So where are we all supposed to go now?
Photo by Jaap Arriens/NurPhoto via Getty Images
An era of the internet is ending, and we’re watching it happen practically in real time. Twitter has been on a steep and seemingly inexorable decline for, well, years, but especially since Elon Musk bought the company last fall and made a mess of the place. Reddit has spent the last couple of months self-immolating in similar ways, alienating its developers and users and hoping it can survive by sticking its head in the sand until the battle’s over. (I thought for a while that Reddit would eventually be the last good place left, but… nope.) TikTok remains ascendent — and looks ever more likely to be banned in some meaningful way. Instagram has turned into an entertainment platform; nobody’s on Facebook anymore.
You could argue, I suppose, that this is just the natural end of a specific part of the internet. We spent the last two decades answering a question — what would happen if you put everyone on the planet into a room and let them all talk to each other? — and now we’re moving onto the next one. It might be better this way. But the way it has all changed, and the speed with which it has happened, has left an everybody-sized hole in the internet. For all these years, we all hung out together on the internet. And now that’s just gone.
Why is this all happening right now? Lots of reasons, actually, most of them at least somewhat defensible. The economy has gone sour, and after more than a decade of low interest rates and access to nearly unlimited and nearly free money, companies are finding their funding sources to be fewer and more finicky than ever. Those investors are also asking for real returns on that funding, so all these companies have had to switch from “growth at all costs” to “actually make some money.” Few social networking companies have ever made real money, and so they’re scrambling for new features and pivoting to whatever smells like quarterly results.
The rise of AI is also sending all these companies into a tizzy. Large language models from companies like OpenAI and Google are built on top of data collected from the open web. Suddenly, having all your users and content publicly available and easily found has gone from a growth hack to capitalistic suicide; companies around the industry are closing their walls, because they’re hoping to sell their data to AI providers rather than have it all scraped for free. Much of Reddit’s current chaos started with CEO Steve Huffman saying that the company realized that the platform is filled with good information, and “we don’t need to give all of that value to some of the largest companies in the world for free.” On Saturday, Elon Musk introduced Twitter’s new login gate and view count restrictions “to address extreme levels of data scraping & system manipulation.”
Add it all up, and the social web is changing in three crucial ways: It’s going from public to private; it’s shifting from growth and engagement, which broadly involves building good products that people like, to increasing revenue no matter the tradeoff; and it’s turning into an entertainment business. It turns out there’s no money in connecting people to each other, but there’s a fortune in putting ads between vertically scrolling videos that lots of people watch. So the “social media” era is giving way to the “media with a comments section” era, and everything is an entertainment platform now. Or, I guess, trying to do payments. Sometimes both. It gets weird.
As far as how humans connect to one another, what’s next appears to be group chats and private messaging and forums, returning back to a time when we mostly just talked to the people we know. Maybe that’s a better, less problematic way to live life. Maybe feed and algorithms and the “global town square” were a bad idea. But I find myself desperately looking for new places that feel like everyone’s there. The place where I can simultaneously hear about NBA rumors and cool new AI apps, where I can chat with my friends and coworkers and Nicki Minaj. For a while, there were a few platforms that felt like they had everybody together, hanging out in a single space. Now there are none.
I’d love to follow that up with, “and here’s the new thing coming next!” But I’m not sure there is one. There’s simply no place left on the internet that feels like a good, healthy, worthwhile place to hang out. It’s not just that there’s no sufficiently popular place; I actually think enough people are looking for a new home on the internet that engineering the network effects wouldn’t be that hard. It’s just that the platform doesn’t exist. It’s not LinkedIn or Tumblr, it’s not upstarts like Post or Vero or Spoutable or Hive Social. It’s definitely not Clubhouse or BeReal. It doesn’t exist.
Long-term, I’m bullish on “fediverse” apps like Mastodon and Bluesky, because I absolutely believe in the possibility of the social web, a decentralized universe powered by ActivityPub and other open protocols that bring us together without forcing us to live inside some company’s business model. Done right, these tools can be the right mix of “everybody’s here” and “you’re still in control.”
But the fediverse isn’t ready. Not by a long shot. The growth that Mastodon has seen thanks to a Twitter exodus has only exposed how hard it is to join the platform, and more importantly how hard it is to find anyone and anything else once you’re there. Lemmy, the go-to decentralized Reddit alternative, has been around since 2019 but has some big gaps in its feature offering and its privacy policies — the platform is absolutely not ready for an influx of angry Redditors. Neither is Kbin, which doesn’t even have mobile apps and cautions new users that it is “very early beta” software. Flipboard and Mozilla and Tumblr are all working on interesting stuff in this space, but without much to show so far. The upcoming Threads app from Instagram should immediately be the biggest and most powerful thing in this space, but I’m not exactly confident in Meta’s long-term interest in building a better social platform.
So if not that, what? There’s a good case to be made for apps like WhatsApp and Signal, which at least bring some extra privacy muscle to the table. WhatsApp has been adding more social features over time, particularly Channels, a one-to-many way for creators and brands to talk to all their followers at once. (Telegram is also doing some interesting stuff in this space.) But that’s not social, that’s a news feed. These are still chat apps, meant for talking to one or a few people at a time.
Discord is probably the tool best-suited to capture users’ social needs right now. It’s definitely the best Reddit alternative we have. It’s a clever mix of chat app and broadcast tool, a place where lots of like-minded people could conceivably hang out and connect. But, uh, have you ever been in a Discord with thousands of people? It’s pure chaos, and requires you to either devote your life to keeping up or resolve yourself to missing everything. Discord’s moderation tools are a mess, too, and everyone’s still mad about changing their username.
For all its mess, the social networking era did a uniquely good job of just putting people together in a single place. You didn’t have to pick a server or declare your interests ahead of time; you just showed up, set a password, and got to work. Because everyone was together, these platforms were able to make it trivially easy to find people you like and content that interests you. They were able to learn about you over time, and proactively show you those people and that content before you even had to ask.
This all, of course, came with huge downsides. Retweets and quote tweets made it easy for good content to travel, but also made it easy to mass-harass anyone on Twitter. Meta’s knowledge of its users makes your Explore page more interesting, and only extends the dossier on you available to advertisers. I’m not sure it’s possible to have the good without the bad, and I think the bad might outweigh the good. (As a white guy in America, I also experience the bad far less than many users, and I suspect I’d feel differently about the end of this era if I weren’t quite so privileged here.) But I can’t help but think it’s possible to at least do better.
Maybe we should all embrace the downfall of social networks, and maybe my (and our) need for a global water cooler is just a vestigial feeling we’ll all be rid of in a few years. But even before this era fully ends, before Twitter and Reddit turn into MySpace and Friendfeed and basically disappear from my life, I find myself longing for what they once were. Still are, maybe, just not for long. I miss everybody, and I don’t know if I’ll ever find them again.
TweetDeck is falling apart after Twitter’s rate-limiting fiasco
Illustration by Kristen Radtke / The Verge; Getty Images
Twitter’s power-user focused TweetDeck interface is experiencing major issues after owner Elon Musk announced limits on the number of tweets users can view daily. The Verge has experienced these issues first hand with our publication’s Twitter accounts, and multipleusersacrosstheplatform (including at least one Twitter Blue subscriber, spotted by TechCrunch) are reporting seeing an empty interface that would normally be filled with tweets. Users are reporting varying aspects of the interface being broken, with notifications, mentions, likes, and lists failing to load.
While users have been reporting issues with Twitter throughout the weekend, problems with TweetDeck are likely to create issues for professional and power users of Twitter, who rely on the service to build their brands, promote their work, and — in the case of many newsrooms — as a reporting tool.
The issues are thought to relate to the new limits on viewing tweets that Elon Musk announced on July 1st. Musk claimed that the limits, which he called “temporary” were needed “to address extreme levels of data scraping [and] system manipulation” on the platform. With the new limits applied, Verified accounts (essentially those subscribed to Twitter Blue) were limited to reading 6000 posts a day, while other accounts would be limited to either 600 or 300, depending on their age.
Musk later announced increases in these limits to 10,000 tweets for verified subscribers, and between 1,000 and 500 tweets for other users. Away from TweetDeck, Twitter’s main website and apps appear to be functional as of Monday morning.
A paying user: This is what Tweetdeck looks like now. You remain a genius product innovator. pic.twitter.com/7TDnP0B39G
The limits were announced a day after Twitter started blocking unregistered users from reading tweets for similar reasons. “Several hundred organizations (maybe more) were scraping Twitter data extremely aggressively, to the point where it was affecting the real user experience,” Musk said as justification for the change. Musk is likely referring to AI companies here, with the suggestion that they’re scraping Twitter to train large language models.
Although Musk is characterizing the limits as an intentional measure, it’s the latest example of the platform’s reliability having suffered in recent months. In early March, Twitter experienced an outage that left many users unable to view images or follow links posted on the platform. The problem is believed to have been caused after the sole engineer working to shut down free access to the Twitter API made a “bad configuration change” that “basically broke the Twitter API.”
When contacted for comment, Twitter’s press line responded with an automated poop emoji, as it has done since shortly after Elon Musk took over the company and gutted its communications department.
Why Car Repairs Have Become So Expensive More technology and creature comforts, as well as a parade of new electric vehicles, have complicated the job after accidents.
Russia Seeds New Surveillance Tech to Squash Ukraine War Dissent Russia is incubating a cottage industry of new digital surveillance tools to suppress domestic opposition to the war in Ukraine. The tech may also be sold overseas.
Tesla Sales Surge as Tax Credits Fuel Demand Incentives and price cuts made Tesla electric cars cheaper than comparable gasoline models. But the company faces growing competition in China, a key market.
Apple considered a finger-worn controller for the Vision Pro
The Vision Pro almost had a dedicated controller. | Image: Vjeran Pavic / The Verge
Apple’s new Vision Pro headset, which is coming later this year, will use hand-tracking and eye-tracking for control, but at one time, Apple considered a finger-worn input device, according to Bloomberg’s Mark Gurman in today’s Power On newsletter.
Gurman said that early in the Vision Pro’s development cycle, Apple tested third-party virtual reality controllers from companies like HTC. Later, it looked into the finger-worn device — indeed, in 2015 a smart ring patent from the company emerged, though at the time seemed more intended as a general wearable device, not something specific to a mixed reality headset.
Apple Vision Pro first look.
In the end, Gurman says, the company decided that using eye-tracking and hand motions to control the Vision Pro was the “more elegant solution.” When our own Nilay Patel tested the headset, calling it the “best headset demo ever,” he thought the eye tracking was solid, and confirmed that you can keep your hands down at your sides or in your lap, tapping your thumb and forefinger together to make selections — no waving your hands in the air or pointing at things here.
Of course, Apple said at WWDC 2023 that the Vision Pro will still have support for external input devices, including third-party game controllers and keyboards, so you’ll still have the option to skip the hand controls for certain types of input.
Apple’s next AirPods Pro may check your hearing health and take your temperature
AirPods Pro may be getting new hearing health features. | Photo by Chris Welch / The Verge
Some AirPods will be gaining a new hearing health feature, supported by iOS 17, that can check yourself for potential hearing issues and may be able to determine your body temperature via your ear canal, according to Bloomberg’s Mark Gurman in today’s Power On newsletter. He also says all of Apple’s new headphones will include USB-C, and that the company is planning new AirPods Pro and AirPods Max models — but he doesn’t think new hardware is coming soon.
AirPods already support audiograms — audio profiles that tell the AirPods where your hearing may be weakest so that they can tune themselves to your hearing abilities. Right now, you can generate an audiogram using the app Mimi, which Apple would be Sherlocking — an infamous Apple tendency to fold third-party features and apps into its operating system — with a built-in hearing test feature.
Other wireless earbuds have similar capability built in as well. The Jabra Elite 75t added a feature in 2020 called MySound that creates custom sound profiles after playing a series of tones in each ear and prompting wearers to tap their screen when they hear a noise, and the Nothing Ear 2 launched with its own hearing test and audio profile feature this year.
Gurman says Apple is also exploring positioning the AirPods as hearing aids, now that the FDA has approved over-the-counter sales of hearing aids without a prescription.
Next, Gurman added that future AirPods could take your temperature with your ear canal. Rumors that Apple would do this go as far back as late 2021, and the Apple Watch Series 8 introduced something similar last year, albeit as a relative temperature that’s only really used for menstrual cycle tracking.
Expanding temperature tracking to the AirPods could be more accurate, he says, and would potentially let Apple more accurately tell a wearer if they’re starting to get sick. Gurman didn’t say whether this would be exclusive to the AirPods Pro, but it seems likely.
The bad news from the Power On newsletter is that the updated AirPods Pro might not be coming soon. Gurman believes the new health features are “several months or even years away,” and that although USB-C AirPods may be coming soon, AirPods Pro themselves are so far on a three-year refresh cycle. A sample size of two is hardly enough to say that’s how it will always be, but the fact that the 2nd gen AirPods Pro were released only last year doesn’t bode well for a new pair this year.
Then again, supply chain analyst Ming-Chi Kuo said in March that he expects a USB-C version of the AirPods Pro to come later this year, in the second or third quarter, and that Apple didn’t seem to have plans for either the original AirPods form factor or the more bulbous AirPods 3. It’s possible the new hardware could only be a new case with the updated charging port, similar to what Apple did when it introduced a standalone MagSafe case for the AirPods Pro in 2021.
Kuo also predicted in January that $99 AirPods and a new set of AirPods Max will come soon, as early as next year.
UpdateJuly 2nd, 2023 9:15AM ET:Added that other earbuds have hearing test features, such as the Jabra Elite 75t and the Nothing Ear 2.
New Uncensored Chatbots Ignite a Free-Speech Fracas A new generation of chatbots doesn’t have many of the guardrails put in place by companies like Google and OpenAI, presenting new possibilities — and risks.
Musk Says Twitter Is Limiting Number of Posts Users Can Read The change on Saturday came as thousands of users complained about getting an error message that they had “exceeded” their “rate limit.”
Valve won’t approve Steam games that use copyright-infringing AI artwork
Image: The Verge
On Thursday, various outlets reported on a June Reddit post from user potterharry97 saying that Steam would no longer be publishing games with AI-generated content. Later in the month, another game dev penned a similar post. Valve says that’s not quite right.
In a statement emailed to The Verge, Valve PR representative Kaci Boyle said the company’s goal is “not to discourage the use of [AI] on Steam; instead, we’re working through how to integrate it into our already-existing review policies.” She went on to say that the company’s current review process takes into account current copyright law, and that “while developers can use these AI technologies in their work... they can not infringe on existing copyrights.”
Boyle added that Steam will refund app-submission credits for any developers whose games were rejected over AI copyright issues as the company refines its review process. It made the same offer to potterharry97.
Valve’s developer submission rules disallow “content you don’t own or have adequate rights to.” Potterharry97 included the rejection message in their post, which said their game “contains art assets generated by artificial intelligence that appears to be relying on copyrighted material owned by third parties.”
The emergence of AI has added a new, complicated dimension to discussions around copyright. The change has lead to lawsuits over things like uncredited open source code usage and infringement of copyrighted artwork, and Valve heading this off with an aggressive policy surrounding AI-created assets isn’t too much of a surprise.
Smart lights, smart mugs, and a handful of other weekend discounts
Nanoleaf only manufactured a thousand units of its Ultra Black Triangles light panels, but they’re still available at some retailers. | Image: Nanoleaf
Whether you celebrate the “Fourth” or not, Independence Day weekend is oft considered one of the best times of the year to save on everything from TVs and laptops to the humble Toyota Camry. That’s still the case, however, with Amazon Prime Day kicking off in a little over a week, we expect many of this weekend’s best discounts to hang around just a bit longer than they might have otherwise.
Take our first deal of the day, as an example. Right now, you can pick up Nanoleaf’s Shapes Ultra Black Triangles at Amazon and Best Buy for $199.99 ($20 off), which matches the lowest price we’ve seen on the nine-panel kit since it made its debut last year. The modular light panels are nearly identical to the originals aside from the fact they appear all-black when dormant, meaning they can still display 16 million colors and work with Amazon Alexa, Apple Home, and Google Assistant. Plus, they’re outfitted with Wi-Fi and Thread radios, meaning they can also act as a Thread Border Router — a vital component of the new Matter smart home standard.
While I’d like to think most students aren’t thinking about going back to school quite yet — the school year only just ended in many parts of the country, after all — I’d be remiss if I didn’t point out that Dell’s XPS 13 is seeing one of its steepest discounts to date. Right now, the school-ready laptop is on sale at Dell for $849 ($250 off) with 16GB of RAM, a 12th Gen Intel Core i7-1250U processor, 512GB of storage, and a USB-C to 3.5mm adapter for those times when a shoddy pair of wireless headphones just won’t cut it.
As one of our top suggestions for both high school and college students, the XPS 13 has a lot going for it. The ultraportable Windows laptop remains one of the best alternatives to Apple’s MacBook Air thanks to its price-to-performance ratio, not to mention its gorgeous chassis, its 16:10 display, and the kind of lightweight build that makes it a dream to tote around from class to class. All the aforementioned specs also render it a great daily driver, even if you’re not someone who is planning on taking a freshman composition class come the fall.
Mosquitos are a seasonal fact of life, whether you live in the South or you’re a Northwest city dweller like me. Thankfully, Thermacell’s E90 Mosquito Repeller is nearly matching its best price to date at Amazon, where you can currently pick it up for $39.96 ($10 off).
The straightforward, rechargeable device is a welcome alternative to the Deet- and- butane-heavy products of yesteryear, and should allow you to thwart pesky bugs within a 20-foot radius for up to nine hours at a time thanks to its built-in battery. Keep in mind that the repeller comes with a single, nine-hour repellent cartridge, though there is a 40-hour cartridge available if you really need to keep ‘em at bay for longer.
A few more cool, summer savings
The Ember Mug 2 continues to be one of our guilty pleasures, one that’s on sale at Woot right now starting at $97.99 ($52 off). The app-controlled smart mug doesn’t do anything other than keep your drink warm, but there’s a small comfort in knowing your coffee is going to be the exact temperature you set it to any time you reach for it in the morning.
iOttie’s wireless car charger—one of our favorite car gadgets — is currently on sale at Amazon for $34.99 ($15 off), which is just $1 shy of its all-time low. Not only is the handy, Qi-equipped mount wide enough to accommodate pretty much any iPhone or Android phone on the market, but it also features a one-touch design that lets you easily snap your phone in place. It’s a steal of a price if you’re okay with basic charging speeds, especially since the few Apple-approved MagSafe mounts still cost north of $100. Ouch.
Anker’s Soundcore Liberty 4 NC are currently seeing their first discount, dropping them to just $79.99 ($20 off) at Amazon when you clip the on-page coupon or to the same price at Anker with offer code WSCPXLFXWN. Although we haven’t tested them yet, Anker’s newest earbuds look to be an excellent value thanks to features like adaptive noise cancellation and ear detection, the latter of which automatically pauses what’s playing when you take them out. We were big fans of the Liberty 3 Pro, and while I imagine the new earbuds might not be on par with the Pro model when it comes to sound, they still off a few features that are rare at this price point.
Apple’s ninth-gen iPad is on sale at Amazon at checkout for $249.99 ($79.01 off) in the 64GB base configuration, matching its lowest price to date. The entry-level model remains a great kid-friendly pick if you’re looking for something less expensive than the latest iPad Air or Pro models, especially since it still packs a 10.2-inch screen, a robust selection of apps, and an A13 Bionic chip that’s still plenty fast enough for most people’s needs. Read our iPad buying guide.
How Amazon Taught Alexa to Speak in an Irish Brogue For Alexa to speak like a Dubliner, Amazon researchers had to crack a problem that’s vexed data scientists for years: voice disentanglement.
A ‘Cage Match’ Between Elon Musk and Mark Zuckerberg May Be No Joke Talks over a matchup between the two tech billionaires have progressed and the parameters of an event are taking shape.
How Amazon Taught Alexa to Speak in an Irish Brogue For Alexa to speak like a Dubliner, Amazon researchers had to crack a problem that’s vexed data scientists for years: voice disentanglement.
The Reddit app-pocalyse is here: Apollo, Sync, and BaconReader go dark
Screenshot by Jay Peters / The Verge
After a month of outrage, protests, and unrest from the community, Reddit has finally flipped the switch to shut down some third-party apps.
Apollo, an iOS app that became a rallying point for the recent protests against Reddit’s imminent API pricing, no longer loads any content from the platform. When I open it up, all I see is a spinning wheel. Developer Christian Selig confirmed to me that Reddit is the one that turned things off, not him: “would have been nice to have been given a time,” he says in an email to The Verge.
Well, looks like Reddit pulled the plug a little early. Apollo started crashing, but I just manually revoked my token and it looks like it fixes the crashing, but no more Reddit access haha. Those folks are fun to the very end!
BaconReader, another popular app, shows an error message for me: “Request failed: client error (429).” When I tap the “Tap to refresh” link, I just get the same error message.
Sync, an Android app, has stopped working too, displaying this message: “Error loading page: 401.” We’ve additionally found a tweet showing an error and Lemmy comments about lack of functionality in a fourth app, reddit is fun (RIF), but at the time we published this article, one Verge staffer could still see content on the app when not logged in. He wasn’t able to log into his account, though.
Users were outraged at the company’s treatment of Selig and the developers of some other popular third-party apps, organizing protests to try and get Reddit to budge. But despite more than 8,000 communities going dark, Reddit held its ground, and now some apps are officially kaput. (Not every app is going away: Narwhal, Relay, and Now will still be available, though they will eventually become subscription-only.)
When reached for comment, Reddit spokesperson Tim Rathschmidt pointed to the company’s fact sheet about its API changes, which was just updated on Friday, as well as a Friday evening post from a Reddit admin confirming that the new API rate limits would be enforced “shortly.” (According to the fact sheet, the rate limits were technically supposed to go into effect on July 1st. I’m not sure what time zone Reddit was measuring that by, but if we’re basing it on US time zones, that means that Reddit decided enforce the limits a few hours ahead of when it said it would.)
This week, I asked Selig if he planned to still use Reddit after Apollo shuts down. “Honestly, not sure,” he said. “I’m certainly using it a lot less.”
The Game Boy that survived the Gulf War has been removed from Nintendo New York
Photo by Sean Hollister
Ever heard the urban legend about how the original Nintendo Game Boy survived a bomb? I have reason to believe that’s not true. But until recently, the flagship Nintendo Store at New York City’s Rockefeller Center housed an original Game Boy that, it claimed, was damaged in a bombing during Operation Desert Storm.
We just confirmed with Nintendo New York that, after many years on display, the Gulf War Game Boy is no longer there. VideoGameArt&Tidbits was the first to report the news; they say a worker told them it was returned to Nintendo’s US headquarters in Washington state.
If it’s true, and if it’s not coming back, we’re hoping that Nintendo will display it somewhere else. But just in case it doesn’t, here are five 4K images of the Gulf War Game Boy for posterity.
I shot these photos with a Pixel 3 when I visited the store in 2019. You can download them, blow them up, freely share them if you like (I didn’t shoot these ones for work.) Just link back if you do please?
Me, I figure this Game Boy probably survived because the back was mostly unscathed — and because, its original owner confirms to The Verge, it didn’t actually get hit by a bomb.
The Game Boy originally belonged to Stephan Scoggins, a ‘90s Nintendo Power reader who asked the Nintendo-owned magazine if he could get a new Game Boy in exchange. At the time, he simply said that it was “claimed by a fire while I was stationed in the Middle East,” when he was a registered nurse serving in Desert Storm.
Image: Nintendo Power
The Nintendo Power origin story,found at this link.
Here’s what Nintendo Power’s editors wrote in July 1991:
When we received Stephan‘s Game Boy from the Middle East, we thought that it was a goner. The back of the unit was in fair condition, but the front was charred and blistered from the heat of the fire. As an experiment, we popped in a Tetris Game Pak, plugged in a Battery Pak, and flipped on the power switch. When we heard its distinctive “Ping!” we couldn’t believe it! The Control Pad and A and B Buttons suffered melt down, but the Start and Select Buttons worked perfectly. Game Boy is even tougher than we thought it was! Of course, we don’t recommend that you subject your Game Boy to trial by fire, but in this case, we replaced Stephan’s Game Boy as a special “Desert Storm” courtesy.
Scoggins tells The Verge that yes, it was a fire. “It wasn’t a bombing, it was that the tent burned down.”
He suspects two different events were conflated. There was a bombing at that location, Scoggins says, but “it wasn’t one we were involved in.” We’ll be speaking to Scoggins more about his Game Boy soon.
Nintendo PR didn’t immediately have a comment on the Gulf War Game Boy. I think we can all agree that, bombing or no, it belongs in a museum.
Why influencers love a free trip — even a controversial one from Shein
Photo by Xavi Torrent / Getty Images
Thousands of influencers post content online with one goal: getting a brand deal. It’s affirmation that they’ve “made it” as an influencer, that they’re interesting enough to be paid to post, and that content creation could be a job. It’s usually a good thing — until, of course, it blows up in your face.
It’s not clear to me why anyone involved in Shein’s recent PR campaign thought it was a good idea to send influencers on a free trip to try to beat the labor exploitation accusations that plague the company. There’s an unseriousness to sponsored content — the cheery, upbeat music, the approval process videos go through before they can be posted — that makes it an insufficient response to workers who say they’re subjected to illegally long workdays and withheld wages.
But that didn’t stop Shein from tapping a handful of creators to visit Guangzhou, China, for a multi-day guided tour of Shein factories and facilities, fancy dinners, and photo ops.
“I expected the facility to be so filled with people just slaving away, but I was actually pleasantly surprised that most of these things were robotic,” one influencer who went on the trip said in a video. “Everyone was just working like normal, like chill, sitting down. They weren’t even sweating.”
Shein got what it wanted, but the influencers quickly realized this wasn’t a typical brand deal — followers and strangers alike were furious over content that seemed to brush past widely reported troubles with the brand. The backlash was swift, primarily directed at a creator that goes by Dani DMC, an influencer with nearly 300,000 TikTok followers whose videos were reshared on Twitter. In a matter of a few days, content from the Shein trip was deleted, defensive responses shared (and then also deleted), and apologies issued.
I was at a loss for words honestly they took her to the model unit equivalent of the sweatshops like this can’t be pic.twitter.com/aJtWcSycEh
— joe biden’s face lift (@thisisnefertiti) June 22, 2023
It should be yet another good lesson for anyone trying to make money through content creation: brand deals will sometimes come back to bite you. And it’s often the individual content creator — not the advertiser — that gets the most heat while having the least amount of support resources.
The incentives to make ads for brands have never been higher. Fueled by breakaway stars like MrBeast, the D’Amelio sisters, and Alix Earle, young people the world over don’t just dream of creating a viral presence out of nothing — many of them live it. Acting like an influencer is so easy it doesn’t even feel like pretending; shilling products and services online to a following of a few hundred has become a tenant of being online.
The work of converting content creators into an army of micro-advertising firms is becoming increasingly streamlined and frictionless. Platforms like Instagram offer creator marketplaces where brands can find influencers to hire for sponsored content. And on TikTok, a new program encourages creators to submit branded content for a chance at making some cash if their video performs well, without guaranteed returns.
Making a living online can be tricky — for most content creators making shortform videos, the payout from the platforms themselves is paltry. Earnings from creator funds, which pay viral personalities out of a pool of money set aside, often come out to be just a few dollars for millions of views. Other rewards programs for shortform video that were introduced to compete with TikTok have now dried up. Apart from potentially lucrative ad revenue sharing programs, many creators rely on brand deals to pay the bills.
The eagerness to make content for brands — and for brands to tap popular creators — has repeatedly backfired. Influencers boosting crypto projects made thousands of dollars, only for projects to turn out to be a scam. Even Kim Kardashian paid a $1.26 million fine after sharing sponcon for a crypto token without properly disclosing it was an ad.
Last year, TikTok and Instagram were littered with sponsored content made for a little-known app called Nate, which claimed to use artificial intelligence to autocomplete online shopping transactions.
Fashion and lifestyle influencers earned thousands of dollars in shopping credits by getting followers to sign up for the app. But the “AI” reportedly ended up just being human workers in the Philippines — users’ checkout information was manually entered by strangers. And in December, Nate ran off with influencers’ earnings, abruptly suspending its influencer program. Creators who had been using and promoting the Nate app aired out their frustrations and announced they would no longer be using the service — in the end, they lost out on what they were promised.
When brand deals go awry, it’s often the individual influencer who becomes the center of the maelstrom, as was the case with the Shein image rehab trip and others. Last month, a different influencer came under fire for referencing a school shooting that happened at her university in a sponsored video for skincare company Bioré. The brand apologized, too, saying it reviews all influencer content but doesn’t “edit or censor” material creators submit. I’m less outraged that a young person who experienced a campus shooting would mention it when asked to create content about mental health. But if you work in marketing for a major brand and don’t see how this could cause problems for everyone involved, you are bad at your job. Content creators are responsible for what they put their name on, but it’s up to the brand to make sure they don’t look like a fool.
Even worse is when a brand partners with an influencer and throws them to the wolves when there’s a public response, as was the case when Bud Light partnered with trans creator Dylan Mulvaney. When Mulvaney was subjected to an onslaught of transphobic vitriol and attacks, Bud Light doubled down, putting marketing executives on leave. As recent as this week, Mulvaney said the company had not even reached out to her since the abuse began.
In hindsight, hiring a smattering of influencers to convince the public you are definitely not violating labor laws was a bad idea. But if it wasn’t this group of people, it would have been another. And without the traditional stopgaps to tell creators how to navigate deals — editors, advisors, or someone to put out the fires — the rest of us will have to keep enduring the ill-conceived sponcon and resulting public outrage cycle.
Apple’s alien show Invasion returns for season 2 in August
Image: Apple
The summer of sci-fi continues over on Apple TV Plus. Just as season 1 of Silo ends and a new season of Foundation is about to begin, Apple has confirmed that its alien series Invasion will be back very soon. The first episode of a 10-episode-long season 2 will premiere on August 23rd, with new ones dropping on Wednesdays. As part of the announcement, we also got some early images of season 2 that include a cool spacesuit and what seems like a pretty war-torn version of Earth.
Invasion originally premiered in October 2021 and told a planet-wide story about an alien invasion, letting viewers see things unfold from multiple perspectives. It took a bit to get going but eventually ended up as a great start to a sci-fi story once it got around to the actual aliens. It also left a lot of questions unanswered. It sounds like the plan for season 2, as is often the case with sequels, is to go bigger, with the story picking up a few months after the season 1 finale.
“It’s a bigger, more intense season that drops our viewers into a wide-scale, global battle from the start,” co-creator and executive producer Simon Kinberg said in a statement. “At its core, the show is about the power of the human spirit and the emotional connections that hold us together especially when facing incredible obstacles.”
Who killed Google Reader?
Illustration by Hugo Herrera for The Verge
Ten years after its untimely death, the team that built the much-beloved feed reader reflects on what went wrong and what could have been.
There was a sign in the Google Reader team’s workspace at the company’s headquarters in Mountain View, California. “Days Since Cancellation,” it read, with a number below: zero. It was always zero.
This was in 2006 or so, back when Google Reader was still growing. Back when it still existed at all. Google’s feed-reading tool offered a powerful way to curate and read the internet and was beloved by its users. Reader launched in 2005, right as the blogging era went mainstream; it made a suddenly huge and sprawling web feel small and accessible and helped a generation of news obsessives and super-commenters feel like they weren’t missing anything. It wasn’t Google’s most popular app, not by a long shot, but it was one of its most beloved.
Within the company, though, Reader’s future always felt precarious. “It felt so incongruent,” says Dolapo Falola, a former engineer on the Reader team. “Literally, it felt like the entire time I was on the project, various people were trying to kill it.”
Of course, Google did kill it. (Google didn’t respond to a request for comment on this story.) Reader’s impending shutdown was announced in March of 2013, and the app went officially offline on July 1st of that year. “While the product has a loyal following, over the years usage has declined,” Google SVP Urs Hölzle wrote in a blog post announcing the shutdown.
Google tried its best to bury the announcement: it made it the fifth bullet in a series of otherwise mundane updates and published the blog post on the same day Pope Francis was elected to head the Catholic Church. Internally, says Mihai Parparita, who was one of Reader’s last engineers and caretakers, “they were like, ‘Okay, the Pope will be the big story of the day. It’ll be fine.’ But as it turns out, the people who care about Reader don’t really care about the Pope.” That loyal following Hölzle spoke of was irateoverlosingtheirfavorite web consumption tool.
Google’s bad reputation for killing and abandoning products started with Reader and has only gotten worse over time. But the real tragedy of Reader was that it had all the signs of being something big, and Google just couldn’t see it. Desperate to play catch-up to Facebook and Twitter, the company shut down one of its most prescient projects; you can see in Reader shades of everything from Twitter to the newsletter boom to the rising social web. To executives, Google Reader may have seemed like a humble feed aggregator built on boring technology. But for users, it was a way of organizing the internet, for making sense of the web, for collecting all the things you care about no matter its location or type, and helping you make the most of it.
A decade later, the people who worked on Reader still look back fondly on the project. It was a small group that built the app not because it was a flashy product or a savvy career move — it was decidedly neither — but because they loved trying to find better ways to curate and share the web. They fought through corporate politics and endless red tape just to make the thing they wanted to use. They found a way to make the web better, and all they wanted to do was keep it alive.
Photo by Chris Wetherell
From left to right: Ben Darnell, Chris Wetherell, and Laurence Gonsalves, three of the early members of the Reader team.
“I think I built a thing”
“This is going to be the driest story ever,” says Chris Wetherell, when I ask him to describe the beginning of Google Reader. Wetherell wasn’t the first person at Google to ever dream of a better way to read the internet, but he’s the one everyone credits with starting what became Reader. “Okay, here goes: a raging battle between feed formats,” he says when I push. “Does that sound interesting?”
Here’s the short version: One of the most important ways that information moves around the internet is via feeds, which automatically grab a webpage’s most important content and make it available. Feeds are what make podcasts work across apps, and how content shows up in everything from Flipboard to Facebook. In the early aughts, there were basically two ways to build a feed. One was RSS, which stands for Really Simple Syndication and has been around approximately forever. The other was called Atom, a newer standard that aimed to fix a lot of the things that were outdated and broken with RSS.
In late 2004, Jason Shellen, a product manager working on Atom projects at Google, called up Wetherell, a former colleague on the Blogger team, and asked him if he could hack together some kind of Atom-based app. “Is there any way you could write just a little thing that would parse Atom, just to show how it works?” Shellen asked. All he really needed was a tech demo, something he could show potential partners to explain how Atom worked.
Wetherell stayed up late one night building a simple app that converted a bunch of websites’ RSS feeds to Atom and displayed those feeds in a Javascript-based browser app so you could click around and read. “And then I tried to make it a pleasant arrangement,” Wetherell says. He called it Fusion. It wasn’t much to look at, but it was fast and worked in a web browser.
Photo by Chris Wetherell
Wetherell’s first prototype didn’t look like much, but it felt like nothing before.
And then the strangest thing happened: as soon as he’d finished the Fusion app, Wetherell started using it to actually read stuff from the sites whose feeds he’d grabbed. He turned to his partner that night and said: “I think I built a thing.” Wetherell sent the prototype to Shellen, who also immediately saw its potential.
In 2004, most people weren’t viewing the internet through a bunch of social networks and algorithmic feeds. Facebook and Twitter were barely blips on the radar. At that point, most people experienced the internet by typing in URLs and going to websites. A few tools like NetNewsWire and Bloglines had cropped up to make it easier to subscribe to lots of sites in one place, but these RSS readers were mostly tools for nerds. Most users were stuck managing bookmarks and browser windows and furiously refreshing their favorite sites just to see what was new. Wetherell’s prototype wasn’t complicated like NetNewsWire, it didn’t crash like Bloglines, and the Javascript interface felt fast and smooth. It immediately felt like a better way to keep up with the web.
Wetherell and Shellen started imagining all the different kinds of feeds this tool could store. He thought it might bring in photo streams from Flickr, videos from YouTube and Google Videos, even podcasts from around the web. Shellen, who had come to Google as part of its Blogger acquisition, saw the possibility for a social network, a single place to follow all your friends’ blogs. “Of course, it was just a hacky list of feeds,” Wetherell says, but there was something about the speed with which you could flip through articles and headlines, the information density, the simplicity of the reading experience, that just worked.
Ultimately, Wetherell ended up spending some of his 20 percent time — Google’s famous policy of letting employees work on just about whatever they wanted, which ironically died about the same time Reader did — building Fusion into a more complete feed-reading product. It handled RSS, Atom, and more. After a while, he wound up showing it to the folks building iGoogle, the company’s recently launched web-homepage product. (iGoogle has since been killed, of course.)
Image by Chris Wetherell
As the Fusion prototypes got more polish, they started to look more like Reader.
In his presentation, Wetherell shared a much bigger, grander ambition for Fusion than an article-reading service. He and Shellen had been talking about the fact that a feed could be, well, anything. Wetherell kept using the word “polymorphic,” a common term in programming circles that refers to a single thing having many forms.
“I drew a big circle on the whiteboard,” he recalls. “And I said, ‘This is information.’ And then I drew spokes off of it, saying, ‘These are videos. This is news. This is this and that.’” He told the iGoogle team that the future of information might be to turn everything into a feed and build a way to aggregate those feeds.
The pitch sounded good, and they got permission to keep working on it. Fusion wasn’t exactly made an official project or staffed like one, but it was at least allowed to continue to exist. Wetherell and Shellen recruited other people working on similar projects in their 20 percent time, and Shellen wrote an official product spec document outlining Fusion’s ambitions. The vision, he wrote, was to “become the world’s best collaborative and intelligent web content delivery service.” It promised to “build a robust web service and best-of-breed user interface for viewing subscriptions” and to produce an API that would let other apps tap into the same underlying data.
In other words, Fusion was meant to be a social network. One based on content, on curation, on discussion. In retrospect, what Shellen and Wetherell proposed sounds more like Twitter or Instagram than an RSS reader. “We were trying to avoid saying ‘feed reader,’” Shellen says, “or reading at all. Because I think we built a social product.”
That was the idea, anyway.
Google goes social
In October of 2005, Shellen announced Fusion to the world at the Web 2.0 Conference in San Francisco. Only he wasn’t allowed to call it Fusion. The team had been forced to change the name at the last minute, after Marissa Mayer — at that point, the Google executive in charge of all of the company’s consumer web services — said she wanted the name for another product and demanded the team pick another one. (That product never launched, and nobody I spoke to could even remember what it was. Mayer also didn’t respond to a request for comment.)
The team had brainstormed dozens of other names: Reactor, Transmogrifier, and a whiteboard full of others. Down near the bottom of the list: Reader, “a name none of us liked,” Wetherell says, “because it does many other things, but… it’s fine.” But somehow, that became the choice.
Shellen in particular still rues losing the fight over the name. Even now, he bristles thinking about the fight and the fact that Google Reader is known as “an RSS reader” and not the ultra-versatile information machine it could have become. Names matter, and Reader told everyone that it was for reading when it could have been for so much more. “If Google made the iPod,” he says, “they would have called it the Google Hardware MP3 Player For Music, you know?”
So Fusion launched, as Google Reader, and immediately crashed spectacularly. The site simply couldn’t keep up with the traffic on the first day. Most of those early visitors to reader.google.com never came back, either. Even once the Reader team stabilized the infrastructure, lots of users hated the product; it had a lot of clever UI tricks but just didn’t work for too many users. “People don’t remember this,” Wetherell says, “but it bombed. It was terrible. We were accused by someone of hurting the share price of Google because it bombed so hard.”
It wasn’t until the team launched a redesign in 2006 that added infinite scrolling, unread counts, and some better management tools for heavy readers that Reader took off. Another newish Google product, Gmail, had far more users, but the engagement with Reader was off the charts. “People would spend, I don’t know, five minutes a day on iGoogle,” Parparita says, “and like an hour a day in Reader.” The team hadn’t been pushed to worry about monetization or user growth, but they felt like they were on the right track.
Reader appealed primarily to information junkies, who wanted a quick way to keep up with all their favorite publications and blogs. (It turned out there were two types of Reader users: the completionists, who go through every unread item they have, and the folks who just scroll around until they find something. Both sides think the other is bonkers.) The team struggled to find ways to bring in more casual users, some of whom were put off by the idea of finding sites to subscribe to and others who simply didn’t care about reading hundreds of articles a day.
One feature took off immediately, for power users and casual readers alike: a simple sharing system that let users subscribe to see someone else’s starred items or share their collection of subscriptions with other people. The Reader team eventually built comments, a Share With Note feature, and more. All this now seems trite and obvious, of course, but at the time, a built-in way to see what your friends liked was novel and powerful. Reader was prescient.
Reader was always a power-user tool at heart, and it appealed to people with a lot to read.
At its peak, Reader had just north of 30 million users, many of them using it every day. That’s a big number — by almost any scale other than Google’s. Google scale projects are about hundreds of millions and billions of users, and executives always seemed to regard Reader as a rounding error. Internally, lots of workers used and loved it, but the company’s leadership began to wonder whether Reader was ever going to hit Google scale. Almost nothing ever hits Google scale, which is why Google kills almost everything.
The bigger problem seemed to be that Mayer didn’t like it: Shellen says she told him at one point that he was wasting his engineers’ careers working on Reader. The team had trouble getting face time in product reviews, and asking for additional resources or funding was a waste of time. Google co-founder Larry Page had been a fan of the app — Jenna Bilotta, a designer on the team, remembers he had this very specific idea about using Reader to research windmill-generated energy — but a few years later, Shellen recalls Reader appearing on Page’s list of Google’s worst 100 projects.
Google’s executives always seemed to think Reader was a feature, not a product. In meeting after meeting, they’d ask why Reader wasn’t just a tab in the Gmail app. When a team decided to build a new email client called Inbox, with promises of collecting all your important communication and information in one place, executives thought maybe Reader should be part of that. (Inbox was eventually killed, too.)
Every so often, a faction of the Reader team was called into a meeting and asked to justify the product’s ongoing existence. It didn’t require many resources, which was helpful; the team only ever got as big as about a dozen people, many of them on loan from other teams at the company. On the other hand, Reader wasn’t a roaring Google scale success, nor did it have a powerful executive championing its existence. It seemed the company got more tired of this side project all the time. Falola still remembers one particularly telling interaction: “We were having some back and forth with some VP at the time, making our petition for why you should keep Reader around, and I remember that VP responding with, ‘Don’t confuse this for a conversation between peers.’”
Threatened by the rise of social networks — namely Facebook and its quickly encroaching seizure of the online ad market — Google became desperate to build its own. It tried to build a social graph called Google Friend Connect, which went nowhere. It decided to build a network around email contacts, where the company already had a head start because of Gmail, but that didn’t make any sense. So the company’s big swing became Google Buzz, an app that tried to combine messaging, social networking, and blogging into one thing. That launched in 2010 and was killed in 2011.
For a while, the Reader team managed to stay alive by promising to be the guinea pig for Google’s other social ideas. It tried the Gmail contacts thing; Parparita remembers that as “the year Reader ruined Christmas” because the feature launched in December and suddenly everyone’s mom and landlord and Craigslist acquaintance could see all the articles they’d starred. (The team scrambled to build sharing management tools quickly after that.) The Reader engineers worked with the Buzz team, the iGoogle team… anyone who needed help.
The tide turned when Google decided not just to build a social product but to fundamentally re-architect the company’s apps around social. Two executives, Vic Gundotra and Bradley Horowitz, started a new project codenamed “Emerald Sea” with plans to build sharing and friend-based recommendations into just about every Google app. It would come to be known as Google Plus, the company’s most direct shot at a Facebook-like product, and Gundotra and Horowitz amassed an empire within the company. “We’re transforming Google itself into a social destination at a level and scale that we’ve never attempted — orders of magnitude more investment, in terms of people, than any previous project,” Gundotra told Wired in 2011. He wasn’t exaggerating.
“As far as I could tell, nobody ever won against them,” Parparita says. “They just got their own way.” There was plenty of opposition to the project, including from the Reader team, but it didn’t matter. The Emerald Sea team worked in a special building, only accessible to a few employees; the secrecy was just one more signal to everyone that this was Google’s top priority.
Gundotra and Horowitz also seemed to pluck any employee they wanted. And they wanted a number of Reader employees, who were some of Google’s most well regarded. “We assembled the Beatles,” says Wetherell, and Shellen calls the team a “Murderer’s Row.” Both singled out Paraprita as one of Google’s best engineers, along with Ben Darnell, a back-end whiz who built much of the product’s underlying infrastructure. Many of these engineers had started working on Reader as a side project, simply because they loved the app. Some had done stints full-time and then gone on to other projects. Now it felt like everyone was being pulled into Plus — and many of them chose to leave the company instead.
And in its effort to build a splashy new social platform, the Reader team felt Google was missing the burgeoning one right under its nose. Reader was probably never going to become the world-conquering beast Facebook eventually became, but the team felt it had figured out some things about how people actually want to connect. “There were people that met on Google Reader that got married,” Bilotta says. “There are whole communities that met on Google Reader that meet up — they fly to meet each other! It was crazy. We didn’t anticipate this being that sticky.” The team was plotting new ways for users to discover content, new tools for sharing, and more. Bilotta urged executives to see the potential: “They could have taken the resources that were allocated for Google Plus, invested them in Reader, and turned Reader into the amazing social network that it was starting to be.”
By early 2011, with the team severely diminished, Reader had been officially put into “maintenance mode,” which meant that an engineer — Parparita, mostly — would fix anything spectacularly broken but the product was otherwise to be left alone. Reader was integrated into Google Plus, sort of, before Plus began its inexorable decline. Despite Google practically force-feeding its social network to hundreds of millions of people, users rebelled against Google’s real-name policy, resented its spam problem, and ultimately could never figure out what Plus could do that Facebook or Twitter couldn’t. “The engagement was so low,” Bilotta says, “that basically within eight months, they realized it wasn’t going to be a product.”
The damage was done for Reader, though. Its core team was gone, its product had withered, and by the end of 2012, even Parparita had left Google. Hardly anyone on the team was surprised when Google announced a few months later that Reader was shutting down for good.
The alternate Reader universe
It’s been a decade since Reader went offline, and a number of the folks who helped build it still ask themselves questions about it. What if they’d focused on growth or revenue and really tried to get to Google scale? What if they’d pushed harder to support more media types, so it had more quickly become the reader / photo viewer / YouTube portal / podcast app they’d imagined? What if they’d convinced Mayer and the other executives that Reader wasn’t a threat to Google’s social plans, but actually could be Google’s social plans? What if it hadn’t been called Reader and hadn’t been pitched as a power-user RSS feed aggregator?
And, of course, there’s the biggest question: what if they’d tried to build Reader outside of Google? It had millions of devoted users, a top-notch team, and big plans. “At that time, outside of Google, VCs would have been throwing money at us left and right,” Wetherell says. Inside Google, it could never compete. Outside Google, there would have been no politics, no crushing weight of constant impending doom. If Google had been driven by anything other than sheer scale, Reader might have gotten to Google scale after all.
But Reader was also very much a product of Google’s infrastructure. Outside Google, there wouldn’t have been access to the company’s worldwide network of data centers, web crawlers, and excellent engineers. Reader existed and worked because of Google’s search stack, because of the work done by Blogger and Feedburner and others, and most of all, the work done by dozens of Google employees with 20 percent of their time to spare and some ideas about how to make Reader better. Sure, Google killed Reader. But nearly everyone I spoke to agreed that without Google, Reader could never have been as good as it was.
Over the years, people have approached Bilotta, Falola, and a few of the other ex-Reader team members about building something in the same vein. Shellen and Wetherell ended up co-founding Brizzly, a social platform based on a lot of the ideas in Reader. Kevin Systrom, once a product marketing manager on the Reader team, went on to found Instagram and, more recently, Artifact, two platforms with big ideas about information consumption that clearly learned from what went wrong at Reader.
For a while, the internet got away from what Google Reader was trying to build: everything moved into walled gardens and algorithmic feeds, governed by Facebook and Twitter and TikTok and others. But now, as that era ends and a new moment on the web is starting to take hold through Mastodon, Bluesky, and others, the things Reader wanted to be are beginning to come back. There are new ideas about how to consume lots of information; there’s a push toward content-centric networks rather than organizing everything around people. Most of all, users seem to want more control: more control over what they see, more knowledge about why they’re seeing it, and more ability to see the stuff they care about and get rid of the rest.
Google killed Reader before it had the chance to reach its full potential. But the folks who built it saw what it could be and still think it’s what the world needs. It was never just an RSS reader. “If they had invested in it,” says Bilotta, “if they had taken all those millions of dollars they used to build Google Plus and threw them into Reader, I think things would be quite different right now.”
Then she thinks about that for a second. “Maybe we still would have fallen into optimizing for the algorithm,” she allows. Then she thinks again. “But I don’t think so.”
European companies claim the EU’s AI Act could ‘jeopardise technological sovereignty’
The letter warns that the strict rules for generative AI outlined in the AI Act have ‘consequences.’ | Photo by Philipp von Ditfurth/picture alliance via Getty Images
Some of the biggest companies in Europe have taken collective action to criticize the European Union’s recently approved artificial intelligence regulations, claiming that the Artificial Intelligence Act is ineffective and could negatively impact competition. In an open letter sent to the European Parliament, Commission, and member states on Friday, and first seen by the Financial Times, over 150 executives from companies like Renault, Heineken, Airbus, and Siemens slammed the AI Act for its potential to “jeopardise Europe’s competitiveness and technological sovereignty.”
On June 14th, the European Parliament greenlit a draft of the AI Act following two years of developing its rules, and expanding them to encompass recent AI breakthroughs like large language AI models (LLMs) and foundation models, such as OpenAI’s GPT-4. There are still several phases remaining before the new law can take effect, with the remaining inter-institutional negotiations expected to end later this year.
With more than 150 european researchers, entrepreneurs, CEO... I am today signing an open letter to warn against the risks for our continent of the latest developments of the EU AI Act draft proposalhttps://t.co/FBZAliKkAs?
The signatories of the open letter claim that the AI Act in its current state may suppress the opportunity AI technology provides for Europe to “rejoin the technological avant-garde.” They argue that the approved rules are too extreme, and risk undermining the bloc’s technological ambitions instead of providing a suitable environment for AI innovation.
One of the major concerns flagged by the companies involve the legislation’s strict rules specifically targeting generative AI systems, a subset of AI models that typically fall under the “foundation model” designation. Under the AI Act, providers of foundation AI models — regardless of their intended application — will have to register their product with the EU, undergo risk assessments, and meet transparency requirements, such as having to publicly disclose any copyrighted data used to train their models.
The open letter claims that the companies developing these foundation AI systems would be subject to disproportionate compliance costs and liability risks, which may encourage AI providers to withdraw from the European market entirely. “Europe cannot afford to stay on the sidelines,” the letter said, encouraging EU lawmakers to drop its rigid compliance obligations for generative AI models and instead focus on those that can accommodate “broad principles in a risk-based approach.”
“We have come to the conclusion that the EU AI Act, in its current form, has catastrophic implications for European competitiveness,” said Jeannette zu Fürstenberg, founding partner of La Famiglia VC, and one of the signatories on the letter. “There is a strong spirit of innovation that is being unlocked in Europe right now, with key European talent leaving US companies to develop technology in Europe. Regulation that unfairly burdens young, innovative companies puts this spirit of innovation in jeopardy.”
The companies also called for the EU to form a regulatory body of experts within the AI industry to monitor how the AI Act can be applied as the technology continues to develop.
“It is a pity that the aggressive lobby of a few are capturing other serious companies,” said Dragoș Tudorache, a Member of the European Parliament who led the development of the AI Act, in response to the letter. Tudorache claims that the companies who have signed the letter are reacting “on the stimulus of a few,” and that the draft EU legislation provides “an industry-led process for defining standards, governance with industry at the table, and a light regulatory regime that asks for transparency. Nothing else.”
OpenAI, the company behind ChatGPT and Dall-E, lobbied the EU to change an earlier draft of the AI Act in 2022, requesting that lawmakers scrap a proposed amendment that would have subjected all providers of general-purpose AI systems — a vague, expansive category of AI that LLMs and foundation models can fall under — to the AI Act’s toughest restrictions. The amendment was ultimately never incorporated into the approved legislation.
OpenAI’s CEO Sam Altman, who himself signed an open letter warning of the potential dangers that future AI systems could pose, previously warned that the company could pull out of the European market if it was unable to comply with EU regulations. Altman later backtracked and said that OpenAI has “no plans to leave.”
Hun Sen’s Facebook Page Goes Dark After Spat with Meta Prime Minister Hun Sen, an avid user of the platform, had vowed to delete his account after Meta’s oversight board said he had used it to threaten political violence.
Is A.I. Poisoning Itself?, Billionaire Cage Fight and Cooking With ChatGPT The missing ingredient in A.I. content might soon be human-generated content.
/cdn.vox-cdn.com/uploads/chorus_asset/file/24764296/IMG_20190809_092407.jpg)
Photo by Sean Hollister
/cdn.vox-cdn.com/uploads/chorus_asset/file/24764297/IMG_20190809_092052__1_.jpg)
Photo by Sean Hollister
/cdn.vox-cdn.com/uploads/chorus_asset/file/24764295/IMG_20190809_092144.jpg)
Photo by Sean Hollister
/cdn.vox-cdn.com/uploads/chorus_asset/file/24764294/IMG_20190809_092125.jpg)
Photo by Sean Hollister
/cdn.vox-cdn.com/uploads/chorus_asset/file/24764293/IMG_20190809_092109.jpg)
Photo by Sean Hollister
/cdn.vox-cdn.com/uploads/chorus_asset/file/24764292/IMG_20190809_092046.jpg)
Photo by Sean Hollister
/cdn.vox-cdn.com/uploads/chorus_asset/file/24764325/origin__1_.jpg)
Image: Nintendo Power
/cdn.vox-cdn.com/uploads/chorus_asset/file/24760521/image_1.png)
Photo by Chris Wetherell
/cdn.vox-cdn.com/uploads/chorus_asset/file/24760514/1a._fusion_prototype___feed.png)
Photo by Chris Wetherell
/cdn.vox-cdn.com/uploads/chorus_asset/file/24760510/2d._fusion___tv_view.png)
Image by Chris Wetherell
/cdn.vox-cdn.com/assets/2147505/20130211-10455160-4-IMG_3044.JPG)
