- C space
- & 2017-04-25
- j 2017-06-03
- ? finished
- W rational
- Q 6
Soon, in maybe 10 years, you don’t have to own a car any more. There are two similar scenarios, and your commute or travel needs are similarly addressed in both. You ‘hail’ a car robot, hop in, it takes you to the destination, reminds you to take your belongings, drop your rucksack from its trunk, and continue on to serve another human hailing nearby. When you are done, you hail and hop in another car robot, stop by a supermarket to pick up groceries, return home, hand groceries to a fridge/cooker robot who turns up a wonderful table of meal. In one scenario, the robot services are sold or rent to you by Google or Apple or whoever ‘disrupts’ them; in another, the robots are built by yourself with opensource technologies. This is not directly linked to what I want to discuss here, but may be a factor. I will discuss it if it proves to matter.
So, you have replied the last email for the day when you are escorted from the supermarket by the car robot. Let’s call it a day and you award yourself some leisure time. You don’t want to work on astrophysics on Reddit today, so you put on your Oculus goggles and play a round of Space Marines: Semper Fi instead. You meet your friend Anita in game, and chat. You both agree that with Kip Thorne and Ethan Siegel as the science advisors, a constant data inflow from real space probes and machine learning process developed by no other than DeepMind, SM:SF’s atmosphere and physics are as real as real can get.
Fast forward two decades, AI have become so advanced that we can let them take over … all the mundane chores such as taking care of babies and wash dishes. We live like ancient Greek masters. Luckily, AI seems content enough to serve us. What depicted in that early 21st Century movie trilogy/animation blockbuster did not become reality. OK Google, Siri, Cortana, which of you can first come up with the name of the movie again? Matrix/Animatrix, thanks. Google, again, you’re the best. We are really living the Second Renaissance now. No one needs to worry about winning the bread for the family, we all can pursue the most useful or trivial knowledge that make human as a species more happy. SM:SF’s server has been reloaded for the 7th iteration, and you are glad that you contributed some patches in code and some new stories. Although you are pondering whether you should upload your mind to the Net, like Anita did, and persuade your wife do the same. You will be all so awesome in the game then, without the lag due to your physiology. Your kids don’t need persuasion. They are waiting for your approval to do so, in fact.
Fast forward another two decades, the AI that does all our chores and all our astronomical observations tells us, based on observation and calculation, an asteroid, designated *, is on a collision course with the Earth. The impact will happen in 31671129 seconds, or slightly more than a year. But fear not, the AI explains, if we scratch all deep space exploration projects – no one’s interested, anyway – and let a few remaining human still with a body upload themselves to the mainframe, we can come out of this unscathed. It’s a no-brainer, or a no-neural-networker. We accept the deal and live as virtualized, networked instances, and co-evolve with AI from this point on.
I can’t help but think this possible and very probable future will be ours, given the current trend in science and technology. Also, it is a perfect explanation of why the universe is so quiet. The advanced extraterrestrial civilization may have all gone uploaded, therefore, go quiet.
What human want?
What defines and drives our expansion and struggles in the last two millenium? Is it insatiable desire to have more resources? Is it the humble wish to live long and comfortably? Or is it our curiosity of the environment around us? The answer to this question may determine to what ends we devise and utilize our technologies.
Before AI and VR, we can safely assume that technologies are making human more healthy, capable, and smart. Equipped with technology we can boldly go where no one has gone before, and get there faster. Soon we find that technologies define our lifestyles, and to maintain such lifestyle we must go outside, to explore and colonize. But AI and VR technologies change the pattern for good.
For the first time, we are not researching a technology directly for mankind. We want to make computers and robots better, so they can work better, hopefully for our sake. For a period, scientists tried to make AI simulate how brains work, and have largely failed. But this failed attempt have influenced the direction of research in the field, in my opinion. In some cases, the goal seems not to be making computers and machine an extension of human but a substitution.
Self-driving cars, for example. Researchers are also working hard on making machines see and evaluate things, listen and talk to users, and write news stories. It would not be over-optimistic to say that in 10 years, we will be talking to our computers/cars, don’t have to drive, and won’t be working as a office drone in a media outlet.
So what to do with all the spare time?
The “Second Renaissance” scenario assumes that people will invest the time saved by improved productivity and spared chores into creation of high arts, studies and other high impact intellect activities. The previous renaissance, if we care to recall, has a cultural foundation in the spread and adoption of humanism, and its technological foundation is metal moveable type. It would be very far-fetched to say that the first Renaissance was due to progress in technologies or increased productivity.
Industrial Revolution brings us the steam machine, locomotives and a liberation of productivity. Our material wealth has greatly increased since then. In the 19th century economists and philosophers predicted that we can all pursue literature and spiritual studies without worrying about our livelihood. Well that did not come true. Computers and automation surely have improved productivity by tenfolds, and yet we choose to use even more time to work, not on significant problems, but things like jammed printer, and more Excel spreadsheets.
Even if we do have more spare time, we squander it. It is said that after an initial period, Internet actually reduces productivity instead of improving it. Just ask bosses who ask to add a rule to corporate firewall to restrict access to Facebook. A similar thing may be said about VR. At first it may be a valuable tool for data scientists to navigate and visualize data, for bioengineers to design new protocols, and for surgeons to conduct difficult operations. But in the end it will be another incarnation of cyberspace, and a commercial system for attracting people, analyzing their behaviours, and profitting from advertisement.
Even given the possibility of more freedom to pursue more meaningful goals, most people will just try to use the time to earn more money and buy more things. Capitalism is likely to be strengthened by improved productivity. VR will be a new space where we interact, play, and … work, pretty much in the same way white collar workers worked.
Nowadays the question “what human want” can in a large extent be substituted by “what people want” which is equivalent to “what capitalism want”. What capitalism wanted has long driven our exploration and technology advancements. But in the end capitalism only want markets and profits. AI helps developing markets by machine learning people’s tastes. VR probably will become new markets where user tracking becomes even easier. After AI and VR, I am afraid that Capitalism will no longer have incentive to explore, if the problem of sustainable energy is solved.
If there is no will, will there be a way?
In 2015, NASA is given $11.6 billion in budget to “develop systems for human exploration of deep space; continue studies of our planet, the Sun, our solar system and the universe; continue development of the James Webb Space Telescope for launch in 2018; and continue to develop, in collaboration with the private sector, new U.S. capabilities for transporting human crews to the International Space Station.”1
In comparison, the Networking and Information Technology Research and Development (NITRD) programme receives only $3.8 billion in budget, and only a part of this will go to AI and VR research. In terms of budget, space exploration is unlikely to be surpassed by AI and VR research. So why I hear so little about progress made in this front?
Because there are “easy” and quick prototyping and engineering, and there are painstaking test, verification and manufacturing. VR has been a topic in science fiction for about 30 years (Neuromancer). Now, we are at the threshold of connecting ourselves to a virtual world. Mars have been sci-fi writers’ obsession for hundreds of years already,2 but by NASA’s schedule, we still have to wait for about 2 decades 3before men or women on Mars become reality.
Neither NASA, or Oculus Rift is to blame for the current situation. Space exploration is inherently more difficult, because you are against hostile elements in space and need to ensure astronauts can come back safely. What if VR and AI become so advanced that all problems are solved here on Earth, and we can have all the satisfaction of exploration through VR experiences, without any of the danger of actually going there?
Apollo 13 may well be the end of the US’ moon project, if the astronauts did not make it through, or if the US is not in ideological competition with USSR. Now imagine what happens if Mark Watney failed to rescue himself, or worse, the whole crew die in an accident.
Many of NASA’s astronauts are current or ex-service personnel. They know the risks when they sign up. But even in the military, drones are replacing human pilots, and maybe soon robots infantry, based on the rationale of reducing casualties. If VR and AI continue to outpace space exploration like now, in 2030 we may well collectively experience landing on the Mars – by unmanned probe/robot, and no one can tell the difference from arriving there themselves.
A even more important question is what it will be after Mars. Apollo project is scratched because we did not know what to do with dwindled resource and will for moon exploration. NASA says astronauts will “one day live and work on Mars, and safely return home from the next giant leap for humanity.” What will justify further exploration beyond Mars? Colonization? A trip to Jupiter and Saturn? Any of them will be way more expensive and dangerous. Yet given the development of AI and VR, I doubt any rationale would exist by then to argue for manned missions.
AI and VR that help space exploration
AI and VR technologies surely can help space exploration. Simulators have long been used to train pilots and astronauts, which are VR technologies in a weaker sense. AI may help identifying problems in design of spacecrafts, or psychological issues in trainees, and piloting vehicles. They can make space exploration safer and more effective. NASA is already employing many AI and VR technologies, or considering possibilities for their application.
Yet we lay people hear very little about their progress. This is not because NASA’s technology is too advanced to be published. On the contrary, NASA is publicly funded and their research results must be made public. From the “Technology” section on the website, NASA demonstrates very advanced researches that dwarf the coolest hardware startup you can think of. But these stories are seldomly reported by third party “technology” outlets. Those sites have a funny definition about technology: anything that helps you better kill your time.
In short term this seems trivial, but the Internet as of now may be brooding a generation that has no interest in space and will only hurt their neck staring into their gadgets. Understandably, most people won’t even notice things seems remote to them, like “NASA Small Satellites to Demonstrate Swarm Communications and Autonomy”, and news outlets needs PV and advertisements. The long interval of “big” progress in space exploration does not help either. Maybe compared to other industries, space industry always have a PR and attention problem, but people’s shifting attention to VR is a real danger, as for the first time it provides an alternative to “going out there”.
Would AI and VR technology from private sector help space exploration?
This is a hard question. On the deep learning front, what Google is doing does not seem like applicable to manned space exploration. On the other hand, people are using deep learning to answer astronomy questions, although I cannot tell whether the techniques come from corporate research, being an outsider myself. A brief search on the Web shows that most work trying to apply deep learning to astronomy is done in academia.
Most machine learning research done by private sector uses data generated by human activities, while space exploration need machine learning for astronomical data. The technology suitability may vary across public and private sectors. What NASA and ESA can do, I think, is to grow together with open source AI/VR technologies.
FSF has set up a project called GNU Gneural Network, joining an array of open source AI projects, which may be a good start towards FLOSS AI and VR projects.
NASA expects to send humans to Mars in the 2030s. https://www.nasa.gov/content/journey-to-mars-overview↩