SoylentNews is people
SoylentNews
Software is making it easier than ever to travel through space, but autonomous technologies could backfire if every glitch and error isn’t removed.
When SpaceX’s Crew Dragon took NASA astronauts to the ISS near the end of May, the launch brought back a familiar sight. For the first time since the space shuttle was retired, American rockets were launching from American soil to take Americans into space.
Inside the vehicle, however, things couldn’t have looked more different. Gone was the sprawling dashboard of lights and switches and knobs that once dominated the space shuttle’s interior. All of it was replaced with a futuristic console of multiple large touch screens that cycle through a variety of displays. Behind those screens, the vehicle is run by software that’s designed to get into space and navigate to the space station completely autonomously.
[...] But over-relying on software and autonomous systems in spaceflight creates new opportunities for problems to arise. That’s especially a concern for many of the space industry’s new contenders, who aren’t necessarily used to the kind of aggressive and comprehensive testing needed to weed out problems in software and are still trying to strike a good balance between automation and manual control.
Nowadays, a few errors in over one million lines of code could spell the difference between mission success and mission failure. We saw that late last year, when Boeing’s Starliner capsule (the other vehicle NASA is counting on to send American astronauts into space) failed to make it to the ISS because of a glitch in its internal timer.
[...] There’s no consensus on how much further the human role in spaceflight will—or should—shrink. Uitenbroek thinks trying to develop software that can account for every possible contingency is simply impractical, especially when you have deadlines to make.
Chang Díaz disagrees, saying the world is shifting “to a point where eventually the human is going to be taken out of the equation.”
Which approach wins out may depend on the level of success achieved by the different parties sending people into space. NASA has no intention of taking humans out of the equation, but if commercial companies find they have an easier time minimising the human pilot’s role and letting the AI take charge, than[sic] touch screens and pilot-less flight to the ISS are only a taste of what’s to come.
Which approach, do you think, is the best way to go forward ??
(Score: 0) by Anonymous Coward on Friday July 03 2020, @10:44PM (1 child)
These spacecraft are just as safe as Elon's autonomous cars. Just hope there are no space police cars for them to smash into at full speed.
(Score: 0) by Anonymous Coward on Friday July 03 2020, @11:03PM
Trust Boeing to deliver the crispy astronauts [npr.org] we crave.
(Score: 4, Insightful) by Barenflimski on Friday July 03 2020, @10:50PM (3 children)
I'm the kind of guy that still likes to turn the knob when I change the volume. I can't tell you how many times I wanted the volume at exactly 7 1/8th of a turn, but the system locked me down to only being able to go to 7 or 7.5. -grin-
In all seriousness, at some point one would think that in long haul vehicles it would be necessary to have the ability to manually control the ship. Being able to turn a wrench, release a valve, or twist open a hatch release seems like things that might come in handy once in awhile as you gently float in a tin can through the solar system at 50,000m/s.
Let's get real. Even the Borg were shining lasers at things and working with materials and they were so superior they would assimilate you. Take head humans!
(Score: 2) by c0lo on Friday July 03 2020, @11:15PM
I reckon adjusting the music volume will be left for you. And that only, because the speed of human reaction is limited to 150-200ms.
https://www.youtube.com/watch?v=aoFiw2jMy-0 https://soylentnews.org/~MichaelDavidCrawford
(Score: 3, Informative) by Snotnose on Friday July 03 2020, @11:25PM
I'm that same kinda guy. Except my 15 year old car has decided the steering wheel volume control now changes radio stations/cd tracks, the change stations/tracks does nothing, my sunroof sorta works, and I'm about the sell the car even though the mechanical s are great.
Computers are great for maybe 10 years, after that stuff just breaks down. You can't really compare the lifetime of an electrolytic capacitor vs a crankshaft/camshaft/piston,piston rod/etc.
When the dust settled America realized it was saved by a porn star.
(Score: 0) by Anonymous Coward on Saturday July 04 2020, @02:54AM
I mean, you jest, but tuning to radio stations by hand I get clear signals, and digital tuners give me crap.
(Score: 1) by fustakrakich on Friday July 03 2020, @10:53PM (5 children)
There should be only two buttons, "Up" and "Down". For the impatient, we can put the "Close Door" button also.
For troubleshooting, under the monitor is a reset button, and as always, the power cord in case all else fails.
La politica e i criminali sono la stessa cosa..
(Score: 2, Insightful) by fustakrakich on Friday July 03 2020, @11:04PM
In all seriousness, all normal operations should be totally automated. Just wait for the green light to come on. Humans should only step in if there is a malfunction, and then manual repair and control should be possible. The ship will always need an engineer. There isn't much to debate over this
La politica e i criminali sono la stessa cosa..
(Score: 2) by driverless on Saturday July 04 2020, @01:05AM (3 children)
Read the book "Digital Apollo", which goes into this in great detail. It was recognised even half a century ago, with the relatively primitive control systems available at the time, that spacecraft control was beyond the capabilities of humans. The Soviets put this into practice and automated as much as possible. The US, who had to deal with macho test pilots who were used to running the show, tried to provide the illusion of pilot/astronaut control while automating as much as they could behind the scenes, in a battle that stretched for years.
Humans are necessary for handling exception conditions, but everyday control needs to be handled by capable automated systems.
(Score: 2) by mhajicek on Saturday July 04 2020, @06:40AM (2 children)
Good luck handling one of those exception conditions when there's no option for it on the touch screen.
The spacelike surfaces of time foliations can have a cusp at the surface of discontinuity. - P. Hajicek
(Score: 1) by khallow on Saturday July 04 2020, @02:48PM (1 child)
(Score: 1, Touché) by Anonymous Coward on Saturday July 04 2020, @04:41PM
"Set SCE to Aux"
(Score: 4, Insightful) by takyon on Friday July 03 2020, @10:56PM (2 children)
A spacecraft failure shouldn't set back the study of a solar system object by 5, 10, 20, 50 years. We should be sending out hundreds or thousands of autonomous spacecraft every year. Lowering launch costs will help make that happen.
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 2) by driverless on Saturday July 04 2020, @01:13AM (1 child)
... to bring ever more star systems under the ruthless control of our one.
(Score: 1) by khallow on Saturday July 04 2020, @02:54PM
Certainly bringing the Solar System under our control is a good start for a lot more than just a stellar empire.
(Score: 5, Insightful) by Runaway1956 on Friday July 03 2020, @10:57PM
Maybe. I kinda sorta expect spacecraft to do a lot of crap for the "drivers". You shouldn't be expected to do a space walk to align some thruster just like you want it.
On the other hand - if you decide that you really DO WANT to make that spacewalk to tweak the thrusters, the software and crap shouldn't get in your way.
Face it: Here on earth, if your car stops working, there are probably hundreds of phone numbers you can call, to get things fixed. Someone can be at your location within half an hour, almost anywhere. They can bring anything from a new tire, to a new car, if you have the funds to pay. Up there? If shit stops working, a rescue mission might be a year away, or more.
If you're going where no man has gone before, you had better be resourceful enough to outsmart, bypass, and even eliminate faulty software, and hardware too!!
"Ground Control to Major Tom, you're not authorized to rescue yourself!! The BMW/JohnDeere/GMC/PublisherGuild consortium won't allow it!"
(Score: 4, Insightful) by MostCynical on Friday July 03 2020, @11:00PM
docking as a service [soylentnews.org]
subscription should be paid before departure, to ensure you can dock when you arrive... proof of purchase/receipt may be requested on arrival.
"I guess once you start doubting, there's no end to it." -Batou, Ghost in the Shell: Stand Alone Complex
(Score: 2) by Lagg on Friday July 03 2020, @11:13PM (4 children)
i.e. Over-abstraction and cruft is making it hard to unit test all possible code paths.
One thing you can give the human brain despite itself and its apparent bandwidth: We're good at patching in realtime when one of our branches creates undefined behavior and generally don't segfault so hard it brings down everything we were in the middle of doing in the heap space.
Not sure how fixable a problem like this is besides with forced minimalism though. Or we could just ban thinking machines and create mentats. Or remove the human risk entirely, stop caring about space junk and "move fast and break things" until some iteration makes it far enough out to send back useful data. That seems to be the overall opinion of the guy in the quote. Still not sure how I feel about it. It's definitely the logical extension of the production-from-HEAD philosophy
Nice job outlining the whole article whoever submitted this btw, i didn't notice my quote above was already in your outline until just now.
http://lagg.me [lagg.me] 🗿
(Score: 2) by takyon on Friday July 03 2020, @11:40PM (1 child)
AI will fix the problem. And probably write the code too.
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 2) by Lagg on Saturday July 04 2020, @12:10AM
I don't even know if you're serious or not because this industry has its own hype-industry. But your other post was pretty grounded to what actually happens in this kind of beta testing.
Granted, non-determinism like that would definitely bring me over to the side of AI being deservedly called that instead of stuff I still consider to be neural nets and FSM on some level. Just multiple layers of them passing increasingly structured data back and forth after doing their own deterministic processing on it.
http://lagg.me [lagg.me] 🗿
(Score: 0) by Anonymous Coward on Saturday July 04 2020, @09:31AM (1 child)
I find it hard to believe you can't have both extensive automation and also the ability to selectively turn things off and have a large degree of manual control when required. Different layers of software abstraction are the perfect tool for the job.
Auto pilot of course has flight plan containing all the steps needed to go somewhere, and recalculates whenever something unexpected happens.
Executing those steps will involve actions like 'add X deltaV in direction Y'
Which breaks down into the functions to orient the craft and functions to fire the appropriate thrusters for a specific amount of time.
Which break down again into functions to prepare the fuel pressure and valves and all other stuff required to operate the thrusters.
And so on, and so on, until you get to the functions that flips the actual bit that manipulate the hardware, like closing a circuit relay to provide power to a specific motor to operate a specific valve. Much like the switches in an airline cockpit.
Of course the same goes for other systems like eg. life support or thermal management.
At each of those stages you could decide at any moment to put the high level program into some sort of standby mode where it will not execute any actions (or only certain types of actions), and you could instead manually call the lower level functions yourself, if the software contains the appropriate interface to allow that. There is no reason a pilot couldn't give the command to 'fire main thruster 500ms' or 'close valve A' at any time.
(Score: 2) by Lagg on Saturday July 04 2020, @04:46PM
Good point, but the problem as I see it - especially because this stuff is really dumb hardware oftentimes compared to what we have on desktops - is like how Windows behaved with ISA and all that non-plug-and-play crap. The 737 Max drama (they're putting it back into production this year btw) kind of informed me that these aeronautics projects have something of a problem in that vein with interdependency buildup/coupling, and I've always reckoned that this "AI" starts getting fucky when you give it unexpected input like that of a human's.
Basically the 737 Max features an entire simulation of the yaw/pitch/pilot-words-i-dont-know to make transitioning to the Max easier for a pilot without recertification. The complexity of simulating the in-flight behavior of an entire other plane model seems complex and incredibly reliant on all those tiny bits of system operating concurrently. Or at least this is how it was explained to me. So I can't imagine the sheer levels of interdependency required to make that stage of "halt execution here for a second" work. Because if the pilot gives input of any kind, you have to also give the code a "step back to this pointer" behavior of some kind so that it can retrain based on what the pilot just did.
I'm not making any claims of deeply understanding how this "AI" works because it's such a hype hellhole at this point, but it seems like behavior inherent to FSMs and NNs are making people like those quoted be like "well shit this is a hard problem to solve, let's remove the human risk altogether"
http://lagg.me [lagg.me] 🗿
(Score: 4, Interesting) by sjames on Friday July 03 2020, @11:49PM
As I recall, Mercury astronauts complained that the whole flight was either automated or controlled remotely and petitioned to have more ability to manually operate the capsule. They started referring to themselves as "Spam in the can" to emphasize their point.
There is also a general trend for things to have less manual controls at hand as they mature. In the Model T, the driver could manually adjust the fuel mixture, choke, spark advance, and if the ignition ran from the battery or magneto. (The driver also HAD to adjust those). The newest cars don't even give you direct control over the starter motor.
(Score: 2) by Mojibake Tengu on Friday July 03 2020, @11:59PM
Real value of any technology usually shows itself when the war lasts longer than the average lifetime of used components is.
In next war, devices and people with their lives dependent on networks and networked services will die first.
Those who survive will adapt properly, sooner or later.
But unlike possible enemies in a war, the Space grants no mercy and never forgives.
There is a reason, why there are flight engineers in aircrafts, if only for the purpose of diagnostics the unexpected situation properly.
Cheaply not having one in a space ship is a crime.
Respect Authorities. Know your social status. Woke responsibly.
(Score: 2, Insightful) by crm114 on Saturday July 04 2020, @12:07AM (1 child)
This just feels like the old story from the subject. A human being is an incredible machine.
AI is just a new word for same old stuff. Binary based computers make binary decisions.
My first thought is Apollo 13 ... when "hey... something completely outside of the parameters just happened."
Human wetware has this incredible ability to adapt, respond, and resolve.
(Score: 0) by Anonymous Coward on Saturday July 04 2020, @11:19AM
https://chinesenewyear.net/zodiac/rat/ [chinesenewyear.net]
(Score: 0) by Anonymous Coward on Saturday July 04 2020, @01:34AM (1 child)
AI and automation are not necessarily the same thing. Did they really mean AI?
(Score: 2) by PiMuNu on Saturday July 04 2020, @07:03AM
> AI and automation
Well, it turns out AI is a very flimsy word that doesn't really mean anything. For example, (military) rockets have been guided by a sort of adaptive algorithm known as a Kalman filter for 5 decades. This is "machine learning" in some sense - algorithm is essentially
while True
assess path to target
update flight controls
(Sorry my html is poor, probably the formatting will get screwed up)
with some linear algebra thrown in. This *is* machine learning and so can be described by some as "AI". AI is a flimsy word that means nothing. AI is not what you think!
(Score: 5, Insightful) by PinkyGigglebrain on Saturday July 04 2020, @05:17AM
until someone dies as a result of the automation failing because the programmers not accounting for something that could have been fixed by flipping the right switch or worse, the control system failing due to a power surge or other hardware failure.
What would have happened to Apollo 12 after it was hit by lightning if Astronaut Alan Bean had not been able to flip SCE to AUX [wikipedia.org]?
The crew of Apollo 13 probably would have died if not for their ability to manually control every part of the electrical systems of the craft.
Some automation is fine, but when it gets right down to it on a space craft you want to have a backup that allows complete manual control over the craft.
"Beware those who would deny you Knowledge, For in their hearts they dream themselves your Master."
(Score: 0) by Anonymous Coward on Saturday July 04 2020, @02:11PM (1 child)
automation is great ... until it isn't.
not sure this is relevant but netflix is super easy to use. i assume lots is automated.
i use netflix via "rub-off-the-number" from a card i paid cash to get.
this works really easy: go netflix domain, find "redeem code" and enter the number (add some captcha) and you're in ... or not.
now i don't know (blackbox) the back-end of netflix and how it does the code redeeming thing but when it didn't work for me ("sorry, try again later") all manner of nasty thoughts went thru my mind; afterall it worked before and everything else, including youtube and torrents were working fine on my computer.
ready to give up and accept the cash as gone "down the drain of non working prepaid cards" i did one final test: i connected to neighbours wifi and tried one last time to have the code number go thru. lo and behold, it worked.
the difference in the network setup was that in mine, all dns lookups went thru dnsmasq which asked tor (torrc:dnsport) whilst my neighbour had the cloudflare or google dns servers configured.
i think it kindda shows how automation can make things intended by automation to be very simple BE very simple but if you try to do something via the automated system for which it is not automated then expect "a red light and an apology "s
i'm sorry dave" like.
in conclusion (?) maybe, automation is good but do not skimp on precise and informatif error messages just because the system is "simple to use"?
(Score: 2) by takyon on Saturday July 04 2020, @02:36PM
There is a long history of using robotic probes in space exploration.
The difference is that they can now make decisions based on available data instead of waiting for instructions, which can save a lot of time under bandwidth constraints and light speed communication delays.
https://www.jpl.nasa.gov/news/news.php?release=2013-259 [nasa.gov]
https://www.nasa.gov/feature/jpl/nasas-mars-2020-rover-completes-its-first-drive [nasa.gov]
https://www.machinedesign.com/mechanical-motion-systems/article/21836761/helicopter-exploration-coming-to-mars-in-2020 [machinedesign.com]
If you look at the Pluto flyby, you can see that some of the photos taken were blanks or off-center, because the pre-programmed routine didn't perfectly match the sizes and locations of Pluto and its satellites. Could a degree of autonomy have helped to get more useful data? I'm not sure, but you only get one chance with a flyby.
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 2) by ilsa on Saturday July 04 2020, @02:45PM (10 children)
The whole notion of "Move Fast and Break Things" is one of the most categorically stupid notions to come out of the computer industry in a long time, if not ever. It's a pathetic rational for dismissing quality engineering, or even just putting 5 minutes of thought into whatever code you're trying to write.
And the result has been the single biggest and worst leap in software quality we've seen in decades.
The fact that people's lives now routinely depend on this growing library of poorly written software scares me to no end. Boeing has repeatdly demonstrated what happens when you half-ass software development. The fact that they are still in business, and that the executive staff haven't been jailed for malicious incompetence, is a travesty.
The only way through this is to raise the caliber of software development across the board. That means using higher quality developers, with higher quality management, writing software in higher quality languages that force the developer to write better code. I am specifically thinking of languages that, by design, allow code to be proven correct mathematically.
LOL Yeah right. Like that's going to happen when crap like Javascript and MongoDB increasingly dominate the computer industry.
(Score: 2, Insightful) by khallow on Saturday July 04 2020, @02:59PM (9 children)
The space industry seems much more suited to the concept actually. Especially on the manned side, a lot of quality engineering has come from this. Reality keeps space engineering more honest because nobody will die if your website is crap.
(Score: 2) by ilsa on Friday July 10 2020, @05:56PM (8 children)
Not yet, but the way software is becoming more and more SAAS oriented, that can easily change.
Also, Boeing has done a fantastic job of demonstrating to the world that people dying is an acceptable risk when cutting corners in your engineering.
(Score: 1) by khallow on Saturday July 11 2020, @06:34AM (7 children)
"Move Fast and Break Things" != "Cutting Corners".
(Score: 2) by ilsa on Friday July 17 2020, @07:24PM (6 children)
> "Move Fast and Break Things" == "Cutting Corners".
Corrected your typo. It is a fundamental attitude shift in software development that, at an absolute minimum, implies prioritizing developer velocity above and beyond all other factors including good user experience and reliability or even other developers. It de-emphasizes actually thinking about what you're trying to accomplish and how it relates to other systems. It doesn't matter how badly users and systems down the line are inconvenienced or outright disrupted. Making half-assed, breaking changes is now acceptable and users just have to live with it.
Since this model has become popular, software quality has taken a massive nosedive. Maintenance has skyrocketed because both developers and end users having to continually maintain constantly changing software subcomponents. Either that, or they give up entirely and stop trying to upgrade, even at the cost of security and reliability, because the effort is too great. Everything spanning from Facebook to Windows 10 to Ansible to Boeing, is objectively worse than software that preceded them.
Move Fast and Break Things is the poster child for the rampant irresponsibility that has engulfed the computer industry and all related industries. It has cost us god knows how much money in damage, and even human lives.
(Score: 1) by khallow on Saturday July 18 2020, @11:58AM (2 children)
And my point all along is that you are wrong here. Software development is not rocketry development. The latter can't get away with a shoddy user experience and poor reliability - because people die and things blow up when they try that.
Contrary to your assertions, the people "moving fast and breaking things" in the aerospace world have been making more reliable vehicles than the people who didn't.
(Score: 2) by ilsa on Monday July 20 2020, @03:34PM (1 child)
> because people die and things blow up when they try that.
You mean like the Boeing 737 Max, which _has_ killed people, and Boeing continues to not give two shits about?
(Score: 1) by khallow on Tuesday July 21 2020, @01:34AM
They might not given two shits about that, but they will give plenty of shits about 400 [go.com] undelivered planes.
(Score: 1) by khallow on Sunday July 19 2020, @12:54PM (2 children)
(Score: 2) by ilsa on Monday July 20 2020, @04:38PM (1 child)
> at least has to deliver a product that works well enough that it doesn't break often for the launch tempo they have
And here is where I think _you_ are "wrong". I don't give a rats ass about their launch tempo. I care about having a product that works, and works reliably. Todays software, isn't. Even if some long running software has a bug or aberrant behaviour that lasts for years, as long as it's predictable it can be mitigated against.
But when the software changes at an absurdly rapid pace with zero QA, that all goes out the window. You are literally walking on eggshells, constantly. There is no way to mitigate. There is no defence. You can't trust anything. Even if the software isn't actually broken, your workload has increased exponentially due to constant retooling. You have to retrain people. You have to redesign entire processes from scratch. This is not a sane way to do anything.
You can tell me I'm wrong till you're blue in the face but that won't change reality. How well would math work if tomorrow 1+1 no longer added up to 2? That is literally the way software works now.
Have you ever used Ansible? They will make breaking changes to the fundamental grammar of the system on a regular basis, which may require you to completely rewrite your playbooks. It is so dangerously unstable that people have written entire toolkits just to manage that instability.
Windows 10 Anniversary edition trashed god knows how many computers, and Microsoft continues to destroy computers to this very day with their shitty updates.
My favourite example is Angular, where the idiot that designed the original version didn't like it anymore and completely redesigned it, completely wiping out all the effort god knows how many people spent learning the original version.
I can cite a near endless list of examples of just how god-awful software has become, most of it revolving around endless sloppy changes that makes it feel less like you're using software and more like you're trying to survive another drunken episode with an abusive boyfriend. It's _exhausting_.
Why do you think Apple is doing so damn amazing despite their exorbitant pricing, their obnoxious "do it our way" attitude and their constant missteps with their hardware? It's because Apple's overall UX is overwhelmingly better than anyone else. It's consistent and relatively reliable. Contrast that with Google and their inability to launch _any_ new products at all because nobody trusts them.
(Score: 1) by khallow on Tuesday July 21 2020, @01:45AM
It's not just exhausting, it's completely irrelevant. SpaceX's control software, for example, has no compare in the orbital launch market both in capability and reliability.
(Score: 1) by khallow on Saturday July 04 2020, @03:24PM
The problem is what can you rely on instead? As already noted, humans are too slow and too inaccurate for a lot of the piloting. And that's assuming they're on board in the first place. Much of what is flying in space doesn't have humans on board.
And comprehensive testing doesn't happen, if you're not flying stuff.