50 tons. Add one person and the 5 kg of consumables per day, for a mission of 300 days, gives another 2 tons. That's $3 billion already, or 2.4 billion Euros. Now add the development costs, support staff, communications, etc. and you'll see that it's a lot cheaper to send a small fleet of redundant probes than to send a single human.
In any case, we have no way to put a human on the same orbit as a comet, much less return the human safely to Earth, so only a robot mission is possible.
I didn't mean to suggest Rosetta should have been manned. I was giving concrete examples of how humans can save missions.
Venus has a highly corrosive atmosphere, and surface probes die within a couple hours. I can't find figures on the cost of historical Venusian surface probes, but I think it's safe to say they're more expensive than Rosetta and tremendously harder to operate from a planet away. I also don't think it's a stretch to suggest that humans could make a 10 or even 100 time multiplier on their effectiveness.
"Saving a mission" is not as important as doing the science. If three robot missions are needed to get the equivalent science from one human-supported mission, then it's still better off to send three robots, because outside of near-Earth space it's going to be a lot cheaper to send 10 robot missions than a single human mission.
It's certainly slower to control a rover on Mars from the Earth, rather than in Mars orbit. On the other hand, 5kg of consumables/day * $50K / kg is a $250,000 per day of operations overhead, just to keep the person alive. If the mission is delayed a few days due to dust storms, that's $1 million doing nothing.
Operations from Earth is slower, but Curiosity has been on Mars for three years. With a 100x multiplier, you propose the same might be done in 11 days. With a maximum speed of 90 meters/hour and assuming 8 hours of operation, that's a maximum of 8 km, or less than 1/2 of its current mission. That's of course excluding the time it takes to make measurements, like the hours needed to drill a sample.
Your 100x multiplier is therefore physically impossible. (A rover that could make more effective use of human time would also be heavier, and Curiosity was about the biggest we could manage.)
And remember, robots don't need to come back. Humans do. The rocket equation really hurts when you need to apply it twice.
My apologies. You are correct. I have difficulties in drawing any conclusions about Venus as I can't think of anything where 3 hours of human time would make a difference, so I used Mars (and to a lessor extent Rosetta) as the comparisons.
I researched various proposed Venus landings, like VISE, but still struggle to find something where a human in the near proximity would make that big of a difference.
That is, assuming landing can wait until Earth and Venus are at conjunction, and that 7° of angular separation is enough for a good single, then they are about 38 million kilometers apart, or 2.2 light minutes, so there's a 4.5 minute lag for ground control on Earth. Compare to Mars, which at best is 54.6 million km from Earth, or 3 light minutes, giving a 6 minute lag.
So I can only assume you're talking about driving a rover, which would require a lot of feedback. But a rover can't go far in a couple of hours, and if it's traveling the entire time then it's not drilling or taking spectrographs .. neither of which require much decision making.
Instead, the blue sky plans are for things like the Landsailing Rover, which use passive wind power to move around. And unlike Mars, it seems that Venus doesn't require much in the way of navigation, with little in the way of geography, so autonomous systems might be fine for most travel.
Really, I struggle your proposal, so I'm trying to give real-world comparisons so I'm not just blabbling negativity on the internet. But do you have any examples of where the science is worth the cost of putting a human on the scene, compared to spending the same amount of money on multiple robotic probes? Because if it means putting 1 rover on Venus for 3 hours or sending 20 probes for multi-year missions to orbit around all of the other planets, plus 3 rovers on the Moon, then I can't see how the human-near-Venus rover mission is worthwhile.
You look at the way the Rosetta/Philae mission went, there were a few things that went wrong with Philae that might have been addressed by a human crew if Rosetta had one.
For example, Philae had to do a fully autonomous landing, but a crew on Rosetta could potentially have piloted the lander in real-time and made course corrections to avoid the crash. More concretely, a thruster designed to keep Philae from bouncing was damaged - a human crew could have repaired it before releasing the lander.