Tiny Brontosaurus posted:

Yes, going out in society is a social activity. Maybe you trudge and sulk and shrink away hissing anytime someone makes eye contact, but I know the cashiers who work at my store and talk to them, and frequently run into people I know or have some reason to chat with the other people there. I wonder just how much of human existence you've decided you're too good for.


That's easy mode, yo. You should develop skills on coping with strangers when you don't have a structured activity and someone else making the introductions for you. And PSA: If you don't teach your kids how to shop for food they'll grow up to be fat takeout addicts like you.

Jesus Horse posted:

France has a fake jobs program
https://www.nytimes.com/2015/05/31/business/international/in-europe-fake-jobs-can-have-real-benefits.html

Owlofcreamcheese posted:

I think this has come up before but is the 2010s the ideal peak of human society or is there any automation technologies from earlier times we should have banned/restricted/made illegal?

Like apparently 350,000 people were employed as AT&T telephone switchboard operators. Should they have had their job protected by law and should we mandate we return that sector to human hands? Or is it only future technology based job loss that is bad and all the ones in the past were good? Are there ones in the future that might be good and are there ones in the past that were bad? What is the metric?

CrazySalamander posted:

https://imgur.com/gallery/gkLFz

Algorithms have granted us a tantalizing look at Harry Potter and the Portrait of what Looked Like a Large Pile of Ash.

Jokes aside, it seems like it was at least semi human directed- they describe training predictive keyboards on the Harry Potter books, so what I presume occurred was the human would pick a letter and hit enter if the word had promise or another letter if they wanted to force it in a different direction. The biggest clue to this effect is the quote "The password was 'BEEF WOMEN'," two words that I would lay money never occurred next to each other in the Harry Potter books, let alone in all caps.

It's a bit irritating that they make it sound like it was completely auto generated- as random as it seems it is most definitely not typical of a markov style walk of this length. It is way too cogent for that.

Owlofcreamcheese posted:

Rich companies won't "help" anyone ever, but giant companies like walmart that rely on there being millions of low income shoppers will always do everything in their power to make sure there is low income shoppers. That doesn't make them your friend if they champion funding to make sure american society doesn't get TOO poor, because they also will champion causes to make sure american society doesn't get too well off for them either. Like there is a reason walmart is one of the world's largest donor of food to food banks while also having a huge worker population on food stamps.

Hot Dog Day #82 posted:

Matt Damon’s Elysium could also be a good look at the future, since it has a lot of super poor people just kind of milling around between working lovely soulless jobs and being policed by an unsympathetic justice system staffed by robots.

Raspberry Jam It In Me posted:

I highly doubt that blue denim jeans are fat people clothes. S and M are always the first sizes to be sold out with everything in the shop and the leftover sale rack is always ~90% clothes in L+. It's just incredibly lovely inventory management. Sentient inventory AI when? Why not done?

Owlofcreamcheese posted:

I disagree with the assertion that there actually is millennia of precedent of tools being used to the detriment of 90% of humanity in any systematic way.

Look at something like world hunger or poverty or access to water or disease or whatever. It's stuff that goes up and down over the years with it seemingly generally going down. Less people are in global poverty today than were 10 years ago. Less people are starving. Less people have guinea worms.

Bad things happen too. Some years hunger goes way up, or disease. The US is a mess and we have way more hook worms than we have in 100 years. There is no absolute rule every single thing improves every single year and nothing ever goes backwards. But there is no rule that everything is bad forever and no one ever uses tools to make anything better.

Someone somewhere is fixing water access. The amount of people without water has gone generally down overall over time. Either some dumb elite that doesn't know better that they were supposed to be satan incarnate and are being nice when they shouldn't be, some elite that is being seemingly nice but has ulterior motives or some non-elite that is fixing their own problem because water is such a widespread technology that it's within their reach now too. There hasn't been some rising tide of people having less and less water as elites grow in power. (nor have elites always grown only more powerful, with that being another thing that sometimes gets better and sometimes gets worse year to year, for both political and technological reasons). I bet in the US there will be less people with access to water in 5 years than there was 5 years ago. I bet globally the opposite.

Yuli Ban posted:

Anyone following news on the front of generative adversarial networks?

muon posted:

but it does not imply that the GAN "knows" how to make a horse/zebra walk like humans can easily imagine. Now we could design a neural net to learn how to do that, but **the only way we know how to do that today requires thousands and thousands of labelled data sets of horses walking** before we could even begin to create that model. Neural networks need to have nuanced semantic understanding and ability to generalize to really be able to help create content in the way you're describing. That's an extremely difficult task that I don't know we'll solve by the 20s.

Yuli Ban posted:

I mentioned this myself, that in the 20s, these algorithms will most assuredly generate increasingly amazing content but they will not understand said content and it's up to you to make sense of it all.

To use the comic example again, I could get the algorithm to generate a person in the style of Jack Kirby; first they have to consume thousands of picture of a person and then thousands of works from Kirby, and then eventually it'll understand how to warp the dimensions of a real human to fit a more stylized version, all to create "Dr. Minsky". I could get the algorithm to make this synthetic comic book Minsky take a variety of poses that I feed into it, or maybe alter his hair color or his wardrobe based on text inputs. I can add speech bubbles (I mean, I could do that really), and then I could create another person in the same manner. GANs will be able to do this by the early 2020s. It seems like a long-stretch, but there's a reason why I'm calling it— mostly due to the fact an actual machine learning expert whose pulse is firmly on the field is saying that what's being done in the labs is about right for the course, and he's still conservative about these things.

They won't be able to actually generate a full comic, start to finish, that makes sense. It's up to me to put every panel in order and fill in the text boxes. It's up to me to decide which panel works and which one doesn't. It's up to me to make the story, the reason why these characters are made in the first place. Sort of as if you have fully automated production of paper, but it's up to you to turn it into a book and sell it. It's true when you say that GANs are limited and will still have limits. I'm just saying that we'll still be able to do amazing things within the confines of these limits.

It's not going to be until the 2030s, at the absolute earliest, that we'll start seeing truly creative AI in any real way. I'd love to be wrong, but that's the way I'm seeing it at the moment.

Owlofcreamcheese posted:

Also if a guy is in a room with a bunch of how to answer Chinese questions flowchart books and answers Chinese questions does anyone really understand Chinese etc etc etc

Owlofcreamcheese posted:

Reading the title and trying to think of which book that is I thought maybe it was either

WWW:Wake which is about the internet coming alive after china seals off it's section of the internet and the perspective of having two individuals is what sparks a thoughtless computer into having a sense of self (and then a blind little girl who gets bionic eyes that don't work but let her see the flow of the internet and that does something else I can't remember because I read this book like 12 years ago).

OR

That it was some short story that I can't remember the name of about a blind man being a world class super genius and science realizing that his brain had rewired to use his visual contrex for problem solving and science figures out a technique to do that repeatably so people end up having to remove their kid's eyes because no amount of study will ever make you as good as an eyeless genius. Until it spreads to nearly any skilled labor and only unskilled laborers have eyes.

Raspberry Jam It In Me posted:

I don't get it. What is it trying to say? Is this a schizophrenia thing?

Trabisnikof posted:

lol if you think these AI tools will be free to use.

Trabisnikof posted:

Of course you stick in all this nonsense about "intentionally horded" to avoid the core part of the criticism, that when a machine replaces a job, rarely can that worker afford to buy that machine outright instead.

It isn't "intentional hording" for expensive machines to replace workers, but that doesn't change the fact that when a job is automated it isn't the worker who will end up owning the machine. Sure, at least the cost of entry to software is lower than mechanical machines, but the one-sided relationship is the same. When software replaces a job the worker can't solve their problems by just buying the software, even if they can afford it.

Tei posted:

I have a problem here.

You guys talk like owning a robot is all would take for a poor to be at the same level a rich guy that owns a robot.

Heres something for you:
https://www.cnet.com/news/facebook-closes-19-billion-deal-for-whatsapp/

Facebook paid 19 billion dollars for a IRC client. Writting a IRC client is not very hard, all it takes is about 2 hours of programming.

http://archive.oreilly.com/pub/h/1968

So whats going on here? Why Facebook paid 19 billions for a "robot" that is really cheap? It seems that you don't care about the code itself (the robot) you care about the captured customers base.

Poors can own robots, the means of production, but thats useless to them. They will still be poors. Theres more here in play that owning the means of productions. Theres also something to say about having a contact with the people that provide raw materials, having bargaining power, having a marketing muscle, and so on.


Lets not be naive. A university student can own a printer, and that will not save him from paying 300$ for a crappy book his teacher wrote. Just owning a means of production don't make people peers from the people with power.

Doctor Malaver posted:

A friend who develops on MS Azure was telling me today how it's insane how many new services they announce on a weekly basis. Face recognition, face mood recognition, text summarizing, text analysis sentence by sentence... He tries to stay up to date but too much is going on. And it all costs like a fraction of a cent per use. And that's just Azure, AWS is bigger and better.

Yuli Ban posted:

And now witness this.
 
Remember that time about a year and a half ago when a man fed Blade Runner into a neural network?
 
Turns out that the neural net was able to remember Blade Runner and managed to transpose its aesthetic onto other films.
The Neural Net That Recreated ‘Blade Runner’ Has the Movie Stuck in Its Memory
The AI that made ‘Blade Runner: Auto-encoded’ transposed the aesthetic of the movie onto other sci-fi classics

Tei posted:

Let me use this space to whine about human memory. Is like a lovely compression algorithm.

Tei posted:

It don't seems to be a huge difference. Or maybe I am too dumb to understand any of this.

It don't sounds like autoencoders work very differently than human memory, only that human memory uses another subsystem (a tokenizer/ a lexer phase) that autoencoders don't include.

Malcolm XML posted:

this is not true

max-pooling gives invariance to a bunch of things assuming you've trained correctly

quote:

you don't need to be perfect to replace a bunch of repetitive jobs esp if you can control for things like rotation and scaling using such tricks as "lenses" and "orienting" like all the fruit classification systems do

Owlofcreamcheese posted:

The future is now

Raspberry Jam It In Me posted:

Tasmantor posted:

Vertical farming also gains in being able to be close to the final destination, don't have to ship it far if it was grown in the 'burbs.

Blue Star posted:

Cant wait to be constantly misgendered by software as it scans my big ugly mannish trans woman face. Gonna be rad.

paternity suitor posted:

Ew, now I'm imagining some sort of app that scans people's reaction to your appearance and gives you a ranking at the end of the day so that you know what to wear...those shoes cost you 2 attraction points Rachel.

Tasmantor posted:

Algorithms have been nothing but good for us so far! If only one could tell me how to dress

Who writes that winner? Apple, Google or Amazon gonna get the right to tell us all how we should dress? gently caress silicon valley I don't think anyone should be able to tell you how to dress, not fashion mags, not highschool kid, not some weird group think you believe in and sure as gently caress not the people who think "move fast and break things" is a cool attitude.

Guavanaut posted:

But like all purely aesthetic choices, what if wearing your shirt inside out becomes the new coolness?

Inescapable Duck posted:

And not even bringing up how automated beauty standards based on celebrities and popular opinion are absolutely going to poo poo on minorities, racial and sexual.

Tasmantor posted:

You don't like t-shirts and jeans, they are comfortable and cheap, they endure for a reason.

What the gently caress is with this thread? Automation is useful at making work easier so we have more time to be humans but all you goony dicks want machines to be human for you so you can capitalism harder.

Like gently caress me, if it's so trivial to dress well that an ai could be made overnight then just loving learn how to dress your self. It will take you ten minutes on YouTube and then day to day, go nuts, ask another flesh bag of you look like poo poo. Sure you got bullied a lot in highschool but that was because you're awkward as hell and kid are cruel Basterds. You're adults now and people like interacting with other people, we are social monkeys, they won't hurt you. The world isn't a game of DND you don't dress to min max on some loving spread sheet, grow up.

I mean at least it's automation of something you are talking about but your analysis of it is "yes I am a social retard, I would love an ai mum to dress me and tell me I am special!". The paperback sci fi about a world of man babies, dressed by all loving algorithms and growing food in their yards by letting a specialised green thumb quad copter do the work, writes itself. It's the saddest most pathetic future you could go for, one where people exist to watch tools live for them instead of work for them.

Somewhere in all the bitching about people only seeing bad futures, with no work but the same work or die society, you all switched to letting the tail wag the dog.