7
votes
Dark side of the industry
Hey I was wondering what if Google is tracking our behaviour and using that neural (whatever the word is) to create artificial human replicating exactly to that human's behaviour..i know it's a weird thought which lead to..what are the dark side of the tech industry which is unheard of, or nobody is paying attention on it
Neural networks don't come close to replicating human intelligence. They seem more to me like an optimization function.
It's all just matrix algebra.
It’s entirely possible human intelligence is equivalent to a bunch of linear algebra, too, though.
Sure, but it wouldn't look like a neural net.
It doesn't quite count as dark, but it's definitely both dodgy and depressing. I read recently about the internet obesity crisis. Websites getting bigger and bigger (ie. more bytes) with more and more tracking built in, while simultaneously having less and less actual content. And how the current advertising model is tied into this.
This is the post I read: http://idlewords.com/talks/website_obesity.htm
To add to that, there is an influx of web apps that are "sold" as desktop or mobile apps by just bundling them with a browser - I'm looking at you Electron.
To be precise, electron add around 30Mb for Windows application (35 for 32bit) and it's definitely not "because electron" but because of the necessary libraries, dlls and whatnot that are needed to interact with the OS the build is for :)
Man, I had to create an application that provide an interface to use a very old device (an industrial scale) but is connected to a modern infrastructure. The previous UI was done with Oracle fucking Forms.
I spent 2 weeks figuring out how to make electron talk with USB and serial ports on windows but apart from that, everything else is done with a lightweight app on a website, built in Angular5.
We can now change the UI and update it whenever the fuck we want as the electron part only allow the app to interact with the hardware connected to the computer and load the online interface.
If you're hating on electron is either because you didn't make enough of an effort to understand how to work with it, or you found a roadblock because you're doing something too niche. If it's the latter, get on with it a help the community growth by publishing your code once you make it work.
Staying there all grumpy or edgy about a technology isn't gonna make you better or help others.
Fair point. Luckily enough I don't have to code it. I have to occasionally use it, which is quite annoying.
Do you want the Matrix? Because that's how you get the Matrix.
With that though..deep down I'm petrified
I always thought the credit card protections were a little, a lot "Handmaid's tale"-ish. The ability to build the profiles that major credit card companies have on each user, which to be honest, is everyone now and to disable them is powerful.
There's no way that could ever happen with our current technology. People often ask me if AI could take over the world and enslave humans or something. To which I usually respond that's like teaching your dog to fetch the newspaper for you and then be worried it's gonna learn how to shoot guns and kill everyone in your house. So you can relax. For now, anyways...
I see what you are saying, but the thing is, you,me,dog,humans,all the living things for that matter has some "Limit" to learn grasp adapt or something similar..but for AI? undefined is the limit.. We can relate it by our recent development in AI, like google Duplex is step 1 of thousands of steps.
That's not exactly how it all works. AI are extremely limited. They are built with one specific goal in mind and that's all it can ever learn. We don't have a single "general knowledge" AI like we see in movies. They all accomplish a very specific task and that's it.
That's because it doesn't learn per se like we do, they're just very good at memorizing data we show them and find similar patterns. If you were to make an AI that recognizes cars in an image, you'd show it many thousands of images with and without cars and tell it "this one has a car" or "this one doesn't". At some point you'd be able to show them a new image and get a reliable answer. That AI won't ever learn English, how to type or how to parse text. That AI won't ever be able to hack into your bank and drain everyone's funds. It was made to analyze pixels in an image and that's all it can ever hope to accomplish.
"In the past five years, six Americans have been shot by dogs." They didn't kill everyone in the house, but still...
WTF lol I guess I need a better analogy now
Actually, I think that's an awesome analogy: AI will harm and kill, but it won't be as bad as in Sci Fi.
All jokes aside, I really like that analogy. It's meant to show that some random chatbot can't learn everything we know and overthrow the government. Sure AI can kill, if they're made to kill. For AI to take over the world we'd pretty much have to make one with that intent and give it that ability. That's very far in the future, if ever.