Platformer has interviewed Twitter employees to figure out why people's For You feeds turned into The Elon Musk show, and they said the company's engineers truly did build a system that benefited their CEO alone. According to the publication, Elon's cousin and Twitter employee James Musk sent an urgent message on the company's Slack on Monday morning. "We are debugging an issue with engagement across the platform," he wrote, calling the situation "high urgency" and asking everyone who can write code to help. The situation? President Biden's tweet about rooting for the Philadelphia Eagles got more engagement than Musk's. 

Apparently, the president's tweet generated almost 29 million impressions, whereas Musk's generated 9.1 million only before he deleted it. Musk's issue with his tweets not getting as much engagement as he would like started before the Super Bowl, however. Platformer previously reported that he fired one of Twitter's two remaining principal engineers because he suggested that Musk's tweets aren't generating as many impressions anymore because people are no longer that interested in what he's saying. 

Around 80 Twitter employees were reportedly tasked to investigate the possible reasons for Musk's waning engagement numbers, and they did consider the possibility that a lot of people had blocked and muted him in recent months. They also looked into legitimate potential technical issues, though, because tweets from users whose posts typically perform well should've been automatically promoted by the website's system. 

The fix they came up with, Platformer says, is to deploy code that would artificially boost Musk's tweets by a factor of 1,000, ensuring that they rank higher than everyone else's in people's feeds. As a result, over 90 percent of Musk's 128.9 million followers saw his tweets, and even those who don't follow the Twitter owner kept seeing his posts on their timeline. 

While Musk didn't speak at length about the issue, he did acknowledge the change in his own way, specifically by posting the "forced to drink milk" meme. He also asked people to stay tuned while Twitter made adjustments to the "algorithm" (his quote, not ours) after users started complaining of seeing all Elon all the time. "Adjustments" reportedly didn't mean removing his advantage completely, though: according to Platformer, Elon's tweets are still artificially boosted, but by a factor less than 1,000. Bottom line, users will still see his tweets a lot, unless they choose to move over to the Following tab instead. 

Starting today, Meta is rolling out a new version of its “Why am I seeing this ad?” tool. The company says the redesigned interface is meant to provide users with more information about how their activities on Facebook and beyond inform the machine learning models that power its ad-matching software. If you’re unfamiliar with the tool, you can access it by clicking or tapping the three dots icon next to an ad on Facebook or Instagram.

Once you have access to the updated tool, you’ll see a summary of the actions on Meta’s platforms and other websites that may have informed the company’s machine-learning models. For instance, the page may note that you’re seeing an ad for a dress or suit because you interacted with style content on Facebook. Users will also see new examples and illustrations that attempt to explain how Meta’s machine learning algorithms work to deliver targeted ads. At the same time, the company says it has made it easier to access its Ads Preferences. You’ll see a link to those settings from more pages accessible through the “Why am I seeing this ad?” tool.

Meta

“We are committed to using machine learning models responsibly. Being transparent about how we use machine learning is essential because it ensures that people are aware that this technology is a part of our ads system and that they know the types of information it is using,” Meta said in a blog post published Tuesday. “By stepping up our transparency around how our machine learning models work to deliver ads, we aim to help people feel more secure and increase our accountability.”

The company notes it worked with “external privacy experts and policy stakeholders” to collect input on how it could be more transparent about its ads system. Meta doesn’t say as much, but the changes likely represent an effort to ensure the company is compliant with the European Union’s Digital Services Act when it becomes law in 2024. The legislation has several provisions that apply directly to Meta, including one that mandates more transparency around how recommendation systems work. The law will also ban ads that target individuals based on their religion, sexual orientation, ethnicity or political affiliation.

More broadly, the changes come after Apple’s ad-tracking changes in iOS 14 significantly hurt Meta’s bottom line. One early report after iOS 14.5 went live estimated only four percent of iPhone users in the US opted into app tracking. Since then, Meta has seen revenue growth shrink significantly. More recently, in combination with its virtual and augmented reality spending, Meta saw its first-ever revenue decline in the second quarter of 2022.

Twitter is making millions of dollars from just a handful of some of its most infamous users, according to a new report. New research from the Center for Countering Digital Hate (CCDH) estimates that Twitter “will generate up to $19 million a year in advertising revenue” from just 10 accounts that were once banned from the platform.

The report looked at the current engagement with 10 accounts that were previously banned for “ for “publishing hateful content and dangerous conspiracies.” The accounts were reinstated after Elon Musk’s takeover of Twitter. The group includes a number of high-profile accounts associated with extremism and conspiracy theories, including those belonging to influencer Andrew Tate, Daily Stormer founder Andrew Anglin, prominent antivaxxer Robert Malone and the Gateway Pundit.

In order to estimate their reach and engagement, CCDH analyzed nearly 10,000 tweets from these accounts during a 47-day period in December and January. According to their analysis, “on an average day, tweets from the ten accounts received a combined total of 54 million impressions,” they write. “Projecting this average across 365 days, the accounts can be expected to reach nearly 20 billion impressions over the course of a year.”

In order to determine how much ad revenue those impressions might generate for Twitter, CCDH says it created three new Twitter accounts that followed only the 10 users named in the report. The authors found that ads appeared about once every 6.7 tweets. Then, using data from analytics firm Brandwatch, which estimates that “Twitter ads cost an average of $6.46 per 1,000 impressions,” CCDH came up with “a total figure of up to $19 million in estimated annual ad revenues across the accounts.”

While the estimates aren’t a precise accounting of how much Twitter might be making from these users, it demonstrates how valuable a small number of highly polarizing accounts can be for the platform. It also underscores how much more Twitter stands to gain by bringing back even more controversial users.

All of the accounts named in the report were once permanently banned from twitter, but were reinstated after Musk said he would offer “general amnesty” to users who hadn’t broken the law. Twitter also recently announced plans to allow even more previously banned users to appeal their suspensions.

At the same time, Twitter’s advertising business has taken a major hit since Musk’s takeover. A number of high profile advertisers have pulled back from the platform, and revenue is down as much as 40 percent, according to reporting fromPlatformer.

The report also points out several instances when ads from prominent advertisers appeared adjacent to offensive and inflammatory posts from these users. For example, a Prime Video ad directly underneath a tweet from Andrew Anglin that states “the only career a woman is actually capable of on merit is prostitution.” The report also highlights an ad from the NFL, which appeared directly underneath a tweet misinformation about COVID-19 vaccines.

“This work confirms that Twitter has been displaying ads next to every one of the toxic accounts we have investigated, despite the fact that the individuals behind them are known to promote hateful views and falsehoods,” CCDH writes.

Twitter has shared more details about the upcoming changes to its API that will require most developers to pay in order to keep using its developer tools. In an update, the company said that there will be “a new form of free access” that will allow “Tweet creation of up to 1,500 Tweets per month.”

That clarification means that many of Twitter’s so-called “good” bots — the automated accounts that tweet everything from historical photos to helpful reminders — will be able to continue on the platform. Previously, the future of these accounts was uncertain as many bot makers said they would not pay for API access.

However, the 1,500 monthly limit on tweet creation could still impact bots that tweet most frequently. Fifteen hundred tweets a month works out to roughly 50 tweets a day, which could be problematic for Twitter’s most active bots. The company also granted a short reprieve for developers about to lose API access, saying that it had extended the current version of the API until February 13th.

After that, developers wishing to access the “paid basic” tier of Twitter’s API will need to pay $100 a month, according to Twitter. The company also confirmed that it would be ending the Premium API, and that subscribers would have the option to apply for the Enterprise version of the service. The company still hasn’t weighed in on what, if any, options will be available to researchers currently using Twitter’s developer tools for their projects.

Something is very wrong with Twitter right now, but it’s not clear exactly what. As of Wednesday afternoon, core features, including direct messages and tweets, are not working for large swaths of users.

While some users are able to tweet, many users are receiving messages that they can’t tweet or as they are “over the daily limit for sending Tweets.” A similar message appears when attempting retweets: “Sorry! You’ve have exceeded your Tweet limit. Try retweet again tomorrow.” While rate limiting can in some cases be an anti-spam tactic, the messages are appearing even for accounts that have shared relatively few tweets.

Screenshot via Twitter

In my case, I’ve tweeted less than two dozen times today, and I’m receiving the errors. According to Twitter, the “technical limit” for accounts is 2,400 tweets and 500 direct messages a day. For now, using Twitter's scheduling function appears to bypass the issue. 

Additionally, direct messages are down for many users. Instead of the normal inbox view, it simply says “Something went wrong. Try reloading.” Previously sent messages are currently inaccessible.

It’s not clear what the source of these issues are. Twitter, which no longer has a communications team, didn’t immediately respond to a request for comment.

Developing…

You now have a very practical reason to subscribe to Twitter Blue beyond long video uploads and the coveted blue checkmark. As of today, Blue members in the US can post tweets up to 4,000 characters long. You can't currently schedule these posts or save them as drafts, but this could be helpful if you'd rather not write a lengthy multi-tweet thread just to share your thoughts.

Free Twitter users can still reply to and quote these tweets. And yes, Twitter is aware that scrolling through a sea of these lengthy posts could be frustrating. Any tweet longer than 280 characters now hides any extra content beyond a "show more" prompt. Blue normally costs $8 per month, or $84 per year.

This probably won't prompt you to switch from social networks like Facebook or Mastodon, where long posts have been available for a while. However, it could be helpful if you'd like to share the same post across multiple social platforms without having to split it up or write a condensed version.

There's certainly pressure on Twitter to make Blue more appealing. Twitter reportedly only had 180,000 Blue subscribers in the US as of mid-January, and roughly 290,000 worldwide. While the revamped subscription hasn't been available for very long, it needs to be highly attractive to enthusiasts if Elon Musk wants to achieve a goal of generating half of Twitter's revenue from subscriptions.

Turkey may be blocking access to Twitter, two days after a pair of catastrophic earthquakes struck the area. Thousands of people are still trapped in buildings in Turkey and Syria, where the death toll is approaching 12,000.

According to Bloomberg, people in Turkey started having trouble accessing Twitter on Wednesday afternoon. Some have resorted to VPNs to use the service. Kemal Kilicdaroglu, the leader of Turkey's main opposition party, has accused the government of blocking Twitter.

It's not clear why the Turkish government might want to prevent access to Twitter amid such devastation. The social media service is still a valuable disaster response tool and users have also been sharing images of the destruction caused by the earthquakes. Twitter does not have a communications team that can be reached for comment.

This would not be the first time that Turkey has stopped residents from accessing social media services. It has also done so during cross-border military operations and terror attacks. In 2014, Turkey temporarily banned Twitter. Users were sharing voice recordings and documents that purportedly showed corruption within then-prime minister Recep Tayyip Erdoğan's sphere of influence. Erdoğan became Turkey's president later that year and he remains in power. His government has faced criticism for its response to this week's disaster.

Of all the once-unthinkable changes Elon Musk had made since taking over Twitter, pulling the rug out from under developers might seem relatively minor. After banning third-party clients without warning, Twitter announced that it would no longer allow any developer to use its APIs for free.

So far, Twitter has communicated very little about the changes, other than confirming a February 9th cut-off date. Musk has suggested Twitter could charge $100 a month “with ID verification,” but hasn’t elaborated. What we do know, is that once free access is shut off, thousands of apps, research projects, bots and other services will stop functioning (or, at the very least, be interrupted). If you’re a Twitter user, chances are this will affect you in some way, and you shouldn’t wait until it’s too late to prepare.

Change logins for apps where you’ve signed in with Twitter

If you’ve ever used your Twitter credentials to sign into another service, the coming API shutoff could prevent you from being able to access your account. That means you’ll want to double check which services you’ve used Twitter as a login for, and change your account info while you’re still able to.

You can check which services have access to your Twitter account via Settings -> Security and Account Access -> Apps and sessions -> Connected Apps (or, clicking here). Depending on the service, you may have to create an entirely new account with different credentials.

Nuke your tweets

There are a lot of good reasons to delete your old tweets. But if you haven’t already, your window for easily doing so is rapidly closing. If you want to wipe your tweet history, check out TweetDelete, TwitWipe, TweetDeleter, or another similar service. Note that there are a handful of subscription-based tweet deletion services. It may be a good idea to hold off on using a paid service for now until it’s clear whether any of these apps will continue to operate after the API changes go into effect.

Find your mutuals on Mastodon

Whether you’re new to the Fediverse or not, now would be a very good time to find all your mutuals from Twitter over on Mastodon. Services like Movetodon and Twitodon enable you to easily find familiar accounts on Mastodon. Rebuilding your following graph on a new platform can be one of the biggest hurdles to getting started, so even if you’re not sure about switching, it’s not a bad idea to use these tools while you still can.

Everything else

Unfortunately, the effects of ending the free API will be much more far-reaching than just these services. The future of Twitter’s bots — the automated accounts that share everything from weather updates, helpful reminders, jokes and emergency alerts — is now very much uncertain. While Musk has indicated that he may allow some bots with “good” content to access the Twitter API for free, it’s not at all clear how this will work, or if paying for Twitter Blue will be a requirement. So far, a number of bot developers have said they will not pay for use of Twitter's API and are planning for their accounts to go dark on the 9th.

The API changes will also have a disastrous impact on the research community. Currently, there are scores of journalists, nonprofits and academic institutions that use Twitter’s API to research misinformation, public health, education and numerous other issues. Requiring these researchers to pay could effectively end these projects, particularly for organizations that are smaller and not as well-funded. For now, it's not clear if these organizations will have any other options available to them. 

Elon Musk may have to think of a lot more ways to make Twitter Blue appealing to potential subscribers if he wants the subscription service to be a major source of revenue. According to The Information, only 180,000 people in the US have been paying for a Twitter subscription by mid-January, and that's apparently around 0.2 percent of the website's monthly active users. The publication said it saw the information on a document, which also revealed that 62 percent of the company's paying users is from the US. That means Twitter has approximately 290,000 subscribers worldwide. 

Twitter Blue costs $8 a month for users who pay via web — or $7, if they pay for an annual subscription — and $11 for those who pay via Apple's or Google's app stores. Since the latter option gives the tech giants a cut of subscribers' payments, Twitter still only gets $8 a month from users overall. With the current number of paying users, the website is only set to earn $27.8 million a year from its subscription services. That said, Twitter has only just relaunched Blue in mid-December last year after a bumpy initial launch a month before that. It's bound to rack up more subscribers, though it remains to be seen if it can achieve the level of growth Musk wants to see. 

As The Information notes, Musk told Twitter employees last year that he wants half of the website's revenue to come from subscriptions. Since the company has to pay over $1 billion a year in interests alone from the loans Musk took out when he purchased the website, the executive is aiming to take in a revenue of $3 billion for 2023. Twitter has to have quite a lot of subscribers to earn half of that from Blue. One avenue the company is considering to earn more from its subscription services is to offer a higher-priced membership tier that allows users to browse the website with zero ads. Twitter is also reportedly planning to charge businesses $1,000 a month for their gold verification badges and an extra $50 a month for each account affiliated with them. 

People in Pakistan can once again use Wikipedia, three days after the country blocked the website over content that regulators deemed "sacrilegious." As TechCrunch notes, prime minister Shehbaz Sharif ordered officials to unblock Wikipedia after determining that the ban was “not a suitable measure to restrict access to some objectionable contents/sacrilegious matter on it.” Sharif's office said in a statement that the "unintended consequences of this blanket ban" outweighed the "benefits."

The Pakistan Telecommunication Authority (PTA) "degraded" access to Wikipedia last Wednesday, warning the site's operators that they had 48 hours to remove certain content or face repercussions. Wikipedia apparently demurred and the PTA blocked the site in Pakistan on Friday.

Meanwhile, Sharif is establishing a cabinet committee to "explore and recommend alternative technical measures for removal or blocking access to objectionable content posted on Wikipedia and other online information sites, in view of our social, cultural and religious sensitivities, on the touchstone of proportionality." The committee is also being tasked with offering other suggestions aimed at "controlling unlawful content in a balanced manner." 

Sharif asked the committee to provide the cabinet with recommendations within one week. That doesn't give the committee members much time to fully assess and analyze the many, many considerations that go into content moderation.