That’s a great headline, but it seems like a heck of a lot of speculation. I am in no way connected to the intelligence community, but if I had such a tool, I would use it to identify targets for surveillance. Surveillance doesn’t mean “drone strike.” Of course, that’s just a bunch of speculation on my part.
Exactly. It’s not like they take the outcomes of the system and say, “the computer says this guy is a terrorist, kill him!”
They would just use that system to focus human surveillance.
The article looks like speculation, but you guys should read Rule 34, by Charles Stross.
I also consider this the goal of technological progress. It would mean a much more socialized economy to make it work though (or at least a dramatic reassessment of how to value human production, which I think it’s even harder).
But as others have said, probably each step in that direction you get a harsh readjustment until priorities are rearranged with the productive reality, like in the previous industrial revolution (which solved the unemployment problem via limiting the franchise to work, namely taking kids and elderly out of the equation and limiting working hours, so the average work per working citizen dropped to about a half -and thus a potential 50% rate of unemployment was taken care of-. The myth that new jobs took the place of the old jobs is that, just a myth).
the comments actually go into a lot more detail (as would a general internet search around the subject to help you dig up details to not draw the (weak) ‘speculation’ tag):
This really could have gone in the Evil™ thread, obviously, but it was a trend in the science thread in terms of AI and where it is going/what it is doing etc. Just because you may not like the tone of an article, or it’s subject, does not make it speculation. The Skynet program is not speculation, nor are innocent deaths due to our drone strikes, these are all real.
I could do a general internet search and come up with results telling me that Obama is actually a lizard man. What I posit is that none of us, including earnest article commentators, know the actual process with which this software was used. It could be used for evil, but it is not required or even particularly helpful for evil to be done so why would it?
Nobody is saying that the program doesn’t exist, that drone strikes don’t exist, or that innocent deaths don’t exist. Get past that - it’s not part of the discussion.
I could not resist. Behold, 30 seconds of Google yields:
True. I think i find it just base pretty disgusting that our military’s are working in area’s where this might be a reality, and if not actually now, certainly look likely for the near future. It’s a dark path to walk imho. Anyway the guardian, that bastion of liberal thought and news takes issue with the article also, and provides a bunch of detail:
‘Has a rampaging AI algorithm really killed thousands in Pakistan?’:
A killer AI has gone on a rampage through Pakistan, slaughtering perhaps thousands of people. At least that’s the impression you’d get if you read this report from Ars Technica (based on NSA documents leaked by The Intercept), which claims that a machine learning algorithm guiding U.S. drones – unfortunately named ‘SKYNET’ – could have wrongly targeted numerous innocent civilians.
Let’s start with the facts. For the last decade or so, the United States has used unmanned drones to attack militants in Pakistan. The number of kills is unknown, but estimates start at over a thousand and range up to maybe four thousand. A key problem for the intelligence services is finding the right people to kill, since the militants are mixed in with the general population and not just sitting in camp together waiting to be bombed.
One thing they have is data, which apparently includes metadata from 55 million mobile phone users in Pakistan. For each user they could see which cell towers were pinged, how they moved, who they called, who called them, how long they spent on calls, when phones were switched off, and any of several dozen other statistics. That opened up a possible route for machine learning, neatly summarised on slide 2 of this deck. If we know that some of these 55 million people are couriers, can an algorithm find patterns in their behaviour and spot others who act in a similar way?
What exactly is a ‘courier’ anyway? This is important to understanding some of the errors that The Intercept and Ars Technica made. Courier isn’t a synonym for ‘terrorist’ as such - it means a specific kind of agent. Terrorist groups are justifiably nervous about using digital communications, and so a lot of messages are still delivered by hand, by couriers. Bin Laden made extensive use of couriers to pass information around, and it was through one of them – Abu Ahmed al-Kuwaiti (an alias) - that he was eventually found.
That’s who the AI was being trained to detect – not the bin Ladens but the al-Kuwaitis. Not the targets so much as the people who might lead agents to them. Ars Technica implies that somehow the output of this courier detection method was used directly to “generate the final kill list” for drone strikes, but there’s zero evidence I can see that this was ever the case, and it would make almost no sense given what the algorithm was actually looking for - you don’t blow up your leads.
Still, we should all keep in mind this question: ‘Would God approve of me using AI to kill other humans? How would that look on my rap sheet when i’m at the Pearly Gates?’
I think the load-bearing part of this question is “kill other humans”, not “using AI”. At least at this point, the “AI” in question is strictly a tool.
Well yes, especially in the Christian context, killing other humans is expressly forbidden by God, still just in case folk see using AI to do the actual killing as ‘wriggle room’ to get by those Pearly Gates, well you’d be wrong. So keep that in mind NSA computer AI researchers (and people like that).
So let’s get back to other science stuff (as this current trend is what it is):
‘Is D-Wave’s quantum processor really 10⁸ times faster than a normal computer?’
We have been following D-Wave’s claims about its quantum hardware at Ars for a number of years. Over that time, my impression has oscillated between skepticism, strong skepticism, and mild enthusiasm.
Back in November, D-Wave issued a press release that basically asked tech journalists to spray paint a finish line just behind their feet and break out a victory flag. It seemed a bit much. But now that D-Wave has declared victory, perhaps it’s time to re-examine the skepticism. What exactly has D-Wave achieved, and does it constitute victory? Either way, where are the company’s efforts focused now?
Of course the best way to judge D-Wave is not by its press releases nor by the specifications and benchmarks glued on the boxes of its processors—these should be treated with utmost paranoid suspicion. Instead, it’s better to look at what researchers who have access to D-Wave hardware are claiming in publications. And despite my suspicions, the paper accompanying that last press release—plus a couple of other papers on the arXiV that were released earlier—is interesting. All together, they paint a picture that says we should finally be cautiously optimistic about D-Wave’s progress.
If you are unfamiliar with what D-Wave does and its history, you could easily read several other articles before continuing. The TLDR is as follows: a long time ago, in a country far far away (Canada), a little start-up company announced a 16-bit quantum computer. This surprised everyone, considering that a lot of researchers were previously very proud of achieving one or two bits. To muddy the waters even further, the alleged quantum computer used an entirely different approach from everyone else’s attempts, one called adiabatic quantum computing.
In adiabatic quantum computing, one does not directly perform operations on individual bits or groups of bits. This is unlike circuit quantum computers, where there are single operations such as a CNOT (controlled not, the fundamental logic operation in quantum computing). Instead, the solution to a problem here is re-configured so that it is the ground state of an energy landscape.
Think of it like this: in an energy landscape shaped like a bowl, a particle can sit at the bottom of the bowl, it can be sloshing back and forth up the sides of the bowl, or it can be anywhere in between. The ground state is one that involves the particle sitting at the bottom of the bowl. For a bowl, this is easy to figure out. But for an arbitrary landscape with multiple particles, the ground state is not easy to determine in advance. So even though we know that our desired solution is the ground state of some energy landscape, we cannot conveniently calculate what that is. Therefore, we still cannot efficiently find a solution.
This is where things get clever for D-Wave. Instead of starting at the desired landscape, the company starts with the bowl and puts all the particles in the ground state of the bowl. Next, it slowly and carefully deforms the bowl to the more complicated landscape we care about (this is called an adiabatic process, hence the name adiabatic quantum computer). If it’s done carefully, the particles stay in the ground state—and at the end of the transformation, we have the solution.
Afterward, to get the answer, we simply read out the state of all the particles. Job done.
Just weird as heck…and cool and interesting too :)
‘New lithium battery ditches solvents, reaches supercapacitor rates’:
Cool article, Zak. Thanks.
Scientists announced Thursday that they have built a single-celled organism that has just 473 genes — likely close to the minimum number of genes necessary to sustain its life. The development, they say, could eventually lead to new manufacturing methods.
If a gene could be removed without disrupting the cell’s ability to live, grow and reproduce, it was deemed nonessential. Out it went. They chucked out a bunch of genes involved in transporting and metabolizing different kinds of food, since the cell they were working with had plenty of sugars to keep it going. They kept in almost all of the genes responsible for reading genetic material and creating more. They also kept a few that would allow the cell to reproduce quickly enough for them to observe that growth in the lab. And they added some specific DNA that could serve as a “watermark,” indicating the lab that had assembled it — the J. Craig Venter Institute.
Based on estimates that biologists have been making for decades, Venter’s team thought the modified bacterium would only need about 250 genes to eke out a bare existence. But, says Venter, it ended up requiring a lot more genes than that.
…minimal cells of various varieties could serve as starting points for engineering cells that digest or produce certain chemicals. “It’s a basic component that we can add things to,” says Venter, like a gene from a deep-sea creature that would allow a cell to eat carbon dioxide and spit out methane — a fuel.
“You could design cells and choose the type of metabolism you want,” says Venter. “If you just have a cassette of those genes that you can just plug in, that will enable design to go much faster,” he says.
Cool stuff, with lots of potential to be used in some really interesting bio-engineering fields of the future. Although I suppose you could also put it over in the What Could Possibly Go Wrong thread, since it could probably also be the basis for some kind of gray goo (or would that be green goo?) scenario.
‘Sweet drug clears cholesterol, reverses heart disease—and was found by parents’:
Pretty cool what a little motivation can discover :)
Great article, unfortunate title. Lots of inventors/researchers are parents :)
With modest theatrics, he pulled open his bookshelf to reveal a secret lab behind it. The first room contained the nuclear reactor.
In the adjoining room there was a well-equipped laboratory. Carl gave us radiation meters and gave us a lecture on radiation safety. He had instrumentation all around the reactor to detect any radiation leakage and assured us that there had never been any radiation outside of the one-ton shielding that housed the reactor. Then, it was time to turn on the nuclear reactor.
Carl, an IT manager at Microsoft by day, had a vision of a private science club to teach students “real science.” He turned his vision into a Friday night program that attracts the brightest minds in the region. Adult volunteers, who are experts in biology, electrical engineering and software engineering also attend the Friday night meetings.
The result is a long list of wins and accolades at the highest level of science competitions. In 2015, this little club that meets in Federal Way won 14 first place trophies at the Washington State Science and Engineering Fair, hauling in over $250,000 in scholarships. They attended the International Science and Engineering Fair and were awarded fourth place among 72 countries.
There’s probably a bit of “what could possibly go wrong” here, but it’s more than outweighed by the “how awesome is that” aspect.
I wonder what sort of permits & inspections are required for that. Their web site makes no mention of regulations. Science fair organizations are usually reasonably good at requiring documentation, at least for things they’ve encountered before.
Mr Burns is probably sponsoring them.
In another thread the fact of India being the powerhouse of mathematics invention was mentioned and these yotube videos cover some of that:
It is a shame the full BBC series History of Indian Mathematics is not available for us all to enjoy, it was rather good (as BBC series like this often are).
Not related to India, but fun and informative, Terry Jones (the former Python guy that liked dressing up as old women) did this documentary on maths a few years back:
Some dudes found a cool mechanism of spider silk and created a synthetic version.