92 comments

[ 3.0 ms ] story [ 152 ms ] thread
From the source report:

"Although NASA continues to improve its process for identifying and managing health and human performance risks associated with space flight, we believe that given the current state of knowledge, the Agency’s risk mitigation schedule is optimistic and NASA will not develop countermeasures for many deep space risks until the 2030s, at the earliest."

https://oig.nasa.gov/audits/reports/FY16/IG-16-003.pdf

Of 135 shuttle missions, 2 resulted in the death of all crew, giving a confidence interval of something like 1-4% for the mortality risk of each flight. Any sane assessment of the first Apollo landing missions would put the subjective mortality risk at 10% at the very least. (Later landings were somewhat safer, subjectively, since the equipment was proven.)

I can't imagine the in-mission mortality risk of a Mars mission would realistically be less than 5%, and 15% would not be unacceptable to me. A 10% in-mission risk equates to something like 5 years of lost life in expectation.

In comparison to this, the risk of cancer seem small or comparable. Yes, you'll need to provide sufficient shielding to prevent aggressive cancers from probably developing during the mission, or ones that will reliably lead to death before 60. But I'm not at all surprised that we may need to accept a much higher cancer risk than we would for a normal occupation. This doesn't strike me as one of the key challenges.

I can't imagine the in-mission mortality risk of a Mars mission would realistically be less than 5%, and 15% would not be unacceptable to me.

Thus the rationalist Mars mission: send a drone to Mars. Give your five favorite NASA crew members one shot in the head apiece with a fairly spun six shot pistol with one bullet chambered. This accomplishes all of your objectives, is cheaper, and minimizes the danger that the mission architects are underestimating the risk.

The general claim is that your argument is flawed in "this accomplishes all of your objectives". I don't claim to be an expert in the field, but I've heard people who do care about and practice planetary and martian science that a manned mission could accomplish more than all the previous unmanned missions combined.
That could be true and manned mission can still be irrational choice.

All manned missions arguments I have seen are somewhat coy in this respect. Yes, geologist can make better choices on spot and take better samples than current drones, but that's not real comparison.

We should compare alternative strategies using opportunity cost.

The money used for manned mission could be used for massive fleet of of drones, satellites and technical development. Missions could be more incremental and cover more areas. Self driving cars are just taking off on earth. How does advance fleet of Mars robocars and satellites mapping Mars compare to manned mission?

Some of the lectures I've seen actually peg this as true, but for the wrong reason.

It is possible to put together automated missions at-scale. But not likely. Because auditors can trim back an automated mission's scope smoothly, no matter how expansive it is, in order to save money. A slightly lighter, cheaper camera, gives way to an even more slightly lighter, cheaper camera, ad infinitum. A hundred units give way to 99 units give way to 30 units give way to 1 unit. But if you try to do that with a manned mission, you start to jeopardize the public heroes you have sensationalized in order to fund the mission. You don't get to send half a human being to Mars.

We've already sent a number of drones to Mars. There are rational reasons for wanting to send humans as well: Humans can do things on Mars that drones cannot, and they pave the way for possible colonization in the future.
> Humans can do things on Mars that drones cannot

Such as? The only thing I can think of is that they can test if the human body can endure life on Mars. Honestly, as cool as sending people to mars is, I don't think it is an idea driven by rationality at this point.

Humans on mars would inspire future generations with dreams of space travel, which will make them more likely to support higher levels of NASA funding?
You mean like how landing on the moon did wonders for NASA's budget?
More like how the landing on the moon spawned SpaceX, Scaled Composites, etc. - It made people keep shooting for the moon.
Sadly, the more useful something is, the more it is underfunded.
>> Humans can do things on Mars that drones cannot

> Such as?

Growing potatoes?

I presume more complex experiments on site, that take time and gave a lot of steps and decisions to be taken. of course this also can be automated, and probably at lower cost overall minus potential casualties, but then there is irrational/emotional aspect to getting to a different planet, which is perceived as major milestone in mankind's aspirations.

mountaineers also go to potentially deadly situations voluntarily, just because the those peaks "are there". nobody gets satisfied enough from someone's else photos and says "nah, that's enough of an experience for me", albeit view doesn't offer much more. Neither they do send drones up there instead of them :) (btw that would be a nice use case for some remote high-adrenaline-vr gaming with androids!)

The dexterity of the human body really should not be underestimated. A human being can perform relatively complex analysis in a lab precisely because s/he can move the samples around and whack the equipment a bit if it gets stuck. If you want a robot to do that without getting stuck - ever, then that becomes extremely expensive or you figure out a way to do it in a single chamber like the viking lander did, which is still way more expensive and starts ruling out more complicated forms of analysis. If some element of this mission breaks down, then the human being can also do in-field repairs - something that even the most sophisticated robots of our time can't do.

Beyond picking up rocks, the human can traverse the terrain more easily and s/he can climb up and down really steep slopes - even ones at a nearly 90˚ angle. Whether you'd want the human to do that due to safety concerns on an alien world is another matter, but the amount of terrain accessible to this being is far larger than most billion dollar robots.

Oh and the human being does all of this while being powered on tuna sandwiches. They might be expensive to send to Mars, but once they're there humans can explore a larger area than all rover missions combined in a single mission.

Sending 1 human to mars requires on the order of 20+ tons of equipment at a minimum, plus consumables. That's a lot of robots, so even if a human was 100 times better than the opportunity rover, sending rovers would still be a better bet. Worse there is a large risk they die before doing anything.

If anything Mars is a great opportunity for automation. It has no environment to destroy and with the right design it could end up exporting a constant stream of goods to humanity with zero environmental cost. Nuclear waste for example is a non issue. We could even try things like a space elevator on mars without risk of harming people on earth.

PS: For comparison the lunar lander weighed ~36,200 pounds and needed a lot less thrust to get into lunar orbit.

What goods would you expect to retrieve from Mars? Or do you mean that it would be a great place to locate industry that would create unacceptably high pollution on earth?
In theory, you send one von Neumann machine and in a few years it has a planets worth of manufacturing to send you anything you want. Sure, shipping would be energy intensive, and you only get ~59% as much solar energy to work with. However, the equivalent of a 2 million meter x 2 million meter solar farm could produce a lot of energy.

Assuming space elevators, you could replace the vast majority of earth manufacturing. Though, with a long lag time before arrival.

<The dexterity of the human body really should not be underestimated.

This is a great point. I haven't studied this topic much, but on the surface it seems to me investing these resources into robotics technology is more interesting. Why can't we have a robot that can do such things? I think we can.

> Oh and the human being does all of this while being powered on tuna sandwiches.

and water and balanced nutrition and oxygen and a very narrow temperature operating range and a very short hibernate mode (cannot really go without water/food for an extended time) and have to sleep every day and sensitive to radiation and cannot operate in low pressure and need to defecate and urinate etc.

A drone/robot can be run on solar power. If there isn't enough energy it can go into hibernate mode for months. And you also have to take into account that a manned mission is planned for something like 2030. Robotics is such a rapidly evolving field that by then they will be so much more sophisticated than today.

Don't get me wrong. I think sending people to Mars will be awesome and inspiring. But I think it wouldn't be a rational decision.

Robots can't reproduce. They can't repair themselves. They can't even fix each other. One day in the distant future we'll build robots that can do those things, I'm certain of it, and then we can build planet-colonizing robot swarms. But even then, the most useful thing they will be able to do is prepare for the arrival of humans. Robots won't be ready to replace humans for the purposes of science and exploration until they're ready to replace us altogether, which is a long way off yet.
That's absolutely true, but I think the hesitancy about sending humans to Mars is that it is such a desert, and so the payoff will be quite limited to most people besides geologists. From the point of view of the general public, it's cool that its on another planet and we could do interesting science, but realistically it doesn't seem worth colonizing.

This is one reason that Europa excites the public imagination despite the hideous problems of distance and proximity to a gas giant - that moon appears to have a thick crust of ice covering a liquid ocean, and thus the prospect of life - possibly alien life, possibly a means of maintaining human life.

If we found evidence of life, even fairly primitive life, there would be plenty of money and public support to study it (though not without controversy). Rocks and canyons, not so much.

Until we have AI (since remote control doesn't work in this case) and far, faaaar better robotics, humans are vastly superior to have on Mars. Well, if they're alive, anyway.

See also: You can probably do things a roomba can't.

The technology required to send people to Mars is not in itself worthless. The motivator is sending people to Mars.
To think of the simplest and silliest example - you can get a shovel and dig as deep as realistically possible. Even our most advanced rovers dig hilariously shallow holes(due to technical limitations) and our science would be improved tenfold if we could simply roll over a couple rocks and punch a few holes through the ground.
A hypothetical human on Mars can do things that the drones we currently have in Mars cannot.

A more interesting question is: how much does it cost to send people to Mars, and if we use the same amount of money to send more robotic probes, what can a hypothetical Martian astronaut do that (equally hypothetical) Martian drone armies cannot.

But it completely ignores the main reason we do this: because we're humans and we're not rational. We have irrational emotions, and we think symbolically. Sending a human to Mars and bringing them back is way less about actually going to Mars and back and way more about humanity making sure we can still beat challenges that face us.
with the speed of robotics advancement, compared to advancement in material science and human preservation technologies, we actually might have a team of base-building Martian robots before being able to send astronauts there.
This probably makes the most sense. I still see sending a human to Mars as a likely precursor though. If we prove it is possible, alongside the advancements that come from the missions, would hopefully dramatically increase NASA fandom and funding.
Maybe a drone can build Starfighter too...
"Rational" is not a property of ends, it is a property of means (and beliefs). All ends are arbitrary.

The robots sent to mars are not useful for any reasonable economic reason. Even insofar as you count speculative basic research as economically useful, there are vastly better ways to spend that money on speculative basic research on Earth.

So why go to Mars? Because either (1) you find the basic research being done about Martian geology particularly compelling or (2) you want to see it explored by humans for its own sake. If you really actually believe in (1), then by all means; send more robots. But I'll bet 100:1 that you personally have no appreciation for the basic science questions, which are mostly very technical and boring. (Hint: grand-seeming questions about life are promoted for publicity and funding reasons, but these are not the ones that actually drive the science, and we are not learning more about the origin of life from going to Mars.)

Rather, I think you'll have to admit we want to go to Mars because it is awesome. You may think sending metal under our control is just as awesome as sending flesh and blood humans, but ultimately this is an aesthetic judgement which is not grounded in rationality. Personally, I would like to see humans walk there, and I'm happy to pay my citizen-share of $100/year over the next decade to make it happen.

I think you should work on quantifying the power of "awesome" in terms of increased planetwide hopefulness and collaboration. Optimism itself is a very powerful and meaningful thing to achieve.

Pretending "aesthetics" don't matter and our hardware isn't very significantly designed to respond to emotion is a huge flaw in "purely logical" reasoning. For goodness sake didn't we already cover this in Star Trek TNG?

I'd happily invest in SpaceX if they let me. Even if it counts as 'wasted' money in the books of some the simple fact that humanity would aspire to walking on another planet is worth backing. Robot exploration just doesn't feel the same to me. Hearing a 'roger' beep still gives me the willies and I was only 4 when the moonlandings happened. Nothing we did after that was as impressive to me, even though I'm sure it was easier to impress my 4 year old self than it will be to impress me today, walking on Mars would certainly qualify.
>"Rational" is not a property of ends, it is a property of means (and beliefs). All ends are arbitrary.

So let's just skip the trip to Mars and spend the money to build high-tech pyramids in the Nevada desert and sacrifice virgins.

Isn't that the whole point of Las Vegas?
One justification for space exploration is to eventually colonize space as a means of decentralization.

It's still far off in the future, but we have to take one step at a time. One of those steps will be sending people on a mission there with a return ticket, eventually establish a continuously manned research base etc. etc.

So there still may be some utility in the long run.

Of course there might be other approaches to bootstrap space colonization, such as the idea of mining asteroids in orbit and maybe sending people up there in the process.

Unfortunately, the thing holding us back from space colonization is not transport, but the technology to make colonies self-sustaining (or at least net economically profitable, even if tied to Earth by an umbilical cord of re-supply). Space transport is a much simpler problem, and investing in it will not much help space colonization.
Transportation is exactly the technology needed to make colonies economically feasible. And as you say, an 'umbilical cord' will surely be necessary for the first colonies -- this is also a problem of transportation.
>this accomplishes all of your objectives

No, it doesn't. It doesn't at all! Establishing human presence on Mars is one of the core objectives of the whole program. It's one of the reasons we care about Mars. Hell, moving gradually toward making additional homes for human beings is part of why we do space exploration in general. Yes, you can gather rocks via robots and send back data and whatever else, but to claim that doing so exhausts the set of reasonable objectives for Mars missions is just silly.

Here's a thought experiment to confirm: How many probes have we flown to how many planets in the solar system? Any idea, roughly, when the first one happened, or the last? You may know, but most folks have no idea. Conversely: where in the solar system have humans been? Any idea when we first got there, or when the last trip was? Unlike all the many highly valuable probe missions we've sent, here, most people actually know something about it. This is because people care about other people and their experiences, in a way that they don't care about scientific results as such. Putting human beings on other worlds is an end in itself, and it is simply not accomplished by sending drones instead.

In comparison, the risk of death climbing Mount Everest is in the 1.5% to 10% range (depending on how you count), and this is something people do for fun. Obviously safe space flight is a good thing, but I think it's reasonable to accept that going to Mars is considerably more dangerous than climbing Everest.
Yeah, this is what I don't get. People do drugs, skydive, go sailing and do 150 MPH on motorcycles for fun, but they can lose their shit if 1 in 100 dies going anywhere in space...

I guess it will change with time...

Because the costs of putting people in space are huge, so you need correspondingly huge buy-in. Dangerous recreational activities only appeal to a tiny minority of people (while quite a lot of people take drugs recreationally, most do so fairly cautiously rather than in heroic quantities or contexts a la Hunter S Thompson).

With space travel, it's both expensive and inaccessible other than to those who could be regarded as among the 'best and the brightest,' and whose loss is felt disproportionately. A better contrast might be that we (in the US) have a moderately high tolerance for deaths in war, because the bar for entry into military service is not so high. But if you recall the early days of the war in Afghanistan, there was a big freakout over the death of Corporal Pat Tillman, who abandoned a promising career as an NFL quarterback to serve his country post 9/11 and then died in a friendly fire situation during combat. It's very hard to become quarterback in the NFL so his death was considered a great loss, even though morally it's no different from the deaths of thousands of other soldiers. So deaths on a Mars mission would provoke a reaction of 'why are we sending our best to die in the cold wastes of space/on the surface of distant planets when we have perfectly serviceable robots?'

I think the way it will change is if privately financed asteroid mining becomes a practical reality, in which case deaths in the line of work will be considered tragic but acceptable (within reason), much like deaths on oil rigs and other high-risk industrial environments.

The "freakout" around Pat Tillman was because he was killed by other US soldiers and the military tried to cover it up.
A great deal of the value to a manned mission to a planet, in terms of tangibles, exists in coming back. Humans can take samples easily, and a return mission would bring them back. Humans would significant contaminate any site they visit, and drones can be considerably more sterile, not to mention cheaper.

When you are talking about spending many, many billions for a single mission, arguably the success of the spacecraft is worth more than the lives of the humans. If the crew is at a significant risk of death, and it is worth noting that a lot of these risks would affect the crew mostly evenly, then the whole mission is in jeopardy.

It's not possible to justify the very high price we pay for human safety within the space program based on value to the mission. If it were only a matter of the mission, we would accept tremendously more risk. Human lives are only valued at a but under $10 million, so even for a 10-person mission to Mars the value of the humans is < 1% of the cost of the mission.

https://en.wikipedia.org/wiki/Value_of_life#Life_Value_in_th...

Because of the diminishing returns to safety spending, it's generally better from an economics perspective to accept an order-unity chance of failure and just send multiple missions. But obviously, this is not how we design human space travel.

Armstrong has said he thought he and Aldrin had about a 50% chance of success for the moon landing. Failure could mean lots of things outside of situations resulting in death, but still, they were taking an awe full risk.

Also, looking at the manned Apollo missions, one out of 12 resulted in the death of the crew (Apollo 1), and one in 6 missions were failures (Apollo 1 and 13). Space flight was a hugely risky endeavor back then.

Something carefully avoided in the journalist article, yet relevant, is a vaguely four year mission would naturally have a fatality rate around mission length divided by average human lifespan so a death rate around 5% would be somewhat normal. You take a vacation of 1/20th your lifespan, 1/20th of you will die.

Before the screaming starts, I am well aware this off the cuff estimate is not the pinnacle of modern actuarial science, and there are a zillion mitigating factors likely to lead to lower fatality rates.

I would be kind of pissed off at having to spend 1/20th of my lifetime mostly on a commute to a vacation, I would better tolerate spending maybe 1/40th of my life immigrating permanently somewhere. Also roughly 1/2 the cancer risk would be nice.

How much stuff you'd have to send to make it permanent vs merely 1/20th of a lifespan is somewhat unclear.

Given the long travel times, the untried machinery, the unknown risks, the known risks, the isolation, I doubt the chances of getting there would be larger than 10%, and returning about 0%.

One possibility might be to send older people, or people with a short life expectancy on earth, with the idea they could do something great with their remaining time. This may be politically impossible, though.

Given the small number of crew needed for an initial programme of missions, it ought to be possible to select from the population of the Earth a number of superbly able and healthy people who simply accept the risk.

I sometimes suspect that the proposers of sending older people to Mars are getting on a bit themselves and wish for a spot they otherwise would be denied. (Nothing personal Mr Bright, I'm sure you'd be an admirable astronaut and I'd happily launch with you if we were given the chance). You don't need to select crew from a subset for whom the calculus of risk of death versus ongoing life on Earth is different in some way. You just need crew who are willing to take the risks, whatever their reasons.

NASA selects superbly fit people because they can. I don't see the crises one is likely to encounter in space are ones that superb fitness is going to help much.

There's no way I'm going in a tin can to Mars :-) so that's not my motivation.

I'm sure there are many young people who would be willing to go on a one way ride to Mars and certain death. I suspect that would be politically impossible, though.

I'm sure there are many people who want to launch me on a one way trip to Mars!
Most if the machinery could be tried on the moon first. If we adopt a Mars Direct approach, a return vessel with in-situ refuelling capability would arrive and recommissioning itself before the crew even left Earth, and we might even trial the return vehicle in an automated there-and-back mission first. With all these factors, I think the chances of success could probably be improved to perhaps 90% arrival and 70% return. Pure guesses though, of course.
I agree that establishing a self-sustaining moon base would be far more achievable and useful, and would make a Mars mission a lot more practical.
"the astronauts chosen to make at least the initial forays into deep space may have to accept a higher level of risk than those who fly International Space Station missions."

Is this surprising to anyone? It's a far greater challenge?

The world is not moved forward by those who seek the safety of their current existence. It is those who dare to throw caution to the wind and embrace the challenge of the unknown.

Just because there is a danger of dieing doesn't mean that it shouldn't be attempted or that there are not those willing to attempt it.

Exactly whose safety are they concerned for? The volunteers who would gladly accept the risks of death, disease, and injury for a chance at glory? If we'd had this attitude to risk in the 1960s, the Apollo program would have ended after Apollo 1 burned up.
Yes, that is exactly whose safety they are concerned with. People who are willing to give their life for a cause may be too willing to do so. There is nothing wrong with making a project as safe as possible, even if you can find volunteers with higher risk tolerances.
"There is nothing wrong with making a project as safe as possible"

This is crazy, it will leave us with space projects that will never be finished. Didn't mars one prove that people are willing to die on Mars to get there first? The cost of getting people to mars is trivial compared to the cost of getting them back. We spend so much time on safety we have become totally immobile in space. The whole NASA budget should be given to the private sector!

OK, fine, what if we changed it to "as safe as practical"?

The point is that those who are willing to die should not set the safety standards. We should surpass their personal risk tolerances, and make it safer than that.

As a parent, I found that what I considered "as safe as practical" was considerably different than what my wife considered safe.
And as we learned after Shuttle disasters, mission failure leading to loss of life has political consequences for the nation and the human space program as well. The risks are not limited to the astronauts.
Crew death and illness would represent a serious threat to the mission, especially when it is coupled with damage to the various machines. This is likely to be much more of an issue with such long duration away from earth.
(comment deleted)
Anyone have historical audits of NASA from 1960? I'll bet they were saying the same thing about moon missions back then.
This is what I don't understand, can someone ELI5 it for me:

"Even nutrition, the report said, could be problematic, because any Mars vehicle will likely be significantly smaller than the International Space Station, and astronauts will be too far away to receive the regular resupply missions they now get on the station."

Why does that have to be the case? Why are we limiting the size of the spaceship to what can be launched from Earth? Wouldn't it be better to construct teh ship in space so it could be a behemoth that had ample supplies, etc. and in effect could be a semi-permanent space station in Mars orbit that could be launched from to the surface with smaller ships?

Fixation on modern implementation of Apollo, go there as a stunt and never come back.

Probably they're specifically talking about some kind of transfer vehicle doing a Hohmann transfer orbit. The idea of sending very small resupply and spare part payloads on a non-Hohmann likely very expensive trajectory is interesting, but it'll never work as a substantial fraction of the mass of the ship, like food will have to be. So you could use the entire delta-V of a major booster to deliver a spare circuitboard to a vehicle 3/4 of the way to Mars but significant masses of food won't work.

Another problem is organizational inertia, if a Hohmann transfer takes 600 days or whatever and a major international mission takes 10 years of administrative planning, then if a super emergency resupply is required on day 400, it won't arrive until 3400, in other words a bit too late to matter. So technically you could waste a normally 20 thousand pound booster on a 10 pound emergency payload of circuit boards and med supplies, but organizationally "we" are too slow to do it.

The short answer is: launch costs and crap rockets.

Building a big ship, right now, means lifting lots of mass into orbit and assembling it there. That in itself will be stunningly expensive. But then you want to actually send it to Mars.

Earth orbit to Mars orbit costs about 6 km/s (see http://i.imgur.com/SqdzxzF.png), which is comparable to the 9.4 km/s needed to get from the surface of Earth to low earth orbit. Now, we are in space, which means we can use slow, efficient burns rather than the fuel-belching behemoths you need for a launch... but right now we're limited to inefficient chemical rockets, which means a hell of a lot of fuel.

The ISS masses about 400 tonnes. Assuming hydrazine as a fuel, with an isp of about 350, that means that the wet mass of your vehicle has to be (dry mass) * e^(deltav / isp9.8) = 400e3 e^(6000 / 350*9.8) = 2300 tonnes. So you'll need 1900 tonnes of fuel. That's 380 Falcon 9s full of rocket fuel. (Assuming my maths are right.)

So you try to make your spacecraft as small as possible. That way you don't need to carry the fuel needed to move the fuel. The rocket equation is a bummer.

(Of course, nuclear engines, are vastly more efficient, change all this. Shame we don't have any. If you were willing to wait long enough, an ion thruster with an ISP of 20000 would require only 12 tonnes of fuel. But you wouldn't live that long.)

Now factor in all the failures on the pad. Lets round to 400 Falcon-9s. And thats just for the ISS. To go to Mars, you also need to get off Mars. Here are some of the challenges to doing that:

Constraints are placed on the MAV by the Martian environment and the entry, descent and landing (EDL) system used. The MAV will have to endure daily temperature swings of up to about 100C with an annual maximum of about 24C and minimum of 111C. (These values were calculated using the NASA Ames Global Climate model for Holden Crater, which was the MSL candidate site with the greatest variation in temperature and is assumed to be representative of a MSR landing site.)Current MSR architecture places the MAV in a sky-crane type lander along with a fetch rover to re-trieve the samples. The EDL system constrains both geometric and mass limits on the system. The baseline design is a solid rocket. However, the thrust of solid rockets is dependent on their temperature. Additionally, any cracking of the fuel grain could lead to a cata-strophic failure."

http://www.lpi.usra.edu/meetings/marsconcepts2012/pdf/4342.p...

Also, and I will have nerd-sniped away any productivity now from you dear reader, all of NASA's papers are public. Here is the whitpaper on a Mars Ascent Vehicle:

http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/2011001...

(be warned, this is going to kill any work for the day)

That's "only" 36 falcon heavy launches (with an expected mass to LEO of 53 tonnes). With it being mostly reusable (at least the two boosters), this isn't quite so daunting an idea.

I mean, we did build a space station. We've done similar things in scale and complexity. And we did it with the ridiculously inefficient space shuttle.

Yes, but realistically one could mine the Moon poles and develop the fuel there...

Better yet expand it grow food on moon and refuel and load up on food in moon orbit.

Does growing food on the moon actually help? Unless you can mine the nutrients for the food, you still need to bring up all of the mass that would go into the food, at which point you might as well just bring up already processed food (which would also be denser in value/weight). The main benefit of growing food in space is that it lets you recycle nutrients. (I am including the atmospheric gasses as nutrients in this).
Yes, you'd have to mine the nutrients there for this to be effective. As you say, there's no point lifting it all from Earth.

Organic life is largely made out of carbon, hydrogen, oxygen and nitrogen (the so-called CHON). There's not much carbon or nitrogen. This is one reason why people are so interested in hypothetical ice deposits at the poles. There's a reasonable chance that there'll be stocks of this elements there (in the form of ammonia or hydrocarbons).

The only real hard limit is likely to be specifically the radiation acquired from Galactic Cosmic Rays.

There is, to a first approximation, nothing you can do about this; And it is in fact up there around our thresholds for dose limit, unless you're satisfied with increasing the mission mass by a factor of a thousand. The only practical thing to do is to adjust our thresholds.

But screw it - we're talking about landing on another planet, about millions of man-years of time investment, about activities plenty of people would donate DALYs to accomplish; About activities they will usually have to donate their entire career and their marriages to accomplish anyway, where the unknown unknown risks of sudden death are expected to be of a higher magnitude in the first place. Thresholds are silly in that context.

For comparison:

Flight attendants who have been on the job between five years and their whole careers have ~600% of the breast cancer (and likely other cancer) risk of the median person.

Regular smokers have 2600% the risk of lung cancer as people who have never smoked.

Pay a smoker a million dollars to sign a contract stating they will never smoke again, send him to Mars, then pay a private investigator to follow him around for the rest of his life ensuring he does not start smoking, and he will have a vastly lower rate of cancer at vastly lower monetary outlay than if he had been rejected at the first astronaut screening and NASA had attempted to mitigate radiation risk from GCRs.

In case anyone else is puzzled: DALY = "Disability-adjuseted life year", i.e. a year you're alive and not disabled.
Humans were evolved to survive in a razor-thin sliver of Earth's atmosphere and nowhere else. We aren't designed for and don't belong in deep space.

We should genetically engineer smart creatures that can fare better out there, creatures with enormous capacities for cancer resistance, stupefyingly impressive energy efficiency, and staggeringly long lifetimes. They don't need all of our gravity-influenced skeletal structure and musculature. The fact that we're air-breathers is a particularly worrisome liability for being a space-faring critter. Little holes in our tin can atmospheres can lead to explosively bad results.

I'd start with something closer to an octopus and then throw (even more) intelligence at it.

surely it would be easier to build better robots?
Maybe, if the robots can self-replicate and/or self-repair and have that creative spark, but that's asking a whole lot of tech that we haven't sorted out yet.

Life has the replication shit already worked out, so if the goal is really to put intelligent life in deep space in a sustainable way, a genetically engineered super-bright octopus species in a space-faring aquarium with environmental selectors for intelligence built in might actually be more effective.

A realistic near-term goal is to establish a space station at the earth-moon L1 point such as

https://en.wikipedia.org/wiki/Exploration_Gateway_Platform

The answer to radiation is that you want 6 feet of rock, ideally all around you, but particularly between you and the sun on solar flares. That can be harvested from the moon or an easy asteroid.

A presence on the moon is a lot more realistic, and if commercially useful water can be found at the pole and converted to fuel, it would make it affordable to send astronauts to mars with radiation protection. At that point we would have extensive experience with "living off the land" in space and won't be doing it on a planet with toxic soil.

At the very least, ISS should have a module or two dedicated to growing food. Right now. I don't understand why they're neglecting this technology.
Just as an experiment? ISS would have to be much larger to be self-sufficient in food. On earth it would take around 4,000 m^2 of solar power and 900,000 kg of topsoil (one acre-furrow-slice) to feed each of the six crew.
It doesn't have to be completely self-sufficient. They have 15 pressurized modules, I don't understand why 1 of them can't be dedicated to advance this vital technology.
It's important to remember that the ISS is a multi-national joint effort. If NASA wanted to convert a non-NASA module to grow plants, it would require the approval of that module's sponsoring agency.

I don't think a lot of people realize how inefficient the operation of the ISS is.

Take the JEM airlock for example. The JEM airlock is a Japanese (JAXA) module that has a novel slide table which is remotely controlled from Earth. So an experiment can be placed on the slide table, and then can be moved into space from Earth - without astronaut intervention. This offloads time from the astronaut's (packed) schedule to the ground control's schedule. However, NASA cannot operate this airlock! Only JAXA can! So, for the experiment to use the table, JAXA must agree to staff their ground control to operate the lock which has been a problem, causing extreme under-utilization of the lock. NASA is more than willing to take full operational responsibility of the JEM airlock to free up astronaut time, but as far as I know, JAXA hasn't been willing to train any NASA controllers.

*This information is going on 2 years old and things may have changed recently.

Hydroponics and aeroponics don't require any soil. Using red/blue grow lights instead of full spectrum roughly halves that solar power estimate, and in space they have no atmosphere to block the sun. The ISS certainly couldn't be self-sufficient as is, but it could be with relatively modest expansion.
New auditor's report say that humankind is not ready to leave the cave. We should all stay in the cave.
It was a mistake to leave the trees even.
Arguably, becoming air breathing has turned to not be all it was cracked up to be.
I've often seen two things conflated when it comes to cancer & Mars.

The first is that if we send people on a one-off return mission their lifetime risk of cancer will increase by X%. This is obviously a concern, but as others here in the thread have mentioned we could get people willing to accept the risk.

The second is that if people stay on Mars for generations you have the risk of cumulative damage to DNA over generations. Now it's not just a problem that instead of living to 80 you die of cancer at 60, but rather that after N generations the population's DNA might be damaged past repair.

Ok, exploring space is risky. How does it compare to the risk of previous exploration? I don't imagine that sailing off into uncharted waters in a wooden boat with whatever provisions you could carry was a particularly safe endeavor either. There are more important things than maximizing safety.
Lately I've been seeing NASA and NSA swapped with each other on a subliminal level once in a while and it always leads to some amusing double takes.