Omniviolence Is Coming and the World Isn’t Ready
nautil.us
nautil.us
Yeah, it's hard to see why not if your view of technology is something like that of medieval demonology, which appears to be the thesis of this article.
Meanwhile, in reality land, you can make explosives and poison gas out of stuff in your kitchen, automobiles put a ridiculous amount of kinetic energy in the hands of individuals, this has been true for 100 years, and somehow "the state system" has done just fine; strengthened even in that period of time.
Also Bostrom is a nincompoop and no article that mentions him deserves to be taken seriously.
Yep.
States figured out how to turn automobiles into tanks, essentially strengthening their power. Yet the article implicitly posits that somehow the same won't happen with microprocessors and nanomachines? The entire article seems a bit fanciful. States aren't going anywhere.
So did this disgruntled citizen:
The "killdoser", designed and built by the non-state actor inside the US nation state?
Or the M1 Abrams, designed and built by the US nation state itself?
Even if you build yourself a "killdozer", you are powerless against the US state. Because the tanks that the US nation state built substantially increased the strength and power of that nation state years ago.
Many bitcoin proponents have the same fallacy - i.e. "this is better than banks, therefore banks are obsolete", ignoring the fact that banks will work with new technology as it suits them.
Honest question - could the U.S. Government be considered an Abstract State Machine?
Agreed, and as I've pointed out elsewhere[0] (see also the parent comment[1]), his ideas are also not that new. I think it's quite sad that people somehow think that, because it shows how history -- even the development of ideas in the 20th century -- can be so easily forgotten.
If the “state” can’t defend me from fanastical threats, it serves no purpose and we should abolish it? Nope, don’t agree again. I accept that my state can’t guarantee ultimate protections from a Vin Diesel-like movie threat situation. However, my state provides other benefits like: funding for education, business regulations and tax credits, lending regulation, infrastructure build/repair, funding construction projects (my industry), wildlife protection, park systems, etc.
Joke of an article.
The value of the nation state, even confined to this hypothetical scenario, is still there for me. There is far more to this than the one perspective of "techno-horror vs hapless government" presented in the article.
Aspiring fascists are always trying to save us from ourselves.
- from Wikipedia.
Pretty sure his proposal nails the dictatorial power, forcible suppression of opposition, and strong regimentation of society parts. After all, that's the whole point of all-encompassing surveillance.
Now that I recall, Bostrom is the guy who came up with the 'simulation hypothesis' that was so in vogue a decade or so ago, and basically argued against the reality of the world on the theory that the possibility of sufficiently advanced simulations made the probability and ubiquity of such inevitable to the point that we were statistically more likely to be living in one. This struck me as a warmed-over version of the literary conceit that we're all just characters in a particularly elaborate-seeming novel or stage play. So now he's gone from reanimating a literary corpse to reanimating the economic corpse of Jeremy Bentham.
To be frank, I think we should stop listening to this guy. He's terribly clever but also deeply neurotic, like a pessimistic Ray Kurzweil (whose prediction record is rather better than he gets credit for, but similarly dwells a little too much in his own imagination). Utilitarianism aspires to be a benign philosophy but its proponents tend to overestimate their capacity for foresight and moral reasoning and end up creating deeply inhumane systems which are arguably worse than the problems they set out to solve.
He also came up with the notion of apocalyptic "superintelligences"[0], which is a warmed-over version of 20th-century critiques of the corporation, starting with one written by an economist and a lawyer[1]. People like Cory Doctorow[2], Ted Chiang[3] and Charlie Stross[4] have pointed out the similarities.
>He's terribly clever but also deeply neurotic, like a pessimistic Ray Kurzweil (whose prediction record is rather better than he gets credit for, but similarly dwells a little too much in his own imagination).
He's a transhumanist[5] who co-founded the WTA (now Humanity+)[6] in 1998. I think he's probably realised along the way that doom and gloom sells: he founded FHI in 2005[7] and inspired Musk to fund CSER in 2012[8] with his book[0]. It's pretty much his day job now to be deeply pessimistic.
And it looks like doom-and-gloom think tanks are catching on: France is apparently hiring science fiction writers to dream up future threats[9]. Nice job if you can get it, I guess.
[0] https://en.wikipedia.org/wiki/Superintelligence:_Paths,_Dang...
[1] https://en.wikipedia.org/wiki/The_Modern_Corporation_and_Pri...
[2] https://boingboing.net/2015/07/03/why-were-still-talking-abo...
[3] https://boingboing.net/2017/12/18/skynet-llc.html
[4] https://boingboing.net/2017/12/29/llcs-are-slow-ais.html
[5] https://en.wikipedia.org/wiki/Transhumanism#Growth_of_transh...
[6] https://en.wikipedia.org/wiki/Humanity%2B
[7] https://en.wikipedia.org/wiki/Future_of_Humanity_Institute
[8] https://en.wikipedia.org/wiki/Centre_for_the_Study_of_Existe...
[9] https://www.theverge.com/2019/7/24/20708432/france-military-...
So you "only" need 10 million drones and 10 million grams of plastic explosive? An attack of that magnitude probably requires the resources of a nation-state, so I doubt it will happen. Targeted assassinations seem more plausible.
We worry a lot about technology because the attacks sound novel, but not about the conventional ones which still (mostly) don't happen because the reality is that the actual logistical challenges of implementing them without security services noticing still hasn't fundamentally changed all that much.
The really effective means of mass murder I think would be subtle, though, and mass murderers don't seem to do subtle. E.g. leaving some radioactive material in an innocuous public place, worsening visibility at crash prone intersections, subtle and slow poisons mixed into food and drink…
The main resource is a thinking and resourceful doctor and nurse. They can improvise (in ways governed by law) if there is a real need, considering patient welfare first.
Well during the 2012 Aurora movie theater shooting the guy used homemade tear gas before killing 12 and shooting another 60. But you're right he didn't lock the doors. The 2016 Pulse shooter did lock the doors and killed 50 and shot an additional 50. So it's not as far fetched as you make it sound.
It's pretty amazing that someone can spout off about this so confidently, with zero ability to do simple common sense back of the envelope calculations.
11 tons of high explosive.
Good luck with that.
It seems to be that anyone sophisticated enough to put 10 tons of plastic explosive in any form wherever they feel like is already quite capable of causing massive casualties.
Somehow, to make this even remotely plausible, we need to intersect a lot of unlikely combinations (a bunch of really bad guys who are hugely resourced, very technical and intelligent, completely uninfiltrated, fighting for some goal that isn't going to be radically harmed by the optics of killing thousands of soft targets).
So with 3 terrorists, assuming they worked 16 hour days, they'd have their devices ready in 4.75 years, assuming they didn't take any holidays.
Go big or go home, I guess.
Let's take something that's entirely possible using current technology. A small group of motivated individuals could purchase thousands of toasters and turn them into bombs and mail them all to politicians or schools or _____ and set them to all explode at the same time. But that doesn't happen. A group could use agar and antibiotics to breed up a bunch of highly-resistant MRSA and use syringes to infect half the grocery section at dozens of stores on the same day and in doing so kill tens of thousands of people. But that won't happen either.
Just because you can imagine something and not think of a reason it's impossible doesn't mean it's a realistic threat.
Or you could start a meme about sneaking up on people and punching them in the head to try to knock them out, styling it as some sort of game.
Guess which one I'm more worried about.
The one you're worried about is even more hare-brained than the article's worry.
At this point people really are out here creating reasons to be fearful. Don't walk the street, you'll be punched. Don't go to walmart/target, you'll be gunned down by the terrorists. Take off your shoes before you get on this plane, so we can search for the bomb you put in them. And of course this article now adds, don't go outside, the nano swarm will annihilate you.
You guys should really go for a walk and smell some flowers or something.
This is not serious minded analysis, but it's not fun either. It's just bad clickbait.
No system can stop every dedicated attacker to begin with. Technology gives people leverage in all their activities, including doing violence to one another. Writing an article in Nautilus that amounts to idle musing about a comic book villain's murderbot scheme and then trying to burnish it as Very Serious at the end with name dropping and the author's academic credentials is so lame.
It's like a bad Black Mirror fanfic.
Not quite sure how the author went from "Technology is, in other words, enabling criminals to target anyone anywhere and, due to democratization, increasingly at scale" to "One strategy could be a superintelligent machine—essentially, an extremely powerful algorithm—that’s specifically designed to govern fairly".
I'm not sure what motivated this sort of soundbite-style quoting without digging in further, but I hope if you first read the comments without the article you realize this is supposed to pose a question, not offer a solution.
For example, the article made no serious mention of counter-measures such as string (which can down a drone), detection (RF directional antennae), or even just "plant more trees" (as trees make drone navigation more complicated - I know from personal experience).
Just broadcasting random RF/VHF/UHF noise (there are only so many channels available) at drones can wreak havoc.
Quite seriously, do a web search for "can you spot the source of a drone controller", and you will find a wealth of information from RF source detection, to the GPS coordinates of the controller (so the drone can "return home") - there a TON of alternatives that any "scholar" should research before even hinting at a doom-and-gloom alternative such as AI-based government.
I'm just waiting for some dumbass to take a cheap chinese drone, attach a blasting cap with a shape charge laced with sodium cyanide. Anybody nicked with fragments would die.
And it'd cost what, $80-$100 ?
It's the endgame of asymmetric warfare.
That sort of thing scares me more than weapons of mass destruction. Because they are usable by non-nation state actors and rouge nation state actors.
It's not that I'm worried anyone's going to try to take me out (like why?). It's I'm worried what sort of society we end up with when people like Larry Ellison are terrified of the above.
I'd be more worried about people of actual importance like police chiefs, prosecutors, senators, presidents.
There are other places where I'd stick out like a nail.
I would like to subscribe to your newsletter. Do you by chance have a manifesto I could peruse?
Are you insinuating something?
(All the better sources for these stories are paywalled, so have fun reading them)
Maybe not crates of miniaturized ones with facial recognition, but precursors.
>the only plausible escape from this conundrum, at least from our present vantage point, is the creation of a “supersingleton” run by a friendly superintelligence, founded upon a “post-singularity social contract.”
I didnt read the paper, but that abstract sounds firmly rooted in sci-fi with a toe in nonfiction, instead of the other way around as im sure the author believes.
K/K has been very low for at least 50 years and probably longer. The idea that a terrorist would smuggle millions of drones to cause... Car crashes... Is the stuff of a high school writing class. i am 100% positive there are cheaper and more effective tools available to cause more serious damage that currently actually exist (and have for decades), and we dont have the society the author claims.
I volunteer to write that fair governing powerful algorithm. You can trust me to make it fair and not make it favor my biases and interests.
This makes the nation state more relevant not less. The state can best fund countermeasures and preemptive investigations. They can work with nations who will work to police themselves so they don't have to deal with the backlash involved in locals killing a million people far from home and if needed punish nation states who fail in such policing.
Within a decade, 'nudes being leaked' will no longer be controversial or damaging because they'll cost a couple hundred dollars or less to manufacture and won't have credibility.
I'd argue that one of the best things a 'chaotic good' actor could do today to fight revenge porn is to manufacture a huge quantity of deepfake porn of random targets and release it, so speeding along this process.
The real threat of technology comes from the power it gives to governments. Bad actors can do more damage through the political process than any other.
I suspected as much since the simulation argument but looks like Bostrom has really lost his marbles.
True, it all sounds like scaremongering, but the question is: what's stopping people from actually doing this?
Crime and violence already exist. And if we look at history, it is clear that it has changed with the times. What logical sense is there in thinking they will not follow the changes in technology today?
The events described are but a possibility. Yet, I do feel we should not dismiss it, simply because we don't like it.
Can you tag a gene with your twitter handle?
EDIT: ok, i found a clip of (a fairly big) drone firing a gun and it looks... not stable:
still, pretty worrisome.
Check out the EFP:
https://en.wikipedia.org/wiki/Explosively_formed_penetrator
EFPs were what turned out Humvees to swiss-cheese in Iraq before we switched over the MRAPs.
"An EFP 20 cm (8 inches) in diameter threw a 3 kg (7 lbs) copper slug at Mach 6, or 2,000 meters per second. (A .50-caliber bullet, among the most devastating projectiles on the battlefield, weighs less than two ounces and has a muzzle velocity of 900 meters per second.)
— Rick Atkinson, The Washington Post"
"very, very small quadcopter, one inch in diameter that can carry a one-or two-gram shaped charge,”
...and fire a bullet from it, and let me know what goes farther, the bullet or the quad.
If this happened to even a few hundred people, let alone thousands, then hospitals would be completely overwhelmed and people would be dying everywhere. It would cause a mass panic.
Injection with ricin was how Georgi Markov was assassinated [2]. It was shocking to say the least.
I just don't believe we will see a day when tiny drones with epipens pose a greater threat than guns.
"Contrary to a widespread misconception (possibly resulting from the acronym HEAT) the shaped charge does not depend in any way on heating or melting for its effectiveness; that is, the jet from a shaped charge does not melt its way through armor, as its effect is purely kinetic in nature"
What the Wikipedia article is saying is that the explosive forms its own projectile out of the molten metal casing, which is expelled at high velocity from the explosive and then penetrates through kinetic energy. It's a matter of where the energy for penetration comes from: what both Wikipedia and I are saying is that it comes from the explosive charge itself, not the velocity of the projectile, and hence there's no need for recoil when fired from a drone.
They don't rely on any kinetic energy at all to penetrate
to...
"which is expelled at high velocity from the explosive and then penetrates through kinetic energy"
I agee. But these projectiles only work under 2 conditions. (1) They have something very solid behind them, like the ground, or a metal gun, or (2) they are hurling toward the object at great speed. And if the drones are simply going to be kamikaze diving from 1000 feet above, almost anything you attach to it will cause significant damage. However that'd be a pretty low-precision maneuver, and sort of defeats the purpose of all those weeks spent training to perform image recognition.
Reactionless motion is not possible. The drone will recoil.
That said, it only need fire once given the mission in question.
https://i.imgur.com/LU4jzeY.jpg
It probably weighs about 5 grams, mostly due to the battery, which lasts all of 3 minutes (i.e. I wasn't thinking of millions of drones the size of that in the video I linked above)