The Age of Entanglement

[black_son_of_gray]

The juxtaposition of these two recent articles is rather interesting (particularly from an Ingenuity Gap / Lights Out perspective), and it seems like this forums has a mix of people along the "techno-utopia" <-> "impending-societal-collapse" spectrum.

The Age of Entanglement

“When the world we have created is too complicated for our humble human brains, the nightmare scenario is not Skynet—the self-aware network declaring war on humanity—but messy systems so convoluted that nearly any glitch you can think of (and many you can’t) can and will happen.”

The Bandwidth Bottleneck

Internet companies are painfully aware that today's network is far from ready for the much-promised future of mobile high-definition video, autonomous vehicles, remote surgery, telepresence and interactive 3D virtual-reality gaming.

I wonder at what point technological possibility intersects with practical implementation, and where that fits in with current mainstream infatuations (e.g. virtual reality, autonomous vehicles, smart devices). For example, it may be possible in a lab to achieve the absurdly high amounts of continuous data collection and processing to coordinate self-driving cars on the road, but what about in the messy real world, where tunnels or weather might subtly degrade communications and ruins the whole thing, or some cars are operating different versions of proprietary software that glitch out when dealing with each other (quality debugging takes on a different dimension when gridlock that shuts down a city or people's lives are at stake). What if virtual reality becomes very easy to implement on an individual scale but just costs a HUGE amount of computing overhead/bandwidth and isn't really economically feasible for all the people on a network that want it?

Behind all of this is the fact that these technological improvements are generally taking place on a backbone of assorted, cobbled together code, frameworks, and infrastructure that it was never designed for. AKA "lock in"

The brittle character of mature computer programs can cause digital designs to get frozen into place by a process known as lock-in. This happens when many software programs are designed to work with an existing one. The process of significantly changing software in a situation in which a lot of other software is dependent on it is the hardest thing to do. So it almost never happens.

Where do all of you sit on the technological optimism spectrum? Are these issues just minor hurdles on the way to spaceships and unitards, or do you think humanity will hit a technological plateau (vs a continuing exponential curve) due to humanities own mental capacities? Any thoughts?

[sky]

Devolution -> Idiocracy

[George the original one]

"Virtual Reality" over the internet is not too tough. In racing and flying simulators, it's been happening for nearly two decades already. The only difference now is the local graphics capability has improved. The trick for success has been to package the data transmissions as small as possible, mostly by preloading the environment and then transmitting only positional information rather than trying to transmit the whole environment.

Technology is applied science. As long as we have not exhausted the combinations of science and applications, technology will march forward. It is far more likely that we'll not have enough resources to accomplish what we can envision. So I'm a pessimistic optimist because I know the resources we can extract are finite.

[jennypenny]

So, you know I'll have a doomer response

Look at the infrastructure of past technologies. How is that holding up? Do we collectively care whether it's maintained, or do we move on to the next new thing and abandon the last? How many of us drive over crumbling roads--roads we are still using and dependent on--and never give it a thought? Look at the state of the power grid, which is arguably the most critical piece of infrastructure in the US. One-third of it is comprised of equipment we don't even manufacture anymore. As a society, we're not good at maintenance. We like new and shiny.

IMO, it doesn't matter anyway. As GTOO said, we don't have the resources to fund ever-more-demanding technologies. We are blindly producing power-consuming technologies, even though it's pretty clear that we're going to hit peak power. In the future, we could be debating whether it's practical to let the masses use power for things like washing machines, or if it's better to conserve most power for technological development or certain industries (like medicine). It's easy to envision a future where small industrial complexes are lit up like Christmas trees while people in the surrounding communities are hand-washing their clothes.

[black_son_of_gray]

Technology is applied science. As long as we have not exhausted the combinations of science and applications, technology will march forward. It is far more likely that we'll not have enough resources to accomplish what we can envision. So I'm a pessimistic optimist because I know the resources we can extract are finite.

Agreed. Resources meaning both raw (e.g. semiconductor material, energy production) and operational (e.g. bandwidth), right? I wonder if the massive expansion of internet-connected devices will ever lead to something like a rolling brownout - because there is a tragedy of the network commons. Companies sell the device, but ongoing drain on public network capacity isn't priced in (is it? I honestly don't know). At some point, if a device transfers huge amounts of data continuously, won't the operational costs of connecting a device over its lifetime cost more than the device itself just in energy to maintain the servers? I imagine high-tech things would shift to being subscription items at that point, with pricing reflecting the rate of energy use. Any computer wizards want to chime in on this?

In the future, we could be debating whether it's practical to let the masses use power for things like washing machines, or if it's better to conserve most power for technological development or certain industries (like medicine). It's easy to envision a future where small industrial complexes are lit up like Christmas trees while people in the surrounding communities are hand-washing their clothes.

I doubt society will be able to prioritize or control power usage in any meaningful way. The washing machine comment reminds me of the current self-driving car optimism. Don't get me wrong, I think it is a neat accomplishment, I just don't see the point in relieving the human in the car of the overwhelmingly easy task of paying attention to your surroundings and occasionally swing your arm a little and adjusting the angle of your foot. Is that the innovation that society would be best of dreaming of right now? A slightly more convenient and efficient usage of a mode of transportation that is ridiculously wasteful of energy? Meanwhile Zika has broken out in Florida, is that a priority?
Interesting point about the patchiness of high-tech distribution. I guess a modern example would be many African communities, where houses may not have running water, but most people have cell phones. I suppose if wealth/income equality in rich countries continues to grow for a few more decades, we could end up with similar weirdness in the US - dilapidated suburban neighborhoods where everyone has a room with "parlor wall"-style screens.

[jennypenny]

Not to get OT, but I love the concept of driverless cars. It's public transit combined with American sensibilities. Every time I see my almost blind and over-medicated 79yo FIL drive off, I think he (and other seniors like him) are the perfect candidates for them. Add in drunk drivers and how many accidents could be avoided just from those two groups alone?

[jacob]

Mainstream infatuations are mostly some extrapolation of the current hot technology. What was the current hot technology in the 1960s? Spaceships! So this is how the future of 2001was perceived. Didn't happen. Go read some Jules Verne for earlier examples. Now, what's the current hot tech of the 2010s? Software engineering and robots ... therefore the future must consist of AI and robotics.

Hmmm ....

In terms of entanglement, [when making predictions about the system of the world] I make a distinct point of seeing whether the complex system is the supporting foundation of something (bad) or whether it lies on top of something (no big deal). For example, the power grid lies below. The airline booking system lies above. If you want to go into politics (ARGH), the public sentiment lies below. Politicians lie above.

The other thing to consider is the strength of the systems couplings. If the coupling is weak, cascades don't happen. E.g. when New Orleans was flooded, it didn't cause failures in Seattle. If the couplings are strong, e.g. housing market to stock market, the system can cascade via a central hub. Another effect is something that ties to every single node simultaneously, such as a Carrington event (chance of ~0.1%ish/year).

TL;DR - I'm not too worried about the lack of bandwidth. The amount of bandwidth being wasted on cute puppies and Game of Thrones means there's a lot of slack. It's similar to the food system this way. That is insofar you have an income over $2000/year. For those below, it's very serious and a leading cause of global/anti-globalization insurrection (commonly known as terr'ism).

[Sclass]

I used to be in conflict about the increasing design complexity in embedded computing systems.

While designing my last product we had to decide whether to home brew part of the system or use an off the shelf operating system for a target processor and its associated libraries. We ended up using the off the shelf since coding up modules for industrial networking and driving display interfaces was just a waste of time. The system went together fast. Our intellectual content was layered deep beneath a mountain of calls to the generic stuff we bought.

On occasion tech support and engineering would come to me and say we had a unit on the South Pole that would quit once a month and need rebooting. This was very inconvenient for the operators. Tech support would be sick of the engineering's inability to fix it and they'd come to me. As if I could trace the bug. I dug into the hardware and code and pulled back in horror as I uncovered layer upon layer of crowd sourced code. I kind of let our young SW guy use this stuff because it made development go faster but at the end of the day, he created something we couldn't understand and couldn't fix.

One little glitch a month. This was a mission critical application and it just wasn't going to work. After personally digging into the kid's code with an in circuit emulator and single instruction debugger I found a divide by zero that happened when a counter overflowed at the same instant a divide occurred via an asynchronous call to a mathematical function. If the divide happened after the overflow there would be no problem other than a ripple on our filtered signal. Under some rare circumstances a crash would result. When I grilled the kid about the lines of code he said it wasn't his fault because somebody else wrote the code and that he couldn't test every possible condition. This drama would play out three times a year. Our customers found the bugs since we never tested our devices under their conditions.

Around the same time we had a huge argument about the cause of the Toyota Unintended Acceleration problem. Remember how they never found an cause? A number of factory recalls may have "flashed" away old versions of firmware. NTSA may be too dumb to know how to get a real answer. I just never could get myself to buy the official story.

So do we go back to stone axes that we understand? No way. Do we go back to excessively long development times while we roll our own? No, we'd die a quick death by bankruptcy.

I just hope that the guys who design things like traffic signals or electronic trading systems are being careful about their ingredients.

[Toska2]

"D. Exler, chief executive of Mercedes-Benz USA, believes that driverless cars are likely to be 'bullied' by human drivers when the two are sharing road space.

Speaking at a motoring conference in LA, Mr Exler told delegates that it human failings are the main factor slowing the development of self-driving cars - more so than technology, insurance considerations or getting even people to accept the very concept of self-driving cars.

“The real issue is humans,” he said.

While driverless cars will be programmed to be law-obiding and considerate road-users, their human counterparts will be much more aggressive."


Skynet will soon know humans are the bottleneck.

I think we will plateau. For an increasing majority, the work necessary to improve the human condition will be "caretaking" of the system and the inhabitants. Think more nurses, psychiatrists, IT people, skilled trades and paper pushers and less genetic researchers or material scientists.

While technology is a very powerful force it's being resisted by human nature. Uber is still a product of lowish urban density. Humans psychologically think "home and a garden" & "buying a house to store wealth". There's parts that we can change about ourselves (aka ERE or materialism) , I don't think voluntary change will happen soon enough to allow technology overcome environmental hurdles.