The SpaceX Thread

[Dmytry]

It'd seem to me that to literally burn a hole through a spysat's imaging sensor (focussing by the spysat's own optics), all you'd need is a $10 000 tracking telescope and a $2000 pulse laser. Assuming the mirror size of 30cm, and wavelength of 1 um, the mirror is 300 000 wavelengths across, so the diffraction limited spread at 1000km distance should be 1000km/300 000 = 3.3 meters. edit: and no adaptive optics, the laser just keeps pulsing and pulsing and it works like the reverse of "lucky imaging". Unlucky imaging perhaps. I presume there's some counter-measures in place.

 

[Dmytry]

Geostationary orbit internet would have rather bad lag... especially with the current idiot way of building websites where there's a multiple back-and-forths until you get anywhere (get the page, get the script, get another script, and another one, then start fetching images, etc). Still better than no internet, of course. I'd love to know in advance when something like that is rolling out for certain, there will be a significant demand for software that would optimize website loading.

 

[Dmytry]

All Chinese do is adding a little yet noticeable inconvenience to unfiltered internet access. Their firewall is very easy to bypass, AFAIK, but those that would bypass it aren't exactly the people who would have their views changed massively by the unrestricted internet - they'd mostly see things they know already.

edit: More to the topic. The Dragon v2 propulsive landing onto a landing pad (with no parachute deployment at all) strikes me as something that may be unnecessarily risky (compared to Soyuz style descent where if the landing cushion engines fail at any point the astronauts still survive with the parachute alone). The helicopter-like accuracy is at odds with safety - if you're coming down with engines alone, there's a point where it would be too late to deploy a chute in the event of failure, and if you're coming down with a chute your landing location is up for the wind to decide.

 

[Dmytry]

tinyMan[/url]"ezh8uay]

edit: More to the topic. The Dragon v2 propulsive landing onto a landing pad (with no parachute deployment at all) strikes me as something that may be unnecessarily risky (compared to Soyuz style descent where if the landing cushion engines fail at any point the astronauts still survive with the parachute alone). The helicopter-like accuracy is at odds with safety - if you're coming down with engines alone, there's a point where it would be too late to deploy a chute in the event of failure, and if you're coming down with a chute your landing location is up for the wind to decide.

This is covered a bit earlier in the thread, if I remember correctly.

Basically, the Dragon v2 has 2x (3x?) the number of thrusters required to land safely and executes a test fire of those thrusters during the descent phase to ensure that a safe landing is possible. Remember, the Dragon V2 has to have enough thrust to out-accelerate a Falcon during ascent for emergencies and requires only a tiny fraction of that power to handle landing.

Hmm, okay. That sounds a lot better then. It's not just the thrust, though, the thrusters would have to fire for a sufficient time to bring the capsule from the terminal velocity to a complete stop. And the thrusters must activate at the correct altitude, with high precison, some failure of what ever sensors are involved would be completely lethal.

 

[Dmytry]

jbode[/url]":2nnx5j80]

Dmytry[/url]":2nnx5j80]More to the topic. The Dragon v2 propulsive landing onto a landing pad (with no parachute deployment at all) strikes me as something that may be unnecessarily risky (compared to Soyuz style descent where if the landing cushion engines fail at any point the astronauts still survive with the parachute alone). The helicopter-like accuracy is at odds with safety - if you're coming down with engines alone, there's a point where it would be too late to deploy a chute in the event of failure, and if you're coming down with a chute your landing location is up for the wind to decide.

Chutes can fail too.

The SuperDracos are hypergolic; as long as the valves work, you should be golden. If they don't, you have the chutes as backup (IIRC, the chutes will always be there, but are not planned to be deployed except in the case of an abort, where all the delta-V in the SDs is spent getting away from the booster). Obviously some effort will be expended to make the SDs super-extra reliable and minimize the possibility of failure.

What I would be more worried about is things such as measuring altitude without fail.

Besides, I don't see how this is any riskier than a dead-stick landing in a brick with wings.

I wouldn't regard that brick with wings as safe either... http://en.wikipedia.org/wiki/List_of_sp ... paceflight . Deadliest thing to ever fly to space.

 

[Dmytry]

smartalco[/url]":2ej8fz3h]

Dmytry[/url]":2ej8fz3h]I wouldn't regard that brick with wings as safe either... http://en.wikipedia.org/wiki/List_of_sp ... paceflight . Deadliest thing to ever fly to space.

Neither of its two failures were caused during descent though (I realize one failure happened during descent, but that was caused by earlier events).

Well the remark was mostly regarding the whole operation. The whole thing is so complex with so many points of failure that a couple post-crash fixes may barely make a dent in the total risk. Don't get me wrong, shuttle is a marvellous and amazing piece of engineering, but it's just a very unsafe approach.

The Dragon v2 has what, 8 thrusters? I'd take a guess that any pair of two opposite engines would be enough to slow it down below lethal landing speeds (even if it isn't a soft and graceful with just two). Someone who understands the math better than me can probably figure it out; each super draco has 16,000 lbf of thrust, and the capsule can weigh up to ~15,000 lb during reentry, all according to wiki. (This also requires a guess as to its terminal velocity, which I'm not even going to try to guess)

Probably needs at least three, and not just any three, to maintain the orientation, unless the thrusters are also vectored. As I said, it's not even so much that a thruster can fail to start, but that now you need very reliable and accurate altitude measurement.

 

[Dmytry]

By the way, how do they measure altitude? Radar altimeter? How reliable that would be, considering it has to work through the heat shield or pop out?

 

[Dmytry]

MilleniX[/url]":1xtu3g0w]GPS isn't a very good source for altitude, and loses positional precision as you get further from the Earth's surface. An inertial system is OK, but it accumulates error as it moves without being able to update against a known point of reference. For altitude measurement, I think the radar altimeter will have to be the primary.

Yeah. Inertial requires you to know your starting altitude. Laser is exactly like a radar altimeter that doesn't work in fog.

For velocity measurements, on the other hand, everything is much easier. GPS can tell velocity based on doppler shift much more precisely than position. INS also only has to integrate once to calculate velocity rather than twice, cutting the error rate.

Depending on how much thrust and delta-v they have available, they could take a kind of 'brute force' approach to making a soft landing on a hard pad: from some relatively high altitude at a point roughly over the target, descend steadily at a constant, fixed velocity. E.g. from roughly 100 meters up, fall at 2.5 m/s.

That'd be 40 seconds at 1g, for delta v of 400 m/s ... multiply by two if you want +-100m tolerance. Quite a bit, but not entirely outside the realm of the possibility. I don't see, though, what sort of altimeter would benefit from that. The barometric altimeter seems like still too much of a risk even if you can absorb +-100m of error. AFAIK, you're not supposed to let your passengers die if GPS doesn't work. And the radar altimeter, when it's working at all, is quite precise, with only a small % error depending on air density etc.

Soyuz only activates it's thrusters at a few meters of altitude, and it uses reflection of gamma radiation from the ground (so it will essentially ignore things like thin roofs, vegetation, etc).

 

[Dmytry]

F16PilotJumper[/url]":1ujvhf75]

** Let me know it it's a faux pas to use ISP for a stage or overall rocket instead of just an engine.

ISP typically refers to an engine alone. For an expendable stage there is an burnout mass (mass of engines + support structure + mass of stage + unuseable propellant) and the total mass at ignition (burnout mass + mass of propellant consumed). Knowing those three numbers, the deltaV of the stage can be calculated using the rocket equation.

Even though Merlin has a lower ISP than RD-180, it has a significantly better thrust to weight ratio - ~155:1 for Merlin vs ~77:1 for RD-180. So an RD-180 stage with similar thrust will have to carry 2x the engine mass as part of its mass at burnout, reducing the deltaV capability of the stage.

Well, when the engine's weight is small compared to the fuel weight, the thrust to weight ratio of the engine itself is not going to make much of a difference to the overall delta-v even compared to some engine of zero weight. Not sure what fraction of the rocket mass is RD-180 typically, but I would guess quite small.

edit: with regards to the 'cheap launches' business strategy of theirs, it is not clear how far down you can get Proton's prices and still make a profit, but I would imagine that if necessary, prices can go down quite a lot while keeping it profitable. It's not like Russians can't switch to 3D printing for some parts; it's just that they have all those highly qualified machinists making a better product (at a premium) and they don't want to put them out of job unless that's unavoidable. Cutting down costs and simplifying production is something that's traditionally held in very high regard over there, it permeates education. The launch prices worldwide strike me as something akin to the internet prices in the US. You can have 10x faster internet for 1/5 the cost, and US ISPs are absolutely capable of doing just that, they just don't want to because they make more profit the way they do things now. So here comes the SpaceX with ultra low launch costs, and then Russia makes some changes to their manufacturing, outsources some stuff to China, and offers slightly cheaper launches, it seems to me that this is something that can happen.

 

[Dmytry]

dio82[/url]":3or3p6ru]Edit:
I don't want to diss SpaceX's work. Their CFD work is top-notch and absolutely cutting-edge technology. I am just a bit baffled that they are investing so much work in "re-inventing the wheel" kind of work.

Well, there's also that it's cool. E.g. I was toying with the idea of working out a much cheaper maglev bearing motor (utilizing the cheap but quite high powered CPUs to control inconvenient configurations), and I'm sure there's simulators for that, but since I'm a programmer I would be heavily inclined to write my own sim just because I can, instead of using existing software, even though business wise it's not smart to do. Also he may have some growth in mind so he wants to get started on doing his own CFD. You can do some extremely amazing work with just a few good people, it's a very minor expense if you have the right people, if you're lucky to have them then best to hire them up than to let the chance pass away.

 

[Dmytry]

CH4 is still cryogenic and not as good as liquid hydrogen... dunno if there's a point unless they manage to cut all the costs so low it begins to matter.

 

[Dmytry]

Even though it is "mildly cryogenic" it is still cryogenic enough you probably won't be taking it long way across the solar system. But yeah it does sound like a ground in the middle between liquid hydrogen and kerosene. Still, my understanding was that the fuel cost was a relatively small fraction of the launch cost even for lh2 fuelled rockets... maybe I'm wrong though.

 

[Dmytry]

To date, a crewed launch escape system had activated exactly once in real world conditions.

This failure is very bad news for SpaceX's manned spaceflight hopes, whenever this exact explosion that happened would have been survivable with an escape system or not. Launch escape is an absolute last ditch effort, which may or may not work depending to whenever you get to predict the failure (which you failed to prevent), and can not be relied on for normal operation.

 

[Dmytry]

The thing is with emergency escape system, is that it doesn't remove the need for a very reliable launcher. An escape system works in unpredictable circumstances and can not itself be expected to have a very high success rate. Also, with the explosion the question is not just whenever people would survive the explosion, but whenever the capsule would not be too damaged to land, and not just this specific explosion but similar explosions occurring at a lower altitude and speed (where the capsule is more at risk of damage by the explosion). Worst case for a capsule is probably when the rocket blows up on the launch pad or immediately after launch, in some asymmetric fashion with pieces tumbling and shrapnel flying. Especially bad if capsule doesn't separate in advance.

 

[Dmytry]

.劉煒[/url]":1771jqhq]No possiblity of using the door and parachuting out?

Well, if it blows on the launch pad then there's no time & everything is in flames. edit: or you mean Challenger? I would think it was tumbling far too fast & little time to do anything.

 

[Dmytry]

Xavin[/url]":ubxvxx48]

Or, hypothetically, if it was a manned dragonrider capsule spacecraft during CRS-7, and could trigger the parachute sequence instead of going plop into the ocean unbraked.

Except that if you had humans on board, you would want that parachute to open regardless of whether the humans were conscious or on the ball, and make sure they didn't open it early or late (that window is going to change drastically based on altitude and speed). Go watch the video of the dragon abort test. I wouldn't want a human that was on that ride anywhere near important controls, and it will only be crazier for an abort mid-launch. Occasionally, humans can do something the programmers and product managers didn't think about, but usually they just make mistakes. We have a laundry list of airliner crashes that happened because the humans though they knew better than the computers, and they were wrong.

Slightly off the topic, I was reading about a plane crash where the autopilot got accidentally disengaged, and apparently in many planes the presumption is that a panicking human who forgot about the autopilot shut off button knows better than autopilot - so if you nudge the steering column too hard (e.g. when picking up something off the floor or going to the loo, or dealing with another emergency), the autopilot will shut off, and in some planes that will not result in a warning sound but only an easy to miss warning light.

 

[Dmytry]

blargh[/url]":vqdyx6se]

Dmytry[/url]":vqdyx6se]
Slightly off the topic, I was reading about a plane crash where the autopilot got accidentally disengaged, and apparently in many planes the presumption is that a panicking human who forgot about the autopilot shut off button knows better than autopilot - so if you nudge the steering column too hard (e.g. when picking up something off the floor or going to the loo, or dealing with another emergency), the autopilot will shut off, and in some planes that will not result in a warning sound but only an easy to miss warning light.

You're probably thinking of Air France Flight 447, about which there has been some quite good reporting recently.

Also, https://en.wikipedia.org/wiki/Eastern_Air_Lines_Flight_401 , https://en.wikipedia.org/wiki/Aeroflot_Flight_593 , and I think I seen another one. The A310-100 way of doing things is particularly weird - the autopilot disengages if you are physically wrestling the autopilot for 30 seconds (it silently counts down those 30 seconds). It seems to me that wrestling vs autopilot without switching it off may not even be indicative of a desire to switch the autopilot off. It's like Clippy, "it seems that you are wrestling the autopilot, would you want to switch it off?", except it just does without asking.

 

[Dmytry]

Well it definitely wouldn't do any good if during an approach with ISS the autopilot just quits because the pilot accidentally nudged the controls. Switching from autopilot to manual is probably fine if very explicit and obvious...

edit: especially noting that not only can you manually dock with ISS, Russians manually docked with dead-as-a-brick Salyut 7 (Remember the docking in Interstellar? That's probably a fictional dramatization of Soyuz T-13 docking to Salyut 7 . It wasn't spinning so hard of course, but it was entirely unpowered so it had been tumbling).

 

[Dmytry]

smartalco[/url]":30ear234]I'm not a plane guy, so these questions are probably stupid/obvious, but I'm going to ask anyway

Technarch[/url]":30ear234]they don't know they're landing with a tailwind

Why don't they? This doesn't seem hard. Get weather report from airport with wind map of the entire area (cover the entire airport area in mini weathervanes). Compare airspeed with ground speed (I'm not sure planes have a reliable way to measure their own ground speed (GPS accurate enough?), but within landing distance of an airport, I'd imagine there's a way to get fairly reliable and fast positional data to extrapolate from). Is not having this just a 'all our technology is 40 years old' thing, or what?

they don't know what to do when the pitot freezes.

I don't understand why a single pitot freezing over should ever be a problem.

My understanding is that the single pitot tube freezing over is not itself a problem as much as an indication of a problem, because all the other pitot tubes are in the same condition and are going to freeze over as well. But yeah it shouldn't be that big of an issue when the GPS is still working and there's weather measurements at the airport.

 

[Dmytry]

Xavin[/url]":5f9ch297]

especially noting that not only can you manually dock with ISS, Russians manually docked with dead-as-a-brick Salyut 7

I'm not saying it's impossible, because it's clearly not, but that doesn't make it good procedure or worth the risk. The computers don't have to be perfect, they just have to be better than humans, which is not hard.

You can do better still by detecting the instances when the computers are doing worse than humans and falling back to human control (In a way that is obvious and clear, none of the tiny light lits up BS).

Let's suppose ISS suffers a major failure where it's attitude control is broken. Things go wrong in space quite often, it's not too far fetched. Nobody's going to be writing software to handle that - best handled by a human acting directly than by a human writing software that will maybe get actually used once. If software was to handle that, it'd be an edge case where software is going to be far far less reliable than it normally is. As of right now, if something goes majorly wrong and ISS ends up literally tumbling, humans can dock to it, and software can't.

 

[Dmytry]

Hat Monster[/url]":3uvbt8ui]

Let's suppose ISS suffers a major failure where it's attitude control is broken. Things go wrong in space quite often, it's not too far fetched. Nobody's going to be writing software to handle that - best handled by a human acting directly than by a human writing software that will maybe get actually used once. If software was to handle that, it'd be an edge case where software is going to be far far less reliable than it normally is. As of right now, if something goes majorly wrong and ISS ends up literally tumbling, humans can dock to it, and software can't.

Bad example! The ISS attitude control is made of a dual-gimbal control moment gyroscope system linked, via software, to the reaction control system (vernier thrusters, lots of them). It's got four CMGs and needs just three to autonomously operate, two and a star tracker (it's got a few of those too). The attitude control jets can be fired independently and are normally purely under software control. The software can compensate for the loss of three quarters of all vernier thrusters and ordinarily the thrusters cannot be manually fired anyway.

Software won't make mistakes in a high-stress environment, can't be distracted and can switch to an alternate plan instantly on equipment failure without any head scratching.

Someone has written the software for that, it's been tested extensively and if it all fails and the ISS goes tumbling, you need the quick reactions of software to dock with it, because humans are crap at it.

Written software for what, docking to a dead space station fully automatic? Citation needed. My understanding is that Russian docking relies on antennas on the station for alignment, so if the station is dead, manual docking is the only option. Also all of that software is old as hell and I very much doubt it can handle a particularly wide range of abnormal conditions (such as station's angular velocity too far above one revolution per an hour and a half) .

I'm not even sure other docking adaptors can work with a dead station or can handle big imprecisions in alignment. Dragon altogether got dragged with the arm.

Also as for reaction times, docking is very slow. Tumbling is a very very slow tumble (else it's going to be coming apart anyway).

Ultimately it is not a part of any plan that ISS would go unmanned, to the point where things break and it ends up tumbling. It's part of random things that sometimes happens to space stations. Like it wasn't a part of any contingency plan that fuel cells would blow up on Apollo 13. Or that Salut 7's battery charging system would fail. Things like this happen sometimes, then you have to make the best of it, and you don't even have a pre-existing plan let alone an incredibly formalized plan that is software.

 

[Dmytry]

ZnU[/url]":3ng9uedr]

Dmytry[/url]":3ng9uedr]
Written software for what, docking to a dead space station fully automatic? Citation needed. My understanding is that Russian docking relies on antennas on the station for alignment, so if the station is dead, manual docking is the only option. Also all of that software is old as hell and I very much doubt it can handle a particularly wide range of abnormal conditions (such as station's angular velocity too far above one revolution per an hour and a half).

It's worth noting that the difficult part of a fully general automated solution to docking is likely building software that interprets and integrates sensor data from multiple sources into an accurate understanding of the spacecraft's surroundings. Doing this with respect to spacecraft docking is a simpler version of the same problem that companies working toward self-driving cars and driver assistance features — including the one that Elon Musk owns and another that owns a chunk of SpaceX — are investing a lot of resources in solving.

It also got a far, far smaller use case than self driving cars. I don't think it's even worthwhile writing a fully automatic software that can handle rare edge cases such as docking to dead space stations (which happens out of lack of money in the first place, so you won't have money for software to handle that case, either). And his capsule gets grabbed with the arm anyway.

 

[Dmytry]

Hopefully it's not the idiot emdrive that he's talking about. These people are big believers in esoterics / pseudophysics.

 

[Dmytry]

There are two obvious things that could be created/manufactured on Mars that Earth will pay for. One is entertainment. The low gravity combined with the inherent drama and newness means there will be a lot of movies, documentaries, reality tv, video bloggers, sports, etc. How much money that will bring in I don't know, but but it will be something.

By space standards any sort of revenue from that is utter peanuts. For the first few years it would maybe just about pay for shipping cameras and extra communication gear necessary for that, as well as for editing. That's about it. Hell have you watched TV recently? Discovery channel: building tree houses. (Not tree houses of that amazon tribe that lives in the trees, no). A good nature documentary a year, about things which are going away forever within next 50 years.

Look at NASA stuff on youtube, look at view counts for the day to day stuff. It really is quite depressing.

edit: Do you remember the names astronauts on Apollo missions other than the first? Here's your answer.

Another example, nobody yet even coughed up the money to even film porn in orbit, not even for the first space porn ever and it being something that'd be genuinely very interesting in addition to being porn (I think they did a bit in suborbital flight once). If you can't fund orbital station with porn (not even 1 delivery for the first ever) you sure as hell can't fund mars base with reality TV.

 

[Dmytry]

Hat Monster[/url]":1i7jp73w]

Dmytry[/url]":1i7jp73w]
Another example, nobody yet even coughed up the money to even film porn in orbit, not even for the first space porn ever and it being something that'd be genuinely very interesting in addition to being porn (I think they did a bit in suborbital flight once). If you can't fund orbital station with porn (not even 1 delivery for the first ever) you sure as hell can't fund mars base with reality TV.

What does orbit add to porn that zero gravity doesn't? Yes, parabolic arc aircraft have been chartered to make porn.

A parabolic flight gives you like 30 seconds of "zero gravity" (bobbling around hitting walls).

Seriously, with the reality tv and such you probably won't even beat Earth-borne "big brother" a few months in. The public interest in something like this vanes very rapidly.

Look at the view counts here. Compare to view counts for, I dunno, reviews of some space flight videogame, or trailers for The Martian, or what ever. I don't know WTF is going on but the public interest level in real space exploration is utterly depressing. And yes I'm aware that there are other ways to view said videos - nonetheless it's clear it's orders of magnitude below gangam style and even gangam style won't fund a mars base.

edit: realistically speaking we're talking of a worldwide hit the first landing on Mars, maybe worldwide hit important discovery of something (life?), followed by public becoming rapidly bored of looking at that place. Long term large viewer base requires entertainment value, and having to do filming on Mars would only decrease entertainment value compared to Earth-bound CGI. That's the best case, worst case there's some sort of disaster with loss of crew and space exploration takes a huge funding hit.

If there will be space economy that's paying for itself, that'll first be in actual space rather than on Mars, mining things out of an asteroid or manufacturing something in zero-g.

 

[Dmytry]

That's the point, the day-to-day matters are dull. NASA's alka-seltzer in water has a lot of views too, and first step on the Moon was pretty popular. But you can't stretch that into some long term export for the Mars colony.

And against some earth-based production, a year after the base's established it's like how many people would pay to read raw spy leaks vs how many would pay to watch a James Bond movie. In the former case you have a few people who value it a lot, and many who don't value it pretty much at all, whereas in the latter you have many people who would pay 20$.

If anything a single video-sending probe into Jupiter atmosphere (sending to a bunch of orbiters or maybe floating recording then sending it back over long time or something of that kind) would be far more valuable for entertainment. We have no references of what it looks like for good CGI. And I bet it looks awesome.

edit: I think the potential entertainment value could actually be higher for orbiting space habitat, because it's in zero-g and shows something that is very hard to replicate well on Earth. Of course, we've had space stations for a long time now, so it's not very new any more. But that novelty won't last for actual Mars base.

edit: or here's another example, how does James Cameron fund his deep sea exploration? With CGI stuff like Avatar, not so much with deep sea documentaries. The main value for entertainment is in inspiration.

 

[Dmytry]

nj_kruse[/url]":1qd3nizb]

Dmytry[/url]":1qd3nizb]If anything a single video-sending probe into Jupiter atmosphere (sending to a bunch of orbiters or maybe floating recording then sending it back over long time or something of that kind) would be far more valuable for entertainment. We have no references of what it looks like for good CGI. And I bet it looks awesome.

Is there a low-radiation zone if you get into a low enough orbit around Jupiter, like there is around Earth?

I think so, for equatorial orbits at least. I used to work on software for rendering clouds and other atmospheric phenomena (this), I always wanted to make a Jovian storm and I'd love to have a reference to work off.

 

[Dmytry]

.劉煒[/url]":1tua8xll]Mars/colonies. The new Sili Valley/etc.

There's gonna be a heck of a brain drain.

Sili Valley is not on the Amundsen–Scott base or something, it's around where you had the actual gold.

 

[Dmytry]

Yeah, exactly, it's not a backup if it can't even survive an economic recession on Earth let alone anything major happening.

From what I gather Musk is concerned about skynet like total apocalypse scenarios. When it comes to nuclear war, Earth is still left far more habitable than Mars can ever be, ditto for any asteroid that is not absurdly huge, and the nearby supernovas and the like would wipe out a Mars colony at lower levels than where they would wipe out people on Earth.

With the skynet nonsense I'm thinking it's 1: quite severely ridiculous at the current tech level, and 2: if it actually happened as described by various AI doomsayers, it would very rapidly take over the entire solar system, probably the neighbouring stellar systems, etc.

The gap between something being not a threat and it out-thinking the entirety of mankind is really huge, the gap between out-thinking entirety of mankind and going to Mars is, well there's no gap.

edit:
IMHO, nothing about this AI mythology makes any sense. Their idea of how AI would end up massively superhuman is "self improvement", which would require that the AI is very massively superhuman to start with. Their survival plans are downright idiotic in one way or the other. Go to Mars, hide in the basement, pay some shady wannabe AI programmer to verbalize his version of laws of robotics. Some even outright talk about attacking Intel (paramilitary style) if it's building brain emulations (never mind that a full blown, working brain emulation is still basically an infant but with energy cost for which you can hire dozens if not hundreds very smart adults, never mind that we can't get a simulated C. Elegans with all of it's 302 neurons to work right).

 

[Dmytry]

RobDickinson[/url]":3ehaf7w3]Sure we have to start somewhere, just questioning if Mars is the best place to do so. We already have a permanent base in space with the ISS though that will never be self sufficient

Yeah, exactly. Orbital base actually has a multitude of plausible paths to economic self sufficiency at least (which would be the first step) - industrial processes requiring microgravity, maybe processing asteroid material (brought there by some solar powered tugs), etc. Mars base, not so much. It's like, I dunno, building up more land at the shore vs building cities at the bottom of the sea at the depth of 5km. Hmm, I wonder what's the cost of making huge steel spheres and sinking them comparing to putting ISS modules up there.

 

[Dmytry]

Megalodon[/url]":usm5enoz]

Dmytry[/url]":usm5enoz]if it actually happened as described by various AI doomsayers, it would very rapidly take over the entire solar system, probably the neighbouring stellar systems, etc.

Musk has said as much. Are you sure you're not arguing with a caricature?

Hmm, you may be right. Still, what sort of Earth borne catastrophe it is that a Mars base is a viable backup for? Keep in mind that it got to become less hospitable on Earth than on Mars (and considerably so) for a Mars base to be a stronger backup than some hardened location on Earth, and I really can't think of anything short of the extreme AI doomsday scenarios that would make it happen (and the latter for a very brief time). edit: is he believing in some unintelligent gray goo by chance?

 

[Dmytry]

RobDickinson[/url]":3610rp30]Probably the biggest real threat is something big hitting earth. Rouge planet or something.

A decent sized asteroid would screw over earth enough to make it tricky to live on but as said more so than mars? doubtful.

I think nuke war is a considerably bigger threat (and due to survivor bias we have really shit one point data on which to base it's probability), but there's absolutely no way it'd make Earth worse than Mars.

 

[Dmytry]

Tom the Melaniephile[/url]":1rx8ld06]

Dmytry[/url]":1rx8ld06]

Megalodon[/url]":1rx8ld06]

Dmytry[/url]":1rx8ld06]if it actually happened as described by various AI doomsayers, it would very rapidly take over the entire solar system, probably the neighbouring stellar systems, etc.

Musk has said as much. Are you sure you're not arguing with a caricature?

Hmm, you may be right. Still, what sort of Earth borne catastrophe it is that a Mars base is a viable backup for? Keep in mind that it got to become less hospitable on Earth than on Mars (and considerably so) for a Mars base to be a stronger backup than some hardened location on Earth, and I really can't think of anything short of the extreme AI doomsday scenarios that would make it happen (and the latter for a very brief time). edit: is he believing in some unintelligent gray goo by chance?

Mars is obviously practice for adapting to/terraforming planets in other solar systems. Lots of space for figuring out stuff we don't know that we don't know yet if we get started on Mars.

However, 99.99% will have a hard enough time wrapping their heads around a self-sufficient Mars colony without discussing the next stage.

Well, the argument was about it being itself a backup. And once again, why Mars? There's a lot of material that is not down a gravity well. I recall reading some scifi that I think got it quite right - there's space economy revolving around asteroids and industry using solar power, and Mars just has an intermittent scientific base, less than Antarctica does today. Which seems to make perfect sense to me.

edit: and as for other stellar systems I kind of think that by the time we can get there the whole point of terraforming will be moot ten times over, with the practical importance comparable to that of building a zoo.

 

[Dmytry]

Megalodon[/url]":qhvkxd8h]
WRT the backup value, I think it largely comes down to the Fermi Paradox and apparent Great Filter. We can't see any short term existential threats but there should be other intelligent life in the cosmos. One explanation for this is that intelligent life just isn't very common, but the Copernican Principle argues we're not special, so if we're not special just by existing in the first place maybe we're headed for some existential threat that we can't see coming and we will also not be special when we're wiped out.

It seems to me you're applying Copernican Principle to the space dimensions but not the time dimensions, thus arguing that given that we are special (as observed), we must be special in time. I never seen a coherent argument why Copernican Principle shouldn't apply equally to the time dimension, I always see everyone implicitly assuming that it shouldn't but I don't see anyone explicitly articulating why.

Various concerns about us not being special really shouldn't be able to lend more credence to one hypothesis why we're observably rare than to another.

I prefer to think of it in terms of volumetric fraction of the universe that is life (at the given moment) - concentration of life, a dimensionless quantity which as we observe is very very small.

The thing is, even a star, any star at all, is very special : most of space is quite far from any star - the concentration of "a shell at habitable distance from a star" in the universe is quite low. The concentration of blobs of liquid water within the above (which never all completely froze or boiled off in billions years) is also pretty damn low.

The next step, you need to have a soup of random sequences of something that replicates and self catalyses in complex enough way that it can evolve (there has to be a road of small improvements from there). And within that soup, a specific string of, say, 11 bases is at best 4x less common than a specific string 10 bases. You see where I'm going with this. The concentration of that first molecule or molecule complex that starts off evolvable replication can easily be astronomically low if it needs even a slightly long sequence that has to be right - it falls off exponentially with required minimum length.

Then there's also the fact that our lineage from first evolving replicator to us is very long, involves great many steps, and probabilities at a few steps being moderately low would multiply up into a very small number.

Basically for the concentration I see something like 10^-20*10^-25*a*b*c... (purely illustrative numbers) and it seems very silly to me to be surprised if any of the a,b,c,... numbers would be as small as ones we know. (Drake's equation just skips the factors we know).

Another thing Musk has talked about with the AI doomsday scenarios is that if it were AIs doing this, then they'd be the ones to spread out into the cosmos so we should still see them, so even if they wipe their biological creators out it still doesn't answer the question.

Yeah, we either see them or don't see ourselves that late in the history of the universe (if they come here).

So your argument that there's no apparent existential threats where Earth isn't its own best backup is undermined by your inability to convincingly argue you understand all the existential threats.

Lack of knowledge cuts both ways. You have to make dubious assumptions about existential threats to argue that spending money on Mars colony is better than spending them on Earthy matters (or anything else). Namely you have to assume that the bulk of existential threat is very far in our future (when the colony becomes self sufficient), rather than some known present day threats.

That is AFAICT in Musk's view is unjustified. Note, at this point please resist the urge to make the "I don't agree so that must not be what Musk thinks" argument, since he talks about it whenever anyone will listen. You don't think it has much backup value, but the robustness of a species' original home wouldn't be unique to us so more diversity is warranted, particularly if you think you can do it for a fractional percent of GDP and accomplish other things worth doing while you're at it.

Lastly, all this Mars activity would be an impetuous to improve our space technology, meaning it would be the first step in spreading out to the rest of the solar system and hopefully eventually beyond. Dismissing Mars because it doesn't accomplish everything you might want from a backup in one step seems like missing the point. Once we can build a self sustaining civilization there we can go anywhere. I can see the argument for going to the asteroids first, but AFAICT a $500k ticket to Mars implies asteroids are pretty easily accessible too so it's not mutually exclusive.

With regards to Mars versus asteroids, it's not as if we're going to Mars (or asteroids) with even a good chance of establishing a permanent base. It is entirely possible to make the whole enterprise go substantially kaput by diverting funds instead of reinvesting them into growth. Seriously there's this talk as if there's no chance whatsoever SpaceX would ever need the cash not to go bankrupt. Some sort of post scarcity mentality being applied to a very resource constrained situation.

 

[Dmytry]

shread[/url]":2uf7glj2]

Statistical[/url]":2uf7glj2]

Yeah, exactly, it's not a backup if it can't even survive an economic recession on Earth let alone anything major happening.

I don't think anyone is saying start a colony and day 1 it is now a backup. It will takes decades for a colony to become self sufficient but you aren't going to get to that point without starting. On a long enough timeline all species becomes interplanetary or they go extinct.

This whole backup meme is an _excuse_ for going to Mars. It's utilitarian window dressing that you add on after you've decided to go because it's there. Nobody here who says they'll go is worried about developing a backup for humans; they just want to go. It's hard for people to let go of utilitarian ideas. It's always easier to sell an idea if it has strong utility and to defend the idea from idiot utilitarians. Going to Mars is just super cool, outstanding, a good thing to do.

Well, yes, that's fair enough, but what I don't like is that it validates the idiot utilitarian viewpoint by playing along, which is very bad for other cool missions (into the atmosphere of Jupiter for example). Or even the high utility things like asteroid stuff.

BTW with regards to it being a 'backup' it seems to me that the number one catastrophic risk is that we have a nuclear war, which would not kill all of us, but it would put us into the situation where civilization has substantially collapsed and there's a lot of ecological niches not requiring human level intelligence, not to mention that easy fuels are exhausted.

edit: it seems pretty weird to me that we're treating this going to the star as a null hypothesis and something which takes a catastrophe to prevent. Plenty of earlier civilizations just crumbled for one reason or the other, and they didn't exhaust worldwide resources. We could just get completely totalitarian world dictatorship and eventually revert to some Mayan style human sacrifices, for all you know. Us going to the stars is just one scenario of many.

Hell, some aliens with reasonably well functioning world government, if they looked at Earth... they'd be seeing the world of Hunger Games except with more inequality and shittier tech.

 

[Dmytry]

Megalodon[/url]":2c5u0ifc]

Dmytry[/url]":2c5u0ifc]It seems to me you're applying Copernican Principle to the space dimensions but not the time dimensions, thus arguing that given that we are special (as observed), we must be special in time. I never seen a coherent argument why Copernican Principle shouldn't apply equally to the time dimension, I always see everyone implicitly assuming that it shouldn't but I don't see anyone explicitly articulating why.

I don't think you've properly understood the argument then, as you're making a case for applying the Copernican principle in time when in fact everyone already is

I don't think you understood what I mean. We both agree that intelligent life has low spatio-temporal concentration (we don't see any).

I think that we can't just conclude whenever if it's due to low concentration on spatial dimensions (due to being far apart) or low concentration on the temporal dimension (due to being short lived).

You somehow invoked the Copernicus principle to privilege the notion of low temporal concentration over the notion of low spatial concentration.

What I am inclined to believe is that the first evolving replicator relies on molecules or molecule complexes which - due to the number of basic building blocks arranged in a specific way - may have very low equilibrium concentration in the primordial soup. I mean, for example if you get RNA sequences arranged at random, a specific 100-base sequence will occur at a concentration of 4^-100 or less.

edit: I.e. what I mean is that you can have absurdly, unimaginably low concentration of specific complex molecules at chemical equilibrium. If the first step to life requires such molecules, Copernicus principle is irrelevant - those molecules could have formed on other planets, and did form thorough the universe, it's just that the distance between them is huge (comparatively to interstellar distances) because their concentration is low. It's trivial that at some point in the emergence of life, we're relying on atoms arranging themselves in a special pattern by mere chance. We don't know how long this special pattern is, but we know that the concentration of such patterns will decrease exponentially with their length (so even a fairly modest length can correspond to ultra low concentrations). We have a perfectly plausible mechanism by which intelligent life can be very rare.

It does not seem plausible to me that very different species of intelligent aliens would very reliably end themselves rather than spread through their galaxy in a relatively short time.

And if that's the case, it's something we wouldn't even be able to change even by completely altering ourselves into a more peaceful or more intelligent or less intelligent (yet more numerous) species, or species that are intrinsically far more cautious. edit: or for that matter, stellar systems that have a second habitable planet.

The civilizations can be made of species less intelligent than we are (if they have more dexterous hands and work better together), and species more intelligent (with less dexterous hands or another handicap that keeps them from starting civilization until they're smarter), species with greater variation in intelligence and other traits or lesser variation, etc.

All profound differences well outside the reach of our ability to steer our course. If the galaxy is teeming with non intelligent life and Great Filter lies ahead of us, we're almost doomed - our only course of action is to try to guess in which direction we may be extreme comparing to other intelligent life and go in that direction further.

Not true in light of the above, as you've already pointed out humanity is unlikely to be wiped out by most of the foreseeable threats

What's with all the wiping out? It can just stagnate itself and regress for cultural reasons. Or, the other life is getting wiped out (or evolution stagnated) within that huge multi billion year timespan it takes to go intelligent (a timespan when stellar parameters undergo considerable change), and we're just the first in the neighbourhood that made it there.

edit: Also, if you're imagining an universe where a lot of intelligent life is springing up and then killing itself, then you're getting intelligent species whose neighbour planet further from the sun is heavier (or whose "venus" is lighter) than what we got, making said planets actually far more useful than asteroids or even habitable.

 

[Dmytry]

Technarch[/url]":n0252os7]

jbode[/url]":n0252os7]\

I'm just not convinced that any of that will translate into a permanent Martian colony, in spite of Elon's stated goals.

I think that a permanent Martian colony could still be made to happen, but it will only be made to happen in spite of the economics. It's a project that pretty much requires a semi-crazed billionaire, and it's not clear that even Elon has enough cash to pull it off. I'm rooting for him, but still.

Yeah, that's my take on this as well. I'm all for space exploration and science. Right now it's great that Musk wants to go to Mars because he's creating all that useful tech. But further down the road we're talking either 1: all his backup talk is revealed to be just talking things up (not improbable IMO), or 2: he's dead serious and starts diverting massive amounts of money towards permanent mars base, which is just bad for his business and would hurt other ventures. Especially when it is compared to focussing that same zeal on asteroids (which could at some point pay for themselves).

 

[Dmytry]

Statistical[/url]":39h4vqm9]

*mass extinctions, >75% of species extinct in each event.

You want to compare extinction rates vs Mars?

That isn't the point. RAID doesn't reduce the probability of a hard drive crashing but it does increase the chance of your data surviving the loss of either drive.

What is the probability of a simultaneous multi-planet extinction event? Probably not zero but significantly lower than a single planet one. On a long enough timeline we spread out among the planets (and eventually the stars) or we go extinct.

The probability of a natural Earth-only extinction event within next 100 000 years is very small to start with, and best addressed by working to protect Earth from said events.

The real problem here is self inflicted extinction events. It's like you're juggling your laptop while standing on an unicycle, on a gasoline spill, while smoking (periodically lighting up another cig). The natural MTBF of the drives is probably not a very big concern then (assuming they're not SSDs for the sake of analogy). Maybe juggling a backup drive as well (simultaneously with juggling the laptop) would help, maybe not.

Seriously, what's with all that vague "existential threat"? We have one specific threat: war. Everything else is rare in comparison.

 

[Dmytry]

Syonyk[/url]":qeeqkpea]

Seriously, what's with all that vague "existential threat"? We have one specific threat: war. Everything else is rare in comparison.

How about ongoing loss of biodiversity, increasingly non-renewable-resource intensive farming for diminishing returns, and the certainty that, at some point, somewhere, if we keep on the current path, we're going to trip over one of the really nasty positive feedback loops with regards to greenhouse gasses (if we haven't already)? There are an awful lot of those identified, and if we keep doing what we're doing, we're bound to hit at least one of them, which will probably set off others.

We keep doing what we're doing, and we don't have a planet that's capable of supporting it's population in 2-3 generations. Having an outpost on Mars changes this exactly none, and it's not long enough out that we're likely to terraform Mars by then. That's a multi-century project, best case.

Yeah, you're right, that's another way we could fail. It's just that it seems less like a threat and more like "it may be the case that we are extreme idiots and our civilizations are too idiotic to have a long lifespan" kind of thing. I also don't think it would make humans go extinct edit: if it's peaceful that is (which it probably won't be, i.e. war again).

The sad truth is that if we can't make it on Earth for self inflicted ecological reasons, it's us being orders of magnitude short of being able to make it anywhere else. Earth would still be far more habitable than Mars after any ecological upset we can do.

edit: with aliens the conditions could be very different easily - aliens with less dexterous hands or aliens (just for an example) living in the atmosphere of say 5% oxygen at 4 bars would have a far higher threshold on their intelligence before they start building things or using fire, so they'll presumably have to be smarter before the civilization emerges. Us we had it rather easy, which is good in some ways and bad in other ways. I think we're actually pretty damn stupid, we lucked out on non-intelligence factors. edit2: it also really sucks that our intelligence has so much variance.

edit3: also the reason for this mass extinction is that we humans are wildly successful. When non intelligent species do this their descendants proceed to fill up the freed ecological niches. If we apply that to humans... picture a planet where anything bigger than a mouse is a hominid. I think there was a scifi book where there was an ecology like this... Larry Niven's ringworld?

 

[Dmytry]

Syonyk[/url]":qfiqvzsm]

Yeah, seems we're mostly in a complete agreement, just a few points though:

I don't think humans will go extinct, but I think a very convincing case can be made that the current fully sustainable carrying capacity of the planet is radically, radically smaller than ~9B people or so.

I don't think it's below 9B people with some sane policies. It's below 9B people when you have extremely unequal distribution of things though. The bottom half of the world's population lives with very very little. If bottom 3.5 billions of people die, that'll only change the resource consumption by a few percent.

Really I have a feeling that just 100 million people, maybe even less, can damage the Earth nearly as much as 10 billions would. Those 100 million people would only have to be moderately wealthier than the top 100 million people are right now.

I certainly think there's value in trying to put humans on Mars, if nothing else for the effect such things tend to have on people's attitudes and global cooperation. But I don't think a "backup population" on Mars is likely any time soon, and unless we make some major changes to our current civilization and approach to energy, nature, water, and the like, we're not going to be around long enough to have a snowball's chance in hell of establishing a self sustaining colony on Mars or anywhere else.

There's a lot we can do, but the bulk of the useful stuff seems to involve going back to closed loop agriculture (which is, by necessity, more labor intensive), and a radical reduction in energy use. Which doesn't always require a radical reduction in lifestyle comfort, though probably a radical reduction in "stuff" - the consumer focused economy is a big problem.

There's also a lot of interesting work that can be done with low energy/low water ways of growing food sustainably, and I intend to work on this quite a bit as soon as I move and have some land to play with - aquaponics seems promising, especially if ground insulated and able to produce most of the year.

None of the stuff we have to do is particularly affected one way or another by cheap launch capacity, though.

Exactly. I think a lot of this approach has to do with the west coast transhumanist subculture. It strikes me as escapist in a sense. The more extreme guys literally preach that we must ignore all real threats to focus on the AI.