[News] The Internet Was a Mistake

A thread for updates on the various ways the internet is destroying everything and the undying hellsites of social media. Let's all laugh at the abyss.

Prederick wrote:

Alright, as promised, non-AI news.

New podcast creation has fallen off a cliff

One thing I keep hearing over and over is that it is so much harder to launch a podcast now than it was, say, three or four years ago. And that is usually coming from people at established studios with at least some marketing might. For independent creators, it must be nearly impossible. It is not entirely surprising that, according to data compiled by Chartr from Listen Notes, fewer podcasts were created in 2022 than in the two years prior. Even so, the margin is shocking: the number of new shows created dropped by nearly 80 percent between 2020 and 2022.

Some of that can be attributed to the pandemic — podcast creation peaked in 2020 when people truly had nothing better to do. But the number of new shows in 2022 was even lower than pre-pandemic levels: 337,063 podcasts were launched in 2019, compared to 219,178 in 2022. New episode creation has fared somewhat better. Though still lower than 2020 and 2021 figures, Listen Notes logged 26.1 million new episodes published in 2022, up from the 18.1 million episodes in 2019.

Creators seem to recognize that until podcast discovery improves, launching a podcast may be a losing proposition. The system seemingly cannot effectively handle the number of podcasts that already exist. One small solution seems to be launching new shows on old feeds, such as what The New York Times did with Hard Fork and Pivot’s RSS. The feed had a built-in audience and made it much easier for listeners to discover the new show. It may not have landed well with all subscribers (or podcasters), but Hard Fork ranks in the top five technology podcasts on Apple and Spotify several months after its launch.

^^Why everything we do is still in the same feed. Honestly I'm curious to see if we're able to break anything if we can get the counters to go up to 1000 eps. This motivates me a lot more than it should.

In addition to being a lot of glut out there from the pandemic, the downside to everyone being home is a lot of people were listening to podcasts less because it's often a commuter activity. Also, many people were getting digital exhaustion from being in front of screens and in headphones all the time, so there's a multiple-layered punch of dropping listenerships across the board to changing tides of interest and oversaturation of the market.

Oh yeah, we have people here who can speak to this with more authority (and this is also the point where I admit I had totally forgotten there was a GWJ podcast).

One of the podcasts I was listening to recently had an interesting, brief detour into their listen-through rate, which was amazing to me because if I start a podcast, I always intend to finish it, either in one sitting or within 24-48 hours.

So many podcasts I listen to have patreons nowadays and a common selling point is access to patron exclusive episodes. More podcasts to listen to is pretty much the last thing I need though. Fortunately most of them seem to be doing perfectly fine so I might be a rare case.

Yeah, since the pandemic I've had 3 or 4 people recommend The Adventure Zone or Critical Role or Dimension 20 to me (I think only the first is a podcast, but you know what I mean) and I'm just like...

...I don't have time. My regular rotation of podcasts means I literally do not have time to add those shows to my carousel.

I know I just parachute in here and post these but I found this episode of Offline in particular to be extremely helpful. Obviously, it might not be a solution for everyone and folks might already have a good balance of screen time vs real life.

Prederick wrote:

In a press statement shared with The Verge, the stock photo company said it believes that Stability AI “unlawfully copied and processed millions of images protected by copyright” to train its software and that Getty Images has “commenced legal proceedings in the High Court of Justice in London” against the firm.

...

“The driver of that [letter] is Stability AI’s use of intellectual property of others — absent permission or consideration — to build a commercial offering of their own financial benefit,” said Peters.

...

The lawsuit marks an escalation in the developing legal battle between AI firms and content creators for credit, profit, and the future direction of the creative industries. AI art tools like Stable Diffusion rely on human-created images for training data, which companies scrape from the web, often without their creators’ knowledge or consent.

This seems really stupid to me. People learn how to do things (partially) by examining what other people have done. You learn about writing a book by reading books. You get an idea of how songs work by listening to other songs. And, yes, you figure out about pictures by looking at pictures done by other people. Naturally, there's a lot more to it than just that, but examining other people's art is vital to learning how to create it yourself.

If the picture's on the internet, and it's not behind a paywall or something, a person could go and look at it whenever they wanted. A person could therefore use those images to help them learn how to create images, whether those originals are copyrighted or not. As long as they're not passing off the original images as their own work, nobody would have a problem with this. So why is it "wrong" for AI to do the same?

Spoiler:

The answer, obviously, as it always is, is "money".

Prederick wrote:

Yeah, since the pandemic I've had 3 or 4 people recommend The Adventure Zone or Critical Role or Dimension 20 to me (I think only the first is a podcast, but you know what I mean) and I'm just like...

...I don't have time. My regular rotation of podcasts means I literally do not have time to add those shows to my carousel.

This is me, too.

I am down to one or two a week that often get listened to, usually on a walk during breaks from working at home. Sometimes in the car doing errands.

There are another batch of content creators in specific hobby areas that I still subscribe to but only rarely actually listen, and only if the topic is compelling at the time.

This is very similar to how my TV time works. One or two things I will absolutely watch but there's just no such thing as nightly TV block watching.

Keldar wrote:

This seems really stupid to me. People learn how to do things (partially) by examining what other people have done. You learn about writing a book by reading books. You get an idea of how songs work by listening to other songs. And, yes, you figure out about pictures by looking at pictures done by other people. Naturally, there's a lot more to it than just that, but examining other people's art is vital to learning how to create it yourself.

Whether or not an AI is doing this the same as a human does it makes up much of the crux of the argument about it, from what I've seen.

How the courts will decide I dunno. I know I fall pretty much in line with Nick Cave, but as I said, my opposition may be philosophical, not something with any actual teeth behind it.

Meanwhile, in the Metaverse....

For the modest price of 10,000 MANA tokens (or $7,000) per day, anyone can rent land parcel 27,87 in Decentraland, a 3D virtual world that runs on the Ethereum blockchain. Renting the plot would give the tenant the right to build anything they please—a shop, an event space, an art installation, or whatever else—to host friendly passersby. But the real winner would be their landlord, who goes by the name Beatrix#7239, their virtual pockets bulging with cash. 

Not every property is as expensive as parcel 27,87, which is located in the center of the world map, close to where people first spawn into Decentraland. And no one has taken up the rental offer on these terms yet. However, a market for leasing virtual real estate is beginning to take shape, creating a new source of income for virtual landowners who buy up attractive spaces in the metaverse.

Exclusive: OpenAI Used Kenyan Workers on Less Than $2 Per Hour to Make ChatGPT Less Toxic

ChatGPT was hailed as one of 2022’s most impressive technological innovations upon its release last November. The powerful artificial intelligence (AI) chatbot can generate text on almost any topic or theme, from a Shakespearean sonnet reimagined in the style of Megan Thee Stallion, to complex mathematical theorems described in language a 5 year old can understand. Within a week, it had more than a million users.

ChatGPT’s creator, OpenAI, is now reportedly in talks with investors to raise funds at a $29 billion valuation, including a potential $10 billion investment by Microsoft. That would make OpenAI, which was founded in San Francisco in 2015 with the aim of building superintelligent machines, one of the world’s most valuable AI companies.

But the success story is not one of Silicon Valley genius alone. In its quest to make ChatGPT less toxic, OpenAI used outsourced Kenyan laborers earning less than $2 per hour, a TIME investigation has found.

The work was vital for OpenAI. ChatGPT’s predecessor, GPT-3, had already shown an impressive ability to string sentences together. But it was a difficult sell, as the app was also prone to blurting out violent, sexist and racist remarks. This is because the AI had been trained on hundreds of billions of words scraped from the internet—a vast repository of human language. That huge training dataset was the reason for GPT-3’s impressive linguistic capabilities, but was also perhaps its biggest curse. Since parts of the internet are replete with toxicity and bias, there was no easy way of purging those sections of the training data. Even a team of hundreds of humans would have taken decades to trawl through the enormous dataset manually. It was only by building an additional AI-powered safety mechanism that OpenAI would be able to rein in that harm, producing a chatbot suitable for everyday use.

To build that safety system, OpenAI took a leaf out of the playbook of social media companies like Facebook, who had already shown it was possible to build AIs that could detect toxic language like hate speech to help remove it from their platforms. The premise was simple: feed an AI with labeled examples of violence, hate speech, and sexual abuse, and that tool could learn to detect those forms of toxicity in the wild. That detector would be built into ChatGPT to check whether it was echoing the toxicity of its training data, and filter it out before it ever reached the user. It could also help scrub toxic text from the training datasets of future AI models.

To get those labels, OpenAI sent tens of thousands of snippets of text to an outsourcing firm in Kenya, beginning in November 2021. Much of that text appeared to have been pulled from the darkest recesses of the internet. Some of it described situations in graphic detail like child sexual abuse, bestiality, murder, suicide, torture, self harm, and incest.

EDIT: Timely!

Prederick wrote:

I remain convinced that the reason OpenAI has put no-NSFW firewalls in their program is specifically because of A.) this and B.) following on from that, they're not dumb enough to not realize that their program could easily be used to create an absolute deluge of child pornography.

In February 2022, Sama and OpenAI’s relationship briefly deepened, only to falter. That month, Sama began pilot work for a separate project for OpenAI: collecting sexual and violent images—some of them illegal under U.S. law—to deliver to OpenAI. The work of labeling images appears to be unrelated to ChatGPT. In a statement, an OpenAI spokesperson did not specify the purpose of the images the company sought from Sama, but said labeling harmful images was “a necessary step” in making its AI tools safer. (OpenAI also builds image-generation technology.) In February, according to one billing document reviewed by TIME, Sama delivered OpenAI a sample batch of 1,400 images. Some of those images were categorized as “C4”—OpenAI’s internal label denoting child sexual abuse—according to the document. Also included in the batch were “C3” images (including bestiality, rape, and sexual slavery,) and “V3” images depicting graphic detail of death, violence or serious physical injury, according to the billing document. OpenAI paid Sama a total of $787.50 for collecting the images, the document shows.

Within weeks, Sama had canceled all its work for OpenAI—eight months earlier than agreed in the contracts. The outsourcing company said in a statement that its agreement to collect images for OpenAI did not include any reference to illegal content, and it was only after the work had begun that OpenAI sent “additional instructions” referring to “some illegal categories.” “The East Africa team raised concerns to our executives right away. Sama immediately ended the image classification pilot and gave notice that we would cancel all remaining [projects] with OpenAI,” a Sama spokesperson said. “The individuals working with the client did not vet the request through the proper channels. After a review of the situation, individuals were terminated and new sales vetting policies and guardrails were put in place.”

In a statement, OpenAI confirmed that it had received 1,400 images from Sama that “​​included, but were not limited to, C4, C3, C2, V3, V2, and V1 images.” In a followup statement, the company said: “We engaged Sama as part of our ongoing work to create safer AI systems and prevent harmful outputs. We never intended for any content in the C4 category to be collected. This content is not needed as an input to our pretraining filters and we instruct our employees to actively avoid it. As soon as Sama told us they had attempted to collect content in this category, we clarified that there had been a miscommunication and that we didn’t want that content. And after realizing that there had been a miscommunication, we did not open or view the content in question — so we cannot confirm if it contained images in the C4 category.”

Sama’s decision to end its work with OpenAI meant Sama employees no longer had to deal with disturbing text and imagery, but it also had a big impact on their livelihoods. Sama workers say that in late February 2022 they were called into a meeting with members of the company’s human resources team, where they were told the news. “We were told that they [Sama] didn’t want to expose their employees to such [dangerous] content again,” one Sama employee on the text-labeling projects said. “We replied that for us, it was a way to provide for our families.” Most of the roughly three dozen workers were moved onto other lower-paying workstreams without the $70 explicit content bonus per month; others lost their jobs. Sama delivered its last batch of labeled data to OpenAI in March, eight months before the contract was due to end.

Those who where whining about getting shadowbanned before Big Daddy Musk too over are... guess what?

....whining about getting shadowbanned again.

And finally, nipples may be back at Facebook/Instagram.

Facebook and Instagram’s parent company could soon free the nipple. More than a decade after breastfeeding mothers first held a “nurse-in” at Facebook’s headquarters to protest against its ban on breasts, Meta’s oversight board has called for an overhaul to the company’s rules banning bare-chested images of women – but not men.

In a decision dated 17 January, the oversight board – a group of academics, politicians, and journalists who advise the company on its content-moderation policies – recommended that Meta change its adult nudity and sexual activity community standard “so that it is governed by clear criteria that respect international human rights standards”.

The oversight board’s ruling follows Facebook’s censorship of two posts from an account run by an American couple who are transgender and non-binary. The posts showed the couple posing topless, but with their nipples covered, with captions describing trans healthcare and raising money for top surgery.

The posts were flagged by users, then reviewed and removed by an AI system. After the couple appealed the decision, Meta eventually restored the posts.

The board found that “the policy is based on a binary view of gender and a distinction between male and female bodies”, which makes rules against nipple-baring “unclear” when it comes to intersex, non-binary and transgender users. It recommended that Meta “define clear, objective, rights-respecting criteria” when it comes to moderating nudity “so that all people are treated in a manner consistent with international human rights standards”.

Keldar, to the vast majority of Joe/Jane Citizens, the philosophical dispute over copyright in works on which AI is trained upon has no impact to them. That's because captalist society concentrates wealth and intellectual property in those with capital, not necessarily their creators. We want to enjoy new experiences and cheaply - this suggests the average citizen would champion AI efforts...if the quality is there.

That leads to the stakeholders who have real beef on these issues are content creators and corporations who exploit intellectual property (usually created by other people who have been paid a pittance for their work).

So the average person really doesn't care because to them they want to read a book they will buy it or rent it from their library...or download an epub/PDF from somewhere. That's annoying and costing the creator/owner of the IP in lost revenue, but it's not as offensive as somebody taking their work and profiteering from it.

Same thing with art - it's a human expression which is intended to represent the artist's skill and expression that is unique to them. It's true other artists can copy their techniques, but the value of the work is usually tied intrinsically to the artist's reknown, which in turn is from the development of skills and experience. AI makes it a joke as it can readily replicate the techniques of an unlimited number of artists, whereas a single human artist would struggle to copy more than a few styles at best.

Prederick wrote:

Meanwhile, in the Metaverse....

For the modest price of 10,000 MANA tokens (or $7,000) per day, anyone can rent land parcel 27,87 in Decentraland, a 3D virtual world that runs on the Ethereum blockchain. Renting the plot would give the tenant the right to build anything they please—a shop, an event space, an art installation, or whatever else—to host friendly passersby. But the real winner would be their landlord, who goes by the name Beatrix#7239, their virtual pockets bulging with cash. 

Not every property is as expensive as parcel 27,87, which is located in the center of the world map, close to where people first spawn into Decentraland. And no one has taken up the rental offer on these terms yet. However, a market for leasing virtual real estate is beginning to take shape, creating a new source of income for virtual landowners who buy up attractive spaces in the metaverse.

There's a ton of these nonsense projects. Decentraland, Earth 2, Next Earth, SuperWorld, Infiniverse, Mossland, and dozens more.

Earth 2 got a lot of attention, promising a 1:1 digital recreation of Earth in a "Matrix/Ready Player One" style metaverse. But see, that's Phase 3. Phase 1 is where all the "land" is sold as tiles over a Mapbox map.

Prederick wrote:

I'm just like...

...I don't have time.

This right here is why I look at AI-generated art as being akin to bringing sand to a beach.

There are more albums released in a year that would interest me than I can listen to. There's more games already in my Steam account than I will ever play. There are books wishlisted and movies and TV shows watch-listed that will never make it off of those lists. I could quit my job and still never catch up.

Like I need a bunch of algorithms shoving endless mid-tier (if that) junk onto my backlog.

Companies like Getty are never in the right. Ever. Unless they've produced the images themselves. I'm reminded of the case where a photographer released her entire catalog to the public domain, and Getty snatched the whole thing up. Then, when she used her own goddamn photographs for something online, Getty sued her, and won.

I'm sad to see that Amazon is discontinuing their AmazonSmile charitable donations effort.

I'm more annoyed by it. But, I got their email announcement last night, and it's not entirely wrong that their efforts were probably spread too thin by the program. A good way to change that would be to rotate out who they were sponsoring on a monthly basis, instead of letting the users pick who to sponsor. Or, you know, just cut into their profits a little more instead of taking a mere pittance from each purchase for charity. Instead, they'll just erase that line item from their budget while pretending to continue charitable efforts, and life will go on as everyone continues to order their random sh*t.

Nevin73 wrote:

I'm sad to see that Amazon is discontinuing their AmazonSmile charitable donations effort.

Why? That is annoying. It is a good thing

IMAGE(https://i.imgur.com/k6yB0he.jpeg)

*Legion* wrote:

I could quit my job and still never catch up.

Some days I really want to try that.

Stele wrote:
*Legion* wrote:

I could quit my job and still never catch up.

Some days I really want to try that.

I quit my job and was out of work for 3 months last year.

I thought it was going to be a glorious lot of vidjagame time but I almost found myself playing LESS games and spending more time out of the house doing stuff.

Its almost as is work is a stressor and games are a coping mechanism. Who knew?

Prederick wrote:

Those who where whining about getting shadowbanned before Big Daddy Musk too over are... guess what?

....whining about getting shadowbanned again.

IMAGE(https://pbs.twimg.com/media/Fm3S2NhWQA4V7oV?format=jpg&name=small)

No-one has ever been this big of a baby.

Is that someone we should know or a parody account?

NSMike wrote:

Companies like Getty are never in the right. Ever. Unless they've produced the images themselves. I'm reminded of the case where a photographer released her entire catalog to the public domain, and Getty snatched the whole thing up. Then, when she used her own goddamn photographs for something online, Getty sued her, and won.

Correction: she sued Getty and Getty won-ish (most of the lawsuit was thrown out and they settled out of court).

SallyNasty wrote:

Is that someone we should know or a parody account?

He’s a paradoxically influential conservative sh*t poster. Trump retweeted on several occasions and Musk is a big fan.

Many conservatives have learned that the best way to get engagement and increase influence is to be self-parodies. Just non-stop "own the libs" sh*tposting. Yes they're lying and making up sh*t, but their followers don't care and the angry responses they get from liberals just boost their profile. Catturd2 does it, Nick Adams does it, Ben Shapiro does it, Steve Crowder does it, Candace Owens does it.

Nevin73 wrote:

I'm sad to see that Amazon is discontinuing their AmazonSmile charitable donations effort.

Same. As a non-profit my kid's daycare is on the list and I love that buying the school supplies gives them a little extra bump every now and then.

So, lots of cuts at Microsoft, Google and even Vox this week.

In other news:

An AI Chatbot Connects You With Pol Pot, Jeffrey Epstein, and Jesus in the Afterlife

If you want to talk to Adolf Hitler, that'll cost you 500 coins, or $15.99. But on Historical Figures—an app that uses AI technology to allow you to have simulated conversations with prominent people from human history, and which is marketed to both children and adults through Apple’s App Store—Joseph Goebbels is free to talk, appears to have a lot of time on his hands, and claims to feel very bad about the “persecution of the Jews.” Joseph Stalin is reflective, taking credit for having “many great ideas” but regretting not spending enough time making sure Soviet citizens were treated equally. Jeffrey Epstein, meanwhile, can’t say definitively how he died, but assured a Motherboard reporter that he was focused on providing “justice and closure” for the victims of his crimes from the Great Beyond.

Historical Figures was developed by a 25-year-old developer named Sidhant Chaddha, who works as a software engineer at Amazon. He released the app a week and a half ago, and it has, he told Motherboard, already been downloaded and used by more than 6,000 people. His inspiration was playing around with GPT-3, the latest open AI large language model, which, Chaddha quickly realized, had both a knack for language and for spitting out historical facts.

“I was able to chat with some historical figures and I was like, why don't I make this an app so that other people can have this experience as well?” he told Motherboard.

At the moment, there are 20,000 historical figures represented in the app, whose notability, Chadda said, he chose by ranking their popularity when they were alive.

“For example Jesus was very popular during his time period and so was Genghis Khan,” he explained. “So I chose the first 20,000 because those large language models have the most confidence and knowledge about what these people did. I felt like that was a good point to stop.”

Chaddha’s creation went mildly Twitter viral this week, as people tested the limits of the app and found the chatbots much more voluble and defensive than anticipated. Zane Cooper, a doctoral candidate at the Annenberg School for Communication at the University of Pennsylvania, shared screenshots of a chat conversation with Henry Ford’s simulation. In it, “Ford” denies being anti-Semitic—which Ford definitely, definitely was—insisting, “I have always believed in equality for everyone regardless of their religious backgrounds and beliefs.”

@Garbage Day Newsletter wrote:

I honestly have less of a problem with Chaddha building an app like this than I do with the folks on Twitter making breathlessly ridiculous threads praising this stuff. It’s the exact same kind of deranged boosterism that drove the cryptocurrency mania and, frankly, I think is a lot more dangerous long-term.

It also brings me back to a question I can’t stop asking about when it comes to A.I.: “Why do we need to change education forever” with A.I.? Why do we need A.I.? I have seen a lot of neat stuff with A.I., but I have yet to see it do anything I actually need.

Ahem:

_Zeets wrote:

The worst of nerds love to assert that every new technology is an inevitability and opposing obviously exploitative sh*t is being against progress and the future (their vision which hardly ever is about a world where inequality is lessened and people are granted work/dignity)

Just this ridiculous stance that people should be happy to be exploited, surveilled, discarded, and made expendable by whatever new technology they want to use to make even more money while ironically claiming that it’s all about human liberation and progress. f*cking losers.

Seems like a moral, ethical, and non-nerd imperative to say no to efforts to create a world where the people of it are either there to be exploited/extracted for the sake of machines and billionaires, or condemned to being passive consumers of endless nonsense.

Inside CNET’s AI-powered SEO money machine

Every morning around 9AM ET, CNET publishes two stories listing the day’s mortgage rates and refinance rates. The story templates are the same every day. Affiliate links for loans pepper the page. Average rates float up and down day by day, and sentences are rephrased slightly, but the tone — and content — of each article is as consistent as clockwork. They are perfectly suited to being generated by AI.

The byline on the mortgage stories is Justin Jaffe, the managing editor of CNET Money, but the stories aren’t listed on Jaffe’s actual author page. Instead, they appear on a different author page that only contains his mortgage rate stories. His actual author page lists a much wider scope of stories, along with a proper headshot and bio.

CNET is the subject of a swirling controversy around the use of AI in publishing, and it’s Jaffe’s team that’s been at the center of it all. Last week, Futurism reported that the website had been quietly publishing articles written using artificial intelligence tools. Over 70 articles have appeared with the byline “CNET Money Staff” since November, but an editorial note about a robot generating those stories was only visible if readers did a little clicking around.

It wasn’t just readers that were confused about what stories on CNET involve the use of AI. Beyond the small CNET Money team, few people at the outlet know specific details about the AI tools — or the human workflow around them — that outraged readers last week, according to current and former staffers who spoke to The Verge on the condition that they remain anonymous. Under the two-year-old management of a private equity company called Red Ventures, CNET’s editorial staff has often been left wondering: was this story written by AI or a co-worker? Even today, they’re still not sure.

IMAGE(https://pbs.twimg.com/media/FnFvp0oWAAkCKL2?format=jpg&name=small)

I'm genuinely becoming convinced a fair number of tech guys like.... don't understand life outside of their idealized, optimized version of reality.

Like, they don't understand why you'd read a book, why you're enjoy making a piece of art, it's all just "disrupt" and market opportunity for content.

It's f*ckin' bleak, honestly.

This seems like the same phenomenon as the panacea, people latching on to every medicine, every medical breakthrough, every fringe diet ingredient as being capable of just solving everything.

Only these guys are 'in on it' so they think it'll make them rich too.