Survival on Mars: Radiation & Temperature Challenges

[1oldman2]

Go

Now that is an honest politician.

Gotta" love his face but you know his budget will be heavy on "pork barrel" spending.

 

[GTOM]

Im not exactly happy to see OFF topic here.

Whether it worth risking human life, i think a major point would be achieve manned interplanetary travel.

 

[1oldman2]

Im not exactly happy to see OFF topic here.

Good point, sorry

Whether it worth risking human life,

Space is always going to be high risk business.

i think a major point would be achieve manned interplanetary travel.

This will undoubtedly come to pass, however not prior to technological advances that rule out one way suicide missions.

 

[Mike the Mechanic]

I suspect that without proper prior preparation a manned spaceflight to Mars will be pointless (though perhaps prestigious in a doomed foolhardy sort of way) until a basic protected and sustainable habitat for humankind has been built there by robots (self replicating or otherwise). Perhaps the cycle of such robot engineering construction will need to start on the Moon, where a larger payload destined for Mars could be launched from that astral body's gravity than could be launched from Earth gravity.

Until such habitat preparation has been done a major concern in the meanwhile surely is what to do with dead bodies in spaceships and on planet Mars where conceivably such could trigger a contamination of its biosphere adversely affecting human life or any other alien life form which may already exist on Mars. Planet Earth is becoming more toxic to humans by the day, and hopefully we can avoid repeating the process on planet Mars, or elsewhere in the solar system.

I wonder if our NASA planners have considered a “transformer” type spaceship which dissembles itself on landing and then reassembles a significant portion of itself into individual functioning robotic modules which could stand alone or complement an internal payload of robotic components?

 

[rootone]

Before any manned mission to Mars is sanctioned. there would need to be placed on Mars at very least:
1. A fully functioning and well tested habitat building which I assume would be assembled by robotic missions.
2. A return journey launch rocket, prepared to be launched at short notice if need be, we can't presume that the vehicle used for the outgoing journey would be in a re-usable condition.

 

[256bits]

Astronauts are selected based on their health (among other things, but if you pick 5-10 astronauts out of 10,000 candidates you can be really picky in every aspect). They are under constant medical monitoring, have access to excellent healthcare in general, and get a good income. While they are also selected for taking more risks than average persons, the chance to die in an accident are not that high compared to medical issues. You would expect astronauts to have an above-average life expectancy.

It is certainly not alarming, and it is not even significant.

https://humanresearchroadmap.nasa.gov/Evidence/reports/Degen.pdf
And they do ( above average life expectancy ), at least for deaths due to circulatory issues.
Page 26 Fig 6 and table 7.

It isn't as if this ( radiation exposure and degenerative tissue effects ) is anything new to NASA.

 

[1oldman2]

It isn't as if this ( radiation exposure and degenerative tissue effects ) is anything new to NASA.

True, the report appears to draw a lot of it's data from historical nuclear events.
Great reading by the way.

 

[enorbet]

Regarding the question

is there anything there worth risking one or more human lifes <sic>

I am not in favor of any sort of battle between robots vs/ manned. They need to coexist IMHO. Please recall that on Apollo 17 that the Geologist, Schmidt, happened to kick the ground revealing an orange substrate of particular interest and import, for just one example of things that humans do that robots cannot be programmed for since programming implies foreknowledge.

There is a growing likelihood that life has existed on Mars at the very least in the distant past and this question is of almost unimaginable importance. It is also very likely that a human can discover and process things in an instant that would simply be overlooked/ignored by robots possibly shortening the search for answers on this immensely important front.

Additionally if we subscribe to the tenet that each of us own our own lives, and considering that millions of us risk our lives daily just by driving to work (as but one example) and also that many explorers from our past risked and even lost theirs, whether from immediate danger or long term unknown danger, why is this even a question? There can be little doubt that all of the known issues will be dealt with to the best of our ability before any (and especially any government sanctioned) exploration is attempted.

 

[rootone]

I am in favour of a human expedition eventually, only saying that a lot of useful work can (well must be) done first by machines.
Work that does have scientific value at no risk to life, and also work that is necessary to minimize risks for eventual human explorers.
I'm not sure about 'a growing likelihood of life having existed on Mars in the past'.
We do know now that there might have been times when the environment would be likely survivable by micro-organisms similar to those on Earth.
That isn't evidence that a form of life actually did exist though.
I understand that the next Mars rover planned by NASA will include some experiments which could detect organic chemicals.
residues which are far more likely to be associated with life than any other process.
If such residues are found then yes, preparing for human exploration should become more of a priority.
On the other hand, If there is complete absence of anything indicating life, then what could a human expedition achieve beyond what could be achieved by exploring the Sahara desert. Glory?

 

[GTOM]

I am in favour of a human expedition eventually, only saying that a lot of useful work can (well must be) done first by machines.
Work that does have scientific value at no risk to life, and also work that is necessary to minimize risks for eventual human explorers.
I'm not sure about 'a growing likelihood of life having existed on Mars in the past'.
We do know now that there might have been times when the environment would be likely survivable by micro-organisms similar to those on Earth.
That isn't evidence that a form of life actually did exist though.
I understand that the next Mars rover planned by NASA will include some experiments which could detect organic chemicals.
residues which are far more likely to be associated with life than any other process.
If such residues are found then yes, preparing for human exploration should become more of a priority.
On the other hand, If there is complete absence of anything indicating life, then what could a human expedition achieve beyond what could be achieved by exploring the Sahara desert. Glory?

Maybe we have to drill deep to find anything, whether life has existed or not, whether there are hidden water caverns or not.

 

[mfb]

They need to coexist IMHO. Please recall that on Apollo 17 that the Geologist, Schmidt, happened to kick the ground revealing an orange substrate of particular interest and import, for just one example of things that humans do that robots cannot be programmed for since programming implies foreknowledge.

Computers are Turing-complete. If computing powers and our knowledge how to program them were not limiting it, we could have computers make the same decision as humans.

We don't know how far away from intelligence (or "intelligence") we are - the current rovers are quite stupid. Currently a human could do in a day what the rovers do in a year, simply because there is no 10-30 minute delay between actions where a rover has to send its data and wait for human decisions. Future rovers might be much more intelligent and do more of those decisions on their own.

On the other hand, If there is complete absence of anything indicating life, then what could a human expedition achieve beyond what could be achieved by exploring the Sahara desert. Glory?

Learn more about Mars and planets in general, try to figure out why Earth got life and Mars not. We also send probes to various other objects where life is not possible at all.

 

[RussB]

I think there are a vast number of reasons to look to not only send an expedition to Mars, but ultimately colonize it.

One is the ostensibly hyperbolic ensurance the survival of the human species. On our planet there have been 5 mass extinction events. And those were caused by relatively natural incidents. Such events are still possible today and we've complimented them with a host of man made possibilities. Weapon based causes ranging from salted bombs, like a cobalt nuke, to engineered bio weapons could all do unimaginable harm. Imagine a deadly virus that spread though air or contact with no immediate symptoms and had an incubation period on the order of years. We're also manipulating our atmosphere with an extremely poor understanding of the longterm consequences. To believe that these sort of things could never possibly happen again or to us is simply hubris. As we expand outwards we give our species the best chances of success and of continuing to evolve.

It's difficult to know what you don't know. A tautology perhaps but important nonetheless! Many of our greatest discoveries been completely serendipitous. Observation of the cosmic microwave background radiation, the microwave, penicillin, radioactivity, pulsars, and countless others. Even relativity was discovered based on serendipitous conditions. Michelson-Morley inadvertently revolutionized physics with something as seemingly benign as measuring the speed of light from different perspectives. Visiting and staying on another planet is many orders of magnitude greater and different than anything humanity has ever done. How can we even begin to imagine, let alone name exactly, what discoveries are patiently awaiting their reveal?

And finally there is the most human aspect. Are you not driven to explore? I write this post half way around the world where I was born as wonder turned into wander. I simply am driven to know what I do not know, see what I have not seen, and experience what I have not experienced. And I think this drive is present in everybody, even if most of us must suppress it to some degree in modern society. That people can be enabled to go and want to go, even with a mutual and sober understanding of the risk involved, is really I think the most fundamental argument for it.

 

[enorbet]

I am in favour of a human expedition eventually, only saying that a lot of useful work can (well must be) done first by machines.
Work that does have scientific value at no risk to life, and also work that is necessary to minimize risks for eventual human explorers.
I'm not sure about 'a growing likelihood of life having existed on Mars in the past'.

There is no such thing as "no risk to life' only reduced risk. Those risks must be weighed against the likelihood of important value, which is itself a debatable quantity and quality. What is the tipping point? How do we determine when machines have done all they can do, efficiently, and that it is time for humans to step in? I completely disagree with mfb's notion that computers are currently "Turing complete" or more specifically that machine and programming will duplicate the human capacity for curiosity and serendipity as well as the ability to work "outside of the box' (respond to the unexpected) given that randomness is always at least tightly controlled if not eliminated in programming of computers whereas it is commonplace in humans. IMHO the need for human exploration exists now and will for quite some time and should only be held back until a reasonable level of safe success can be provided, which should be aggressively pursued.

We do know now that there might have been times when the environment would be likely survivable by micro-organisms similar to those on Earth.
That isn't evidence that a form of life actually did exist though.
I understand that the next Mars rover planned by NASA will include some experiments which could detect organic chemicals.
residues which are far more likely to be associated with life than any other process.
If such residues are found then yes, preparing for human exploration should become more of a priority.
On the other hand, If there is complete absence of anything indicating life, then what could a human expedition achieve beyond what could be achieved by exploring the Sahara desert. Glory?

While discovering evidence of Life may be "The Holy Grail" I think it must be remembered and recognized that understanding "Not A" also helps describe "A". Additionally while I think I understand the point you are trying to make of "low value targets" are you unaware of or have you dismissed how much could be (and has) been learned about Earth and Life on it by studying the Sahara Desert? It wasn't always a desert. Parts of it were seabed and have important fossils exceptionally well preserved in that now desert environment. There are the ruins of ancient cities and evidence of climate change as well as Plate Tectonics. How can anyone dismiss this as useless or low value? We are still learning how our solar system formed and any extraterrestrial data greatly improves that process.

In the case of Mars better understanding of the volcanic period, it's origin and end (why is Olympus Mons so large?) , as well as the also obvious origin of Valles Marineris and the loss of atmosphere and lack of magnetic field are almost as high on the list as Life. Ultimately, of course, knowing whether it could ever be possible to colonize Mars will one day be of extreme importance. We won't know if there is immediate "practical" importance until it is well explored. Who knows what important mineral deposits exist there? There are so many questions some of which beg for answers. The degree to which we can respond to that begging is directly proportional to our technology to provide safe travel for humans which includes more intelligent, better programmed machines as well.

 

[mfb]

I completely disagree with mfb's notion that computers are currently "Turing complete" or more specifically that machine and programming will duplicate the human capacity for curiosity and serendipity as well as the ability to work "outside of the box' (respond to the unexpected) given that randomness is always at least tightly controlled if not eliminated in programming of computers whereas it is commonplace in humans.

They are Turing complete, this is a mathematical statement and easy to prove.
If we stay at the mathematical level: assuming our universe follows some laws (sounds like a reasonable assumption), a computer can in principle simulate a whole human brain particle by particle. I am highly confident that such a detailed simulation is not necessary to produce intelligent programs, but this is just a matter of efficiency, not a fundamental limitation.

If you want randomness, you can add a random number generator. I don't see how that would help - a computer can also evaluate how useful kicking the ground can be, in the same way the astronaut did that unconsciously. Unconsciously does not mean randomly. Anyway, it is nothing computers cannot do.

 

[enorbet]

My apologies, mfb, I really didn't imagine you would use "Turing complete" in the strict definition sense since I don't see how it applies as some computational milestone being an abstract and not really a measure of a computer, or it's programming language, of recreating or competing with a human brain. I am aware that a rat's neuron was possibly successfully modeled but the Human Brain Project after massive funding and many years was an abject failure and has implied any such accomplishment is still a long way off, and even then will occupy considerably more space than a human body.. While this is likely, as you say, possible "in principle" I don't yet see any evidence that we are actually close to even passing the Turing test, let alone when such a test is over an extended period of time or, specific to this thread, practical in space exploration as a human equivalent replacement.

Beyond such sterile exercises and recognizing that it is also theoretically possible that almost nothing is truly random, even so-called random number generators, only possessing of a high degree of unpredictability and referred to as quasi or pseudo, it seems we agree that "kicking the ground" wasn't random in the strict sense, but I have little doubt that you know what I meant... or are you actually stating that a machine today or in the very near future can respond to complex, unexpected events? For example it is still a big step for self-driving vehicles from navigating a known course with extremely few opportunities for surprises than to be turned loose on unfamiliar public highways with rapidly changing traffic patterns, weather, accidents, and all the other low predictability events humans routinely contend with. AFAIK they can't even yet handle short shuttling trips on known routes with moderately limited access but still subject to the vagaries of human traffic and weather patterns.

It should be obvious considering that it has been some 50 years since computers beat Chess Masters that machines can beat humans at very specific tasks but by definition, Adaptive Software Development is also limited like this (from wikipedia) - " The characteristics of an ASD life cycle are that it is mission focused, feature based, iterative, timeboxed, risk driven, and change tolerant." but that is very different from a human lifetime of experience and adaptive behavior, implying that humans and computers may have specialties at which each may excel but not overlap since they will always behave differently both in theory and in practice.

The bottom line is that it will be a long time before it is possible to pack even just the computational power of a human in a similar sized package let alone duplicate the multiple areas of expertise and responsiveness to low predictive events, adaptability, that comes with human biology. For an interesting exercise try to calculate just how much and what kind of computational power would be required o provide the Apollo 13 solution... all of it, not just data checking. We can say that we now combine silicon and biology in space exploration but the time delay that increases if human are stuck on Earth is increasingly limiting.

 

[mfb]

How else can you mean "Turing complete"?

I mentioned limited computing power and our limited understanding of how to make intelligent programs several posts ago. But there is no indication that a somewhat intelligent computer would be impossible.

or are you actually stating that a machine today or in the very near future can respond to complex, unexpected events?

I don't say that, but I say it is possible, let's say in 20+ years. Can you prove that it is impossible?

 

[GTOM]

How else can you mean "Turing complete"?

I mentioned limited computing power and our limited understanding of how to make intelligent programs several posts ago. But there is no indication that a somewhat intelligent computer would be impossible.I don't say that, but I say it is possible, let's say in 20+ years. Can you prove that it is impossible?

Im not sure whether it could or not, but i bet that it would require a server room, and in case of a space travel, a number of spare parts, redundancy. And well I am not sure it is good idea to send a strong Ai to another planet.

 

[mfb]

This is not a binary thing. Better algorithms to avoid obstacles help Curiosity to be faster than previous rovers already, and the rapid development of self-driving cars here on Earth will probably help the Mars2020 rover to be even better. That alone speeds up exploration a lot already. A human-scale intelligence on Mars would be amazing (a human-scale intelligence on Earth would be amazing as well!), but it is not necessary.

 

[Chronos]

Sending humans to Mars at our present state of knowledge and technology would be an incredibly foolish act. We are nowhere near ready. We still need probes and remote data gathering devices to properly assess the myriad risks involved. No sane scientist would volunteer to explore the mariana trench by strapping him/herself to a cement block and shoving it over the side. And the sea is a forgiving mistress compared to the great abyss of space. We have insufficient knowledge to even properly assess the adquacy of martian resources to facilitate human survival. Unless we have the technology and resources sufficient to send mission appropriate supplies along with our intrepid explorers, we should at least furnish each adventurer with a little golden phonograph album, vis-a-vis Voyager, so future generations can better appreciate the scope of our hubris, ignorance and negligence.

 

[enorbet]

How else can you mean "Turing complete"?

Frankly since it confused me why the strict term would be brought up in this context I actually thought you were referring to some loosely defined milestone of raw computational power as a result of Moore's Law iterations.

I mentioned limited computing power and our limited understanding of how to make intelligent programs several posts ago. But there is no indication that a somewhat intelligent computer would be impossible.I don't say that, but I say it is possible, let's say in 20+ years. Can you prove that it is impossible?

No. Nor would I try especially in this context. There is often a huge gap between possible and practical. It is my estimation that there is a fundamental difference if not in how the human brain works vs/ binary computers at the very least in the manner of programming and that the two will likely never be equivalent. Thus, both have applications in different areas where one is more efficient than the other. There is no one winner. The likelihood is that cooperatively working tenders the greatest, most efficient benefits seems highest.

In the case of space exploration there are far too many benefits to discovering how to create safe environments for humans under extreme conditions as well as too many benefits in the exploration to entirely abdicate to machines.

 

[enorbet]

ping Chronos - It has been noted even by most proponents of human exploration here that we are not yet ready to send humans to Mars for exactly the reasons you have stated. I am of the regarded opinion that those who insist we are ready or will be in 10 years or so are either unaware of the actual and numerous difficulties or falling prey to fantasy or wishful thinking. However I also noted that we are applying extremely little effort in using our machines to address human survival issues. Instead machines have been used to supplant human exploration rather than also assist it and this seems foolish and shortsighted to me.

 

[mfb]

Sending humans to Mars at our present state of knowledge and technology would be an incredibly foolish act. We are nowhere near ready.

No one plans to do that. And we don't have any system that could do so. I don't see the point of arguing against something no one proposes.

It is my estimation that there is a fundamental difference if not in how the human brain works vs/ binary computers at the very least in the manner of programming and that the two will likely never be equivalent.

You can build different computers, if necessary.

As far as I am aware, Elon Musk is the only one with a 10 year estimate, and I think even he knows that this is unrealistic. Space agencies usually say 2035+. I don't count Mars One here as I always saw them just as PR gag.
Previous Mars missions were launched knowing that manned missions are still far away, and that initial missions will bring all the life-critical systems with them - there is no need to spend a lot of payload on things only relevant for human spaceflight. The proposed 2018 Dragon mission should be different, it will land a much heavier payload than earlier missions (several tons, instead of 900 kg) with a spacecraft that is (apart from mission-specific modifications) human-rated. No crew will use a Dragon capsule to fly to Mars, but it is still a big step forwards.

 

[enorbet]

You can build different computers, if necessary.

Of course but to what end if you accept that humans offer something that computers can't duplicate, whether size, cost, or adaptability? Your point in an earlier post about faster rovers also doesn't address the point that humans ultimately require control over rovers (whether ideal or not) and the only way to overcome communication time currently is to have humans closer.

As far as I am aware, Elon Musk is the only one with a 10 year estimate, and I think even he knows that this is unrealistic. Space agencies usually say 2035+. I don't count Mars One here as I always saw them just as PR gag.

Or possibly he is taking a tip from the JFK playbook and issuing a challenge to kickstart the process or demonstrate widespread interest and support.

Previous Mars missions were launched knowing that manned missions are still far away, and that initial missions will bring all the life-critical systems with them - there is no need to spend a lot of payload on things only relevant for human spaceflight. The proposed 2018 Dragon mission should be different, it will land a much heavier payload than earlier missions (several tons, instead of 900 kg) with a spacecraft that is (apart from mission-specific modifications) human-rated. No crew will use a Dragon capsule to fly to Mars, but it is still a big step forwards.

How much payload is too much to begin gathering data on what will actually be needed for human survival so planning is even possible? It is my opinion that NASA is still nursing deep wounds over public deaths during 2 Shuttle missions (as well they should given how obvious in retrospect their carelessness was on both of those - examples - cold temperatures affect seals, KE=0.5mv^2) which had far greater shame than the waste on the initial Hubble mirror. It may be instructive to note that as horrific as it was the public soon moved on from the Apollo Oxygen Fire Tragedy largely because they didn't see it occur. Michael J. Adams' death in the X-15 project was all but unknown to the public as are 10s of other space exploration deaths not in public view.

The point is that it seems as if funding is too closely tied to (squeamish and fickle) Public Relations, both pro and con, when it needs to be stated and understood that exploration is by nature both dangerous and necessary.

 

[RussB]

As far as I am aware, Elon Musk is the only one with a 10 year estimate, and I think even he knows that this is unrealistic. Space agencies usually say 2035+. I don't count Mars One here as I always saw them just as PR gag.
Previous Mars missions were launched knowing that manned missions are still far away, and that initial missions will bring all the life-critical systems with them - there is no need to spend a lot of payload on things only relevant for human spaceflight. The proposed 2018 Dragon mission should be different, it will land a much heavier payload than earlier missions (several tons, instead of 900 kg) with a spacecraft that is (apart from mission-specific modifications) human-rated. No crew will use a Dragon capsule to fly to Mars, but it is still a big step forwards.

Dragon 2018 is not sending several tons, it's sending 8-10, making it a near magnitude order of payload contrasted against NASA. And proving the viability of retropulsive landing will open the door to far greater to come. Elon's timeline is also 8 years, not 10. He plans the first human launch in 2024, and I do think he believes it is still achievable. And while organizations like NASA must remain conservative, even if only for the sake of politics, their actions are also in line with the 2024 timeline as well. For instance as early as 2008 they commissioned a study on the optimal energy mechanism for a sustained colony on Mars (Cooper et al). Given the predictably unpredictable (but large) gains they would reasonably expect in renewable energy resources in the three decades from the time they commissioned that study, it'd be somewhat illogical for them to make such expenditures if their timeline was actually the mid 2030s. Similarly, they're also currently doing terrestrial tests of things like growing potatoes on Mars when further information about precise conditions that ought be expected would enable such experiments to be far more accurately designed and informative. NASA has also already announced one of the key tests for the Mars 2020 intentions to deploy ISRU tests for on-site production of oxygen + fuel. Again, so many of these experiments are putting the horse before the cart if you're really looking at a mid 2030s deployment.

I suppose the most telling thing will be in what SpaceX decides to deploy in 2018. So far as I know they've announced nothing, but if they're sticking to 2024 then that means they have exactly 3 windows for launches, including 2018, before humans go. And we'd need proof of a large number of technologies, and redundancies, before then. Food, habitation/life support, and energy production at the bare minimum. The longer they're tested for, the better. If they are not racing towards deployment and tests of such technologies even as early as 2018, then it's a safe bet to say that Elon probably does not believe in the 2024 timeline.

 

[mfb]

Of course but to what end if you accept that humans offer something that computers can't duplicate, whether size, cost, or adaptability?

I don't think there is anything that humans have that is fundamentally irreproducible with technology.

Your point in an earlier post about faster rovers also doesn't address the point that humans ultimately require control over rovers (whether ideal or not) and the only way to overcome communication time currently is to have humans closer.

Faster rovers are a (large!) single step. We'll see more in the future.

How much payload is too much to begin gathering data on what will actually be needed for human survival so planning is even possible? It is my opinion that NASA is still nursing deep wounds over public deaths during 2 Shuttle missions (as well they should given how obvious in retrospect their carelessness was on both of those - examples - cold temperatures affect seals, KE=0.5mv^2) which had far greater shame than the waste on the initial Hubble mirror.

The Dragon mission is not a NASA mission. There is not just NASA. I don't understand your first question here. What do you mean by "too much"?

@RussB: The mass doesn't seem to be fixed yet, so I said "several tons". Musk said "an order of magnitude more than Curiosity".

Again, so many of these experiments are putting the horse before the cart if you're really looking at a mid 2030s deployment.

Not really, development cycles can be long: what if a project shows some critical issue, and you need a follow-up mission to test the system again? Then you need at least two years for the next mission, and if you need more development time you might miss that launch window and get 4.5 years delay in total.

MCT won't be ready for a 2020 launch (statement from Musk somewhere), so the earliest MCT launch would be 2022, arriving at Mars 2023. That gives ~1.5 years to see a MCT on the surface before you launch humans to Mars. That is not sufficient time to produce enough fuel to launch the MCT again - you would send humans to Mars without testing the Mars ascent stage! It also means you have to test fuel production on site with the 2018 mission - testing it with the 2020 mission and implementing it on a large scale for a 2022 launch is ... optimistic.

SpaceX has a long history of delays. They are fast, but not as fast as their plans.

 

[GTOM]

I don't think there is anything that humans have that is fundamentally irreproducible with technology.

Fundamentally no, however, human behavior is undeterministic, thus unpredictable, i sure wouldn't want to build such a supermachine (a computer with a different working principle). I don't expect deterministic machines to be truly able to reproduce human innovation, outside the box thinking, creativity anytime soon.

 

[enorbet]

I don't think there is anything that humans have that is fundamentally irreproducible with technology.

Even if I agree that might possibly be true, it begs the questions "when?" and "at what cost?". We already have some qualified people willing to go once there is a reasonable margin of safety. What would cost more? - providing a reasonably safe environment or recreating human adaptability and expertise with similar size and power requirements? I can't see how a proposal that we "stick with machines only" has a better cost/benefit ratio or faster timeline, let alone historical value and stimulus on funding.

Faster rovers are a (large!) single step. We'll see more in the future.The Dragon mission is not a NASA mission. There is not just NASA. I don't understand your first question here. What do you mean by "too much"?

My point regarding "too much payload" simply refers to how little it would take to gather at least some data on the means to provide human safety as opposed to what appears to be sole commitment to dependence on machine exploration.

Additionally, the example of Schmidt (geologist-astronaut) was not only "irreproducible" then but I contend that it is now and will be for a fairly long time - certainly longer than 20-30 years. A human on the ground can make important observations (humans are innately curious) , respond to serendipity, and make immediate decisions. The only negative aspect to Apollo 17 employing humans was the lack of time in the field caused by biological needs compared to machines, given the state of technology then.

It seems ideal to me to have humans in situ with machines to do the follow up work to expand time in field specifically at sites with the most value to human knowledge. Reporting back to humans who are not present and in direct contact, including that inescapable delay, or giving machines human levels of autonomy cannot possibly compete, nor would many people welcome such machine autonomy. That issue is fraught with controversy at the very least, as well as concern and even abject fear. It's not going to happen anytime soon.

 

[mfb]

I don't think there is anything that humans have that is fundamentally irreproducible with technology.

Even if I agree that might possibly be true

The alternative would be Vitalism.

it begs the questions "when?" and "at what cost?".

Those questions will find answers within the next decades. I don't claim to know the answers today.Exploring the chemical composition of all sorts of stuff contributes to the preparations for a potential manned mission already, and it is the standard research program.

A human on the ground can make important observations (humans are innately curious) , respond to serendipity, and make immediate decisions.

A computer can compare new objects to old objects and report unusual results to humans or directly take a probe. That is something they can do today.

[...] or giving machines human levels of autonomy cannot possibly compete

Absolutely no possibility? Do you really think you know it better than the experts looking into exactly this?

 

[1oldman2]

In pondering the article (http://www.nature.com/articles/srep29901) that I mentioned in post #188, I got to wondering about whether or not there were any Solar events during the Apollo flights, I followed each as they took place and couldn't remember any flairs occurring, although it was a popular topic in the "what could go wrong" department.
In short order I came across this somewhat dated but very relevant article. I suspect on Mars an underground radiation shielded environment would be just the thing for a major solar event if humans were involved.

From, http://history.nasa.gov/SP-368/s2ch3.htm
Solar-Particle Radiation
"No major solar-particle events occurred during an Apollo mission. Although much effort has been expended in
the field of solar event forecasting, individual eruptions from the solar surface have proved impossible to
forecast. The best that can be provided is an estimate of particle dose, given visual or radio-frequency (RF)
confirmation that an eruption has occurred. A system of solar-monitoring stations, the Solar Particle Alert
Network (SPAN), provides a NASA-sponsored network of continuous data on solar-flare activity. SPAN
consists of three multiple-frequency radio telescopes and seven optical telescopes. The network gives data for
determining the severity of solar-particle events and the resultant possible radiation hazards to crewmen. After
the appearance of particles is confirmed onboard a spacecraft , protective action can be taken.

In terms of hazard to crewmen in the heavy, well shielded Command Module, even one of the largest solar-
particle event series on record (August 4-9, 1972) would not have caused any impairment of crewmember
functions or ability of the crewmen to complete [108] their mission safely. It is estimated that within the
Command Module during this event, the crewmen would have received a dose of 360 rads* to their skin and 35
rads to their blood-forming organs (bone and spleen). Radiation doses to crewmen while inside the thinly
shielded Lunar Module or during an extravehicular activity (EVA) would be extremely serious for such a
particle event To monitor particle activity, a nuclear particle-detection system (figure 3) was designed to have a
relatively narrow acceptance angle. It measured the isotropic proton and alpha particles derived from solar-
particle events."

 

[rootone]

I do think there are some things humans can do which would be very difficult for a machine to emulate.
As a way-out -there example, an astro-biologist might recognise a rock formation similar to a stromatolite, or even some kind of fossil which would not seem very interesting to a fully autonomous machine.
(Not that I think these kind of things have much chance of existing on Mars. but it's not impossible)

 

[enorbet]

I don't think there is anything that humans have that is fundamentally irreproducible with technology.

The alternative would be Vitalism.

It seems you have taken the point of view of the theoretical, the "in principle", as opposed to my view of the practical and efficient. I'm not claiming there is some fundamental difference between what we consider to be alive and what doesn't fit our limited definition of Life. I'm saying that it is entirely impractical to program a computer exactly in the manner that humans are programmed and that each has it's advantages and drawbacks.

Additionally, and possibly for another thread in another section, is the fact that there is a fundamental difference in computational methodology in that, so far, computers are strictly limited by design to function in binary and that we don't know yet exactly how the human brain functions, not even what consciousness is and why there was apparently an evolutionary advantage to it. While theoretically it may be possible that a computer can achieve consciousness, it is not yet an accomplished fact nor do we know if that would be a net gain, at the very least to humans. It is also not known if human lifetime's worth of slow, largely uncontrolled programming would be of any benefit to a binary brain(s). We also can't discount that there exists billions of humans, each with different sets of programming who can communicate and learn from each other in a manner that is very different from mere data sharing.

As for the "when?" and "at what cost?" that we may get some answers in the coming decades is interesting for certain but since these important questions are unanswerable today I submit it is impractical to rely on that hope instead of using what we now do know. We know that Man can survive for some time in Space and on another world and that humans bring something to the table that machines cannot currently duplicate and while it is presently deemed too costly to extend that to Survival On Mars, that is a matter of economics and the assumption that our money is better spent elsewhere.

Absolutely no possibility? Do you really think you know it better than the experts looking into exactly this?

"This" being autonomous computers... presumably able to alter their own programming and that of other computers? With sincere respect I must ask, "What experts?" Since such computers don't yet exist and nobody fully understands how a human brain functions (as noted above), how can anyone be a true expert other than "in theory" and in this case "theory" is much more like "hypothesis" since there is extremely little relevant test data as nothing yet exists upon which to apply testing.

Plus, again with the practical, just how long do you suppose it will be before there can be any agreement to allow not one, but many computers to actually learn from and reprogram each other in any human-like, autonomous manner and possibly to build others even more powerful? It may well come to pass but I strongly suspect it will be quite far in the future and fought "tooth and nail" the whole way just as Elon Musk, Stephen Hawking, and other influential, expert people are already engaged in "putting the brakes on" such a proposal. If you think public outcry altering NASA's plans from a few deaths was to be expected (and possibly justified), just wait till there is even one death remotely connected to self-programming, autonomous computers occurs. One doesn't have to be an expert at anything to calculate how that movie ends, even if only for a time.

It seems we have some fundamental disagreement and I'm not at all sure why. Are you categorically opposed to humans in Space? not recognize the value of a tight team of both? or just basically believe that machines are superior, or soon enough will be, in every way?

 

[GTOM]

I thought a bit about the human vs computer issue (i write my current thoughts here, if the topic continues, create another thread)
So in case of a present day neural network emulation program, we don't know how exactly they react in unforeseen situations, most likely they fail (i saw the gif, robodog meets banane on the floor) but we can be sure, they will never came to such conclusions, that self driving car throws out its passangers to help its survival.
But if they can react to unknown situations, connect far away things like humans do, they can come up with such conclusions. It might be helped by a supervisor code like Asimov's laws, but a strong AI can still circumvent them, but my passangers weren't humans, they were experimental dolls.
So, if we want to have programs that can be more human like, we have to be prepared to assume, they will be unpredictable, and we expect the opposite from machines.

So while i don't say, that "true" AIs are theoretically impossible (i have doubts that they could be fitted into an android body rather than a large server room) but i definitally wouldn't want to use them to explore unchecked by humans on the site.

Of course it perfectly makes sense to use low level AIs to explore a really good site and construct a base before humans reach Mars.

 

[Chronos]

Point well taken. It appears we are all in basic agreement..

 

[mfb]

It seems you have taken the point of view of the theoretical, the "in principle"

I had this view the whole time and expressed this multiple times (e.g. with "fundamentally irreproducible"). That is the only level where we can make clear predictions.

computers are strictly limited by design to function in binary

You can simulate non-binary systems with binary systems, if necessary.
There is no need for consciousness, that is an ill-defined concept anyway.

We currently do not have robots that perform human-like research tasks on Mars, but we currently also do not have humans on Mars. Both could be viable in the future, so both options should be investigated.

"This" being autonomous computers... presumably able to alter their own programming and that of other computers? With sincere respect I must ask, "What experts?" Since such computers don't yet exist and nobody fully understands how a human brain functions (as noted above), how can anyone be a true expert other than "in theory" and in this case "theory" is much more like "hypothesis" since there is extremely little relevant test data as nothing yet exists upon which to apply testing.

There are AI experts in the same way as there are experts working on systems that can bring humans to Mars.

There is no need to make the AI truly human-like to make it better at exploring Mars. Just better in performing research is sufficient.

I'm not opposed to humans in space, and I support the efforts to plan a manned mission to Mars, I don't see how you got the opposite impression. I just don't see where you get the confidence from to say that robotic missions will never be able to do tasks humans can do.

 

[enorbet]

Thank you, mfb, I see your position much more clearly now and find we aren't as far apart as I once concluded. My point is quite small and not projected far into the future, being reduced to current exploration of Mars and the fairly immediate future. We apparently do agree that there is some fundamental difference between humans and machines, whether we conclude it is from the difference in how they are programmed or how they compute. There is currently also a difference in the size and mass relative to adaptive computational power. If there was no difference, there would be far less reason to even consider sending humans... at least until machines and humans have progressed to the point that machines have achieved and been granted the value afforded to things "Alive". So my point has been simply that I think it would be wise to include more research into what is needed to provide a relatively safe human environment in Mars exploration, rather than what appears to me to be more like exclusive substitution.

As for my confidence I didn't say nor do I subscribe to "never", especially "in principle'. My point was simply one of practicality. I can't escape that I majored in Engineering not Theoretical Physics or Computer Science. Short of "never" I don't assume any level of confidence past much more than a decade or two. That too is the nature of good predictions, right? By definition, the further in the future we predict, the greater the margin for error.

Tangent - One possibly important area of difference between men and machines is the response when predictions are in error. I wonder how any machine now or in the near future would deal with fundamental errors such as discovered by Vera Rubin in galactic rotation, or the various Nobel Laureates seeking to measure how much expansion was slowing down. It would take a savvy programmer to "if then this" such deep prejudice. It may be a cause for pause and check in humans but may cause a computer to crash :)

Anyway, good conversation... made me think a lot.