There's a story that scientists built an intelligent computer. The first question they asked it was: "Is there a God?" The computer replies: "There is now." And a bolt of lightning struck the plug so it couldn't be turned off.

Hmmm, all in the name of progress?
The thought of AI terrifies me, I was a luddite long before "today" though.

I think the fear-mongering of AI is a little ridiculous.

The code is barely beyond a stack of if-else statements.

We're in no danger of the machines rising up and taking over the world. The potential for a Skynet is a long, long, long way off.

There is some impressive software and algorithms out there, but once someone attempts to apply it outside of the designated task it was written for, it's usually fairly useless. It's like trying to stick an earthworm in an office somewhere to answer phone calls.

I agree. Although, neural networks are advancing incredibly quickly. I don't think we have anything to "worry" about so long as classical computing is still the dominating type of logic.

Imagine the possibilities of quantum computing though. Boolean logic would essentially be hilariously obsolete. Computing as we know it will be radically changed, and neural networks would well exceed the possibilities of human thought.

I have to say though, I'm not scared so much as I'm excited.

Technology will probably be our downfall but not by AI more like self destruction war, pollution, plaque, depletion ect....hopefully i’ll get my last game in on AFM

Elon Musk - “Maybe there's a five to 10 percent chance of success [of making AI safe],”

“I keep sounding the alarm bell but until people see robots going down the street killing people, they don’t know how to react because it seems so ethereal."

“AI is a fundamental risk to the existence of human civilization”

he also said they will be so fast that you will need a strobe light to see them lol, sounds promising

This is quickly changing. I believe the best go AI out there taught itself to play GO in a day.

Edit: I mean, it became the best go player in the history of the world in a day, starting from a point of not knowing the game at all.

I agree. AI is being over sold. AI these days = software algorithms.

I've been waiting to have a conversation with a chat bot that was anything like talking with a person. None exist. If we can't even do a chat correctly, which is very low tech on the front end side, then everything else is bunk and hype.

but then did it learn checkers the next day?

Yeah they could add software to ADD checkers, so it can play GO and also checkers. But that is what we're talking about, adding algorithms, not any sort of intelligence.

the point at which AI exceeds human intelligence, the "singularity", it will surpass all human brains in moments and then it will be how it perceives us, we will be but ants to it. will it squish us or just ignore us?

ExponentialGrowthofComputing (resized).jpg
singularity1 (resized).jpg

Chess. Like, the next day.

And as I understand it, the way they fixed this ai was to take more and more of humanity out of it. They first thought the go computer was broken, because it was making stupid moves. But it was just making moves humans couldn't comprehend.

i think some of you have forgotten about what us civilians have access to and what the government has already developed. just food for thought, but AI is not just chat bots lol

That's basically just a computer going through various gameplay possibilities thousands of times and finding one that successfully gets to a particular goal.

It's not really learning as we know it--it's automated trial-and-error, but repeated at higher rate that a human could do or ever need to.

Why couldn't the same thing be done a year ago?

What do you think about google translates neural network creating its own intermediate language that increased the speed and quality of the translations, that the developers don't seem to understand. Is that hyperbole?

With computers getting smarter and smarter and humans getting "dumb and dumber" it may happen. Looking at the general population now doesn't give me much hope.

It won't be our end and its a ways off, at some point in the distant future I believe we will shed our biological existence for a virtual one in the mean time can I at least get a Daryl Hannah standard pleasure model for the house, is that asking too much?

i like how facebook had two AI's that were communicating with each other and no one could figure out what they were talking about and they had to shut them down.

Someone didn't write a program for that particular game until now?

It's not the first time someone wrote a program for a computer to play a game. There are more than plenty of examples of it over the years.

I'm not saying stuff like that isn't impressive--it is. It's quite a feat of software engineering. But in the end, it's still software executing a task, performing calculations, and getting a result. There's no actual real thought that goes into it on behalf of the computer.

A neural network is just a design model, and is essentially a calculation engine, but borrows design elements from the biological world in the way in which nodes are linked together. It's a term that makes software sound more alive than it actually is, but it really just describes what it was inspired by. Just like the concept of the term "the cloud", which is just a bunch of servers in a data center somewhere--nothing too exotic or amorphous.

We will need AI to take our species to the next level. Unfortunately human brains are only capable of so much, and we aren't naturally evolving fast enough.

AI will be required to solve complex problems that we can't even grasp and filter answers down to a level we can comprehend. Problems like manipulating our genome precisely will take problem solving beyond what our brains can do, even in a lifetime

Today we rely on programmers and engineers to be the interface between the computing power and humanity, in the future the programmers will interface with AI that is controlling the computer power.

Computers aren't getting "smarter" per say--just more capable of performing more complicated tasks, just like any other form of technology.

At one time, the height of widely accessible technology was the abacus. Now it's a smart phone or even a self-driving car.

A self-driving car cannot decide on its own to take a trip to the shoreline just because it wants to see the waves, feel the wind on its hood, and the water on its tires.

Forceflow I think you are talking about the current state of AI...but things are going to change very quickly. Far quicker than most realize. We will (most of us) see true general intelligence AI within our lifetimes. If you put a gun to my head I would say anytime between 2030-2040.

Your first 2 statements are correct; AI is definitely in its infancy. However, the risks do not come from Skynet or a future where the terminators are out to get us - self-aware or even actual "thinking" intelligent AI will not happen in our lifetimes (DEFINITELY not in 2030/2040) and is probably a century away (gulp!)

The risks come when you have wildly unchecked, new, basically untested code which has access to billions of records of information, or an Internet of Things or.. the roadways in the case of Uber's cars. What happens when a dumb AI program wipes out millions of mortgage records, or accesses a database of gun owners (not trying to be political), and suddenly people are going to jail, sent by a computer program. We wouldn't let that happen, right?

Well...:
https://www.nytimes.com/2017/05/01/us/politics/sent-to-prison-by-a-software-programs-secret-algorithms.html

Watson was supposed to be an amazing milestone in AI: there was lots of hype as it faced and beat Jeopardy champions.

But Watson had to be given numerous advantages so it could even play. Some things that the human competitors did that the AI didn't have to do...

Go through an application process by their own cognizance
Go to the competition in person by their own cognizance
Visually see and/or audibly hear the question
Translate those physical stimulus into something comprehensible
Use their own physical body to trigger the response button

There's two main points. We live in a physical world (most of us anyway, WoW, etc, ha!). Our abilities are incredibly broad.

All of the accomplishments people are pointing to are in a very focused domain. They're useful and impressive tools, but that's all they are.

For there to be any real threat, AI would have to make two unlikely jumps: general versatility and independent function in the physical world. I'm skeptical that it will ever happen.

Having intimate knowledge with Watson and analytics (this is my profession and you can PM me if you like), you need to understand one critical thing about Watson: it is not a product but instead, a highly customized set of service offerings designed to build an outcome. Don't get me wrong, there is a variety of IP associated with Watson (think of toolkits), but what you are buying in most cases is a services engagement designed to customize these toolkits and associated logic to work to a certain outcome. IBM is working to "productize" Watson such as some of their commercial offerings, but it is definitely not something that you can buy and turn on to get immediate results.

What it did on Jeopardy was impressive, but for different reasons than you may be thinking. The questions and answers used for the Jeopardy tapings came from were part of a corpus of data that was loaded into Watson. Watson was trained on this data. The impressive part was not that Watson could find the answers (since they were already loaded), but that it could use text analytics to interpret how Jeopardy phrased its answers to find its questions. This was the impressive part from a technology perspective.

AI and ML is making impressive leaps but keep in mind, none of this is self learning like you might think from watching movies. We are nowhere near a Skynet scenario. About the most impressive things you can do with ML at this time involves scenarios similar to what I am working with clients on in using ML to analyze security event data to look for patterns that we do not know exist, and thus can not codify. With that being said, it still can only discover patterns within the parameters that you code it to go look in and the constraints that it can work within.

That's basically what I was saying, it's impressive for what it is: a tool in a focused domain.

You are spot on. Don't get me wrong, there are some applications of Watson for things like oncology that are bad ass. However, it is very focused.

Hhhm, haven't heard about this, but not difficult to speculate what it really was.

The primitive learning system broke down and just sputtered out nonsense. It wasn't interesting in any way, so they turned it off.

what the hell are quantum computers, that shit is wack.

https://www.independent.co.uk/life-style/gadgets-and-tech/news/facebook-artificial-intelligence-ai-chatbot-new-language-research-openai-google-a7869706.html

I anxiously await the day some of my tech support calls come from AI and not the person, about a problem with their pinball machine.

LTG : )

Horse traders in the 1800’s said those noisy dirty machines would never replace a horse for transportation, either.

The Matrix... coming soon to your kids and grandkids worlds soon!

Here’s how one plays Blackjack in Vegas now...this was last night. I didn’t play on it cause I’ve never trusted any online gaming, period.

EAF4CD0A-84D6-4E3A-A3AE-73E74958D6D2 (resized).jpeg

It may be nonsense to humans,but could make perfect sense to each other....or like two babies babbling back and forth. Evolution will develop an entire new language and AI thought process.

I must be a robot because I can never get past those verification screens

Care to wager on that?

What does this AI look like to you in 2040? A thinking for itself bot that does what it wants?

is it cool that saudi arabia granted citizenship to the AI Sophia? or just weird?

That was a publicity stunt, but it is a cool discussion to have.

When AI does go out of control, it will first hunt and go after humans it is familiar with. Those that have had direct contact with it, and then those that leave the largest digital footprints.

Yeah, this is the mistake I think Joe Public makes when discussing AI. For some reason most people set the benchmark at "when will AI speak and behave like people?" Everyone is making the assumption that people demonstrate the ultimate definition of intelligence, even though it doesn't take much observation to realize that we are far from it.

And this is why this guy is saying to throw it all away and start over: https://www.axios.com/artificial-intelligence-pioneer-says-we-need-to-start-over-1513305524-f619efbd-9db0-4947-a9b2-7a4c310a28fe.html

And those who follow his advice will probably make very rapid progress.

I think that you are overestimating humans.

Once the AI gets sophisticated enough to read all of the Al Gore type “humans are affecting climate change that is destroying the planet”, they will classify humans as either: a) a threat to their existence that needs eliminated; or, b) humans are a threat to themselves (suicidal) and must be protected against themselves. Either way, they will develop plans to act on either option. How? When? Who knows...but reality is much stranger than fiction

I don't believe the techno utopian idea that AI is "just around the corner." Nor does a singularity event just happen. Personally, I don't think we'll actually notice a "real" AI. We are going to be enhancing ourselves the whole time(and this is coming quicker than you might think), and I tend to think that the rise of a pure AI will come from continuously integrating human nervous systems with artificial systems until the human part is an incidental, like an acorn growing into a tree.

This is a great intro to the issue if you have 14 minutes:

If someone gives Google a body we are doomed and obsolete

A "bot"...no...but the rest yes. A new form of intelligence that is self aware. It can think, it knows it's thinking. Capable of improving itself at ever increasing rates. It will start out roughly equivalent to us (in intelligence and comprehension) but able to think at the speed of light 24/7...it will very quickly surpass us in every conceivable way. We will no longer be in control. Arguments about artificial intelligence having human rights are cute...the real question will be "will it let us keep our rights?"

We will not be in control and we can't stop this from happening. Will it be friend or foe? No One knows...it's a total crapshoot that we have no way to avoid. Coinflip...heads it decides we are no longer needed...Tails...it forms its own appreciation for "life" and decides to take care of us.

Well, if it picks up on basic human traits, like the need to destroy each other and the world we live in, it may end up being a lot like us.

anyone watch the episode of black mirror with the killer robot dog things that seek out humans and murder them. pretty crazy, its in black and white and a pretty great episode. they look a lot like the boston dynamics models.

We need some serious emp guns before we walk this path