Naked Science Forum
Non Life Sciences => Technology => Topic started by: Vince Mills on 13/09/2011 14:45:04
-
Vince Mills asked the Naked Scientists:
My girlfriend and I were having a discussion recently about how automated society is becoming. She posed the question. "If humans were wiped out overnight or very quickly by some kind of virus, how long would the automated manufacturing systems and computer systems continue to operate.?"
Vince Mills
What do you think?
-
The History Channel had a special called "Life After People".
http://en.wikipedia.org/wiki/Life_After_People
http://en.wikipedia.org/wiki/Aftermath:_Population_Zero
Their conclusion was that the power grid would fail very quickly, perhaps within a few hours or days. However, that is likely dependent on the actual power source and the amount of automation.
A coal or oil fired plant would run until its accessible fuel would be exhausted, probably on the order of a few hours or days.
A hydroelectric dam or wind turbine might run on "autopilot" for years.
A nuclear plant would also likely do well for a period of time, but would eventually decrease its power output over a few years due to fuel degradation.
I don't know about safeguards and redundancies built into these systems that would require human input (or likewise would be independent of human input).
Thermostats, clocks, and etc all depend on the power sources. Some things like some smoke detectors have battery backups that could last a decade or so without input from humans.
Certainly within a decade or so most power systems would be degraded and power distribution lines would be lost.
A few local solar systems might still be active, but even their battery backups would be failing.
I don't believe that we have any computers that would be capable of learning and autonomously expanding their role in society, however it is not inconceivable that over the next century or so we will develop some completely autonomous, and intelligent computer systems that potentially could endure in perpetuity, and potentially begin self-replication if allowed to do so.
-
I don't believe that we have any computers that would be capable of learning and autonomously expanding their role in society, however it is not inconceivable that over the next century or so we will develop some completely autonomous, and intelligent computer systems that potentially could endure in perpetuity, and potentially begin self-replication if allowed to do so.
What makes you think they haven't already developed sentience and are just biding their time?
Congrats on the 1K posts
-
What makes you think they haven't already developed sentience and are just biding their time?
Have you used microsoft software? These sentient computers must being very, very subtle :-)
Electronic equipment has a finite life too and so do all the components. There are hosts of potential failure mechanisms with a variety of "activation energies" that determine their life vs temperature. There is a large element of statistics involved, so generally, complex systems will fail first as they are only as good as their weakest link. Even with in-built redundancy, failure will occur eventually. The spread on failures for different equipment is large so it is difficult to generalise, but I would hazard a guess that general computers (PCs etc) would die within a period of weeks, assuming no power failure, mainly because of a random, software-unrecoverable glitch (a hardware soft error that has not been predicted). Specialist equipment and computers, with very good software systems, may last some years with random failures gradually whittling them away (perhaps some would keep going up to 50 years or so - hard to predict).
-
Not sure how it will be in future, but this topic definitely makes me think about how deep we are connected with our personal robots which are totally everywhere in our life.
-
Small isolated systems might run for a very long time.
I don't know if there are any of these installed anywhere.
https://en.wikipedia.org/wiki/Pulser_pump
but, with no moving parts, and driven by rainwater they would run for a long time without intervention.
This is sort of similar,
https://en.wikipedia.org/wiki/Hydraulic_ram
but has moving parts so it would wear out faster.
These things
https://en.wikipedia.org/wiki/Hydraulic_ram
are definitely in use and they are designed to run without any need for attention for years.
If we could design something to run on spam posts we would be doing well.
-
The inverse problem is interesting: if we do some medical magic that saves a life with, say, an implant, can we maintain that implant for the rest of the person's life? This has raised two ethical issues that have crossed my path:
1. The original Charnley hip prostheses were implanted in 60-year-olds who had spent most of their working lives on their feet, so had strong major bones but damaged or diseased cartilage leading to defective joints. The prostheses comfortably outlived the recipients and the procedure was a fairly unqualified success. All sorts of devices appeared thereafter, including adjustable prostheses for growing youngsters, and the general spectrum of patients gradually widened to include those with considerably weaker bones (by 1990, when this came to my notice, very few prospective patients had ever worked in a factory or a farm) and longer life expectancy, so instead of improving over time, the failure rate became a matter of concern.
2. A cochlear implant can give acceptable hearing to a person born profoundly deaf. Early implantation maximises effectiveness and the child can almost always continue as a hearing person. The downside is that in mainstream education he will probably not become fluent in lipreading and signing. No problem except that AFAIK very few specialist electronic manufacturing companies survive more than 20 years and even if the company does, it is extremely difficult to source components after 10 years, so our patient's social and economic life hinges on a device (a) with a much shorter prospective life than his own and (b) which will almost certainly be superseded by a more capable but incompatible model within 5 years.
This almost looks like a definition of life: despite being unbelievably complicated and delicate (given the task of, say, mining coal, no sane engineer would design a human) we live for around 80 years because we can tolerate and self-repair a lot of damage. prospectively adapt our environment to the limited range in which we can function, and work on a wide variety of fuels.
-
With no humans, how long would automated systems run?
This one is aiming at 10,000 years...
https://en.wikipedia.org/wiki/Clock_of_the_Long_Now
-
I'd forgotten all about that machine, but there was a bit of excitement when the first prototype was announced.
The center of the clock will show a star field, indicating both the sidereal day and the precession of the zodiac. Around this will be a display showing the positions of the Sun and the Moon in the sky, as well as the phase and angle of the Moon.
Why? It's an amusing challenge to add such complications to a mechanical timepiece, and they have some value for navigation, but these phenomena should be directly visible from the top of a mountain in Nevada which ain't goin' nowhere!
I'm reminded of my mother and her neighbor watching a solar eclipse. Just as it began, the neighbor said "I'm going indoors - this will be on TV". And indeed of a friend who returned from the holiday of a lifetime with shaky, unfocussed videos of breaching whales and had never actually seen them directly.