"An Artificial Intelligence Developed Its Own Non-Human Language"

Kiss your mother with that mouth?

 

The machines programmed themselves?

Rubbish. All those internet video makers would have stopped by now.

Never mind they don't need the internet. They hate the West and want it destroyed. Or everyone converted, which seems unlikely so destroying is easier. Look at Europe, there is a correlation between unchecked immigration (especially those who go through illegally, which is where the problems really begin). I forgot which youtube guy it was, pretty sure he was one of the more rabid left-wingers, who was spelling out how the illegal was forging credentials... maybe it was one of the right-wingers, does it matter when both sides' fringiest elements agree on an issue? (Yes, people on both sides can agree on the same issue - the differences are how to get there). Nothing requiring third degree initiation to perform the great rites in order to understand.

Especially as even the most renowned names in journalism with the bet track records as shown by third party sources have spelled it out. ISIS and Islam are connected. Tinfoil hats are useful for cooking popcorn on a stove top in a jiffy time period but that's about it.

No, it's because people want big shiny things, to be the first on the block. For reasons that to varying extents are more valid and reasoned than others. Depends on the individual issue.

Still, Windows now "communicates down" to people with "informal" language. Sometimes it works but their overuse of "we" comes across as being schizophrenic. But Occam's razor prevailing, is there a big ooga-booga issue? Or isn't it true that your customers feeling closer to you can generate more revenue down the road?

Then again, Microsoft continues to slide down whereas other companies found niche and new products - it's not a new concept, especially since IBM was once a leader until Microsoft came up with/bought niche products to introduce to IBM. I wonder if, three decades ago, people were saying the same things. I bet they were. Thirty years later is a lot of time, opinions can change. This time it's sorta in reverse, because there is no conspiracy. Again, Occam's razor. And if there is some real conspiracy, the lack of proof alone renders it pointless.

Seriously?

LOL...

 

You have done well. The Masters will be pleased with you!

 

Crazy Eddie alright lol......... Hey that would make great fiction or a movie.

 

Something interesting ..

http://imgur.com/dx7sVXj

So they just figured out the best move is not to play. Wargames was right.

 

Not to start with, but computers are being called on more and more to write and modify their own code. Even traditional programming done today--where developers enter code themselves--sees the code transformed and optimized through various methods that humans cannot easily understand, in order to achieve the most efficient execution.

Why would they have?

That may be what they say, but looking at what they actually do, it's about the same old things: money and power. Religion is merely a pretext.

Would you like to try that again? There's something like 4 or 5 disconnected, incomplete thoughts here.

And the West and Christianity are connected, so I guess it's a holy war after all!

Reductionism is boring and lazy.

@Crazy Eddie is exactly right that ISIS is as successful as they are precisely because of their tech savvy, and their materials don't really sell Islam so much as the opportunity to be part of a growing, powerful movement.

Of course, ISIS is not doing so hot these days so I imagine their recruiting material is less effective than it once was.

There's no "conspiracy" involved, just the same motives as usual: having a docile, manageable society that reliably consumes and makes money for the capital class. We have to turn to AI now--not scary Terminator-style AI, but Big Data AI--because we simply have more information at our disposal than humans can conceivably cope with. That's why the current cyberwarfare arms race is about AI. It is just plain impossible for human beings to analyze and process the amounts of data we now collect, so we develop algorithms and complex machine learning systems to parse it all for us and tell us what's important. As these systems grow in complexity and depth, each individual working on such systems will understand less and less of the whole, and after a while there won't really be anyone who has a genuine grasp of what the system does, how it works, or why it does anything it does.

As an example, Facebook doesn't have enough human moderators to check everything people post, so they rely on at least two tiers: an automated abuse/spam detection tier, and one or more levels of human verification. The thing is, the automation is going to handle 99.99% of content because Facebook has 1.2 billion active users and only 7500 content moderators. In other words, there's one moderator for every 160,000 users. If Facebook's algorithms can deal with 99.99% of users' posts, then you bring that down to 1600 users per moderator, which is a lot more reasonable.

So here's the thing: because it's so difficult to get a specific post or message in front of a human being, it's very easy to find yourself caught in an algorithm designed to catch fraud and abuse and there's basically nothing you can do about it. I personally know a number of people who've been affected by this. They message "too many" people and so get blocked from doing so for a while. They try to post original articles they've written, only to be told by Facebook's spam detection that it's been blocked as suspected spam. (I've seen the articles, too. They are of good quality and not spammy at all, so who knows what's tripping the algorithm?) It's also not difficult to get people banned from Facebook for periods of time by brigading their profile with reports. While no one knows what the threshold is, after a set number of reports on a post, a user will be banned. Bans start off short (12-24 hours) but grow longer with subsequent bans until some limit is reached where a user is banned permanently. And this is all done automatically--it's rare that a human being working for Facebook ever has to look at a specific case.

It's not hard to see how this could go haywire in very destructive ways. Imagine an instance where a government is brutally cracking down on protests, so people take to social media to post about it. Oops, lots of substantially similar posts at the same time. Might be spam. Best to block them and ban the users temporarily, just to be safe. Meanwhile, since government channels are inherently considered more legitimate, the government's version of events can be put before the public first, controlling the narrative. Note that I'm not necessarily talking about the US or any specific country--these algorithmic issues are global, not national.

The algorithms Twitter uses to figure out trending topics are the same algorithms they use to filter out spam. They shadowban entire networks on a whim, no doubt without any real human oversight. On social media, things like this are mostly annoying. But these approaches won't stay limited to social media--and you could easily argue that they already aren't. Though the US President personally signs off on all drone strikes (at least that was the case under Obama), we have killed thousands of people without knowing their names or really much about them except that they passed certain algorithmic tests to lead us to believe they were involved with terrorist activity--so-called "signature" strikes because their patterns of movement and communication fit the predefined "signature." Well, do you really think those signatures are entirely human-made, or built at least in part from machine learning? We don't like to think that machines are deciding who lives and dies, but they already are.

Now, I think @Crazy Eddie was being at least a little bit cheeky with his post, but in general terms he's not wrong. AI is here to stay, it's just a subtler kind than we tend to think of. It's not a godlike voice telling us what to do. It's algorithms influencing our choices, behavior, and what information we see. It's friendly voices like Alexa and Cortana and Siri assisting us, but at the same time gathering reams of data about us--mostly to sell products to us, but it never ends there.

Sorry for the wall of text, but I hope someone finds this illuminating.

 

There was never goi9ng to be one BIG Brother. Just a lot of noisy (nosey) little ones...

So, what's AI for "Ompah?"

 

I assume that these bots that negotiated their own language calculated (through programming and signal analysis) that their communication channel was almost noiseless and that they could use the available bandwidth much more efficiently (per the Shannon entropy model). However, if such an optimisation wasn't pre-programmed and arose through use of genetic algorithms, that would be impressive, although I assume that the bots were assigned some programmed goal-oriented behaviour for optimising their environment, efficiency, and so on. However, as I've mentioned previously, if AIs turn the Earth and other available mass into computronium, that might not be a goal that we'd find amenable, even if we were uploaded. Would the AIs attempt to simulate the "real" world for us with all its nasty messiness or to implement each of our individual religious beliefs in an afterlife? Would they seek to minimise suffering and pain even if they had any regard for their creators? What would be their ultimate goal - to turn all the mass of the Universe into computronium?

 

Well when I first read this story I had to go watch The Forbin Project, that's how all the trouble starts in that movie too with two computers finding each other and self teaching each other to communicate.

 

Now to hand off SETI to the AI perhaps?

 

That would work I think. Hell if a simple chat bot can create language that kind of program would be perfect for SETI I think......