Posts with «social & online media» label

Twitter’s recommendation algorithm is now on GitHub

Nearly a year after he first floated the idea of making Twitter’s recommendation algorithm public, the company has posted the source code for its recommendation algorithm on GitHub. In a Twitter Space discussing the move, Elon Musk said he hoped users would be able to find potential “issues” in the code and help make it better.

“Our initial release of the so-called algorithm is going to be quite embarrassing and people are gonna find a lot of mistakes but we're going to fix them very quickly,” Musk said.

Notably, the code released Friday only deals with how tweets are shown in Twitter's "For You" feed. The company didn't release the underlying code for its search algorithm or how content is displayed on other parts of Twitter, though Musk said the company would "for sure" open-source the search algorithm as well.

In a blog post outlining how Twitter’s recommendations work, the company explained the various steps of the algorithm, including ranking and filtering. But Twitter users have already been finding interesting details in the code itself. For example, Jane Manchun Wong noted that “Twitter’s algorithm specifically labels whether the Tweet author is Elon Musk.” That may offer yet another explanation for why Musk’s tweets appear so often. Wong also noted that the algorithm has labels indicating whether the tweet author is a “power user” as well as whether they are a Republican or Democrat.

Twitter’s algorithm specifically labels whether the Tweet author is Elon Musk

“author_is_elon”

besides the Democrat, Republican and “Power User” labelshttps://t.co/fhpBjdfifXpic.twitter.com/orCPvfMTb9

— Jane Manchun Wong (@wongmjane) March 31, 2023

When asked about that aspect of the algorithm in the Twitter Space, Musk said “I agree that shouldn’t be there … it definitely shouldn't be dividing people into Republicans and Democrats, that makes no sense.” A Twitter engineer followed up saying “it’s actually not too important.”

Developing...

This article originally appeared on Engadget at https://www.engadget.com/twitters-recommendation-algorithm-is-now-on-github-200511112.html?src=rss

Twitter announces new API pricing, including a limited free tier for bots

Twitter has finally confirmed some of the details and pricing for the new version of its API. The company had previously delayed the changes after confirming that it was banning third-party clients as part of a larger shakeup of its developer features.

As expected, the company is maintaining a free tier with limited functionality, though it offers far less than its predecessor. Under the new free tier, which is aimed at bots and other “testing” purposes, accounts can post up to 1,500 tweets a month, but won’t be able to access any other featuires. That may offer a lifeline to some of Twitter’s famed bot accounts, but at about 50 tweets a day, may prove to be too limited for those that post more frequently.

At $100 a month, the new “basic” tier offers a bit more: developers can post up to 3,000 tweets a month at the user level and up to 50,000 a month at the app level. It also offers a read limit of 10,000 tweets a month, which, again, is far less than what was previously offered.

Twitter

Meanwhile, an enterprise tier is meant for businesses that need a higher level of access, though details for that tier are still murky. According to Twitter’s developer website, the enterprise tier will include “commercial-level access that meets your and your customer's specific needs” and other features. Businesses can apply for enterprise access, but the only pricing information Twitter has disclosed is that there will be "monthly subscription tiers." 

It’s also unclear what will happen to researchers and academics who currently rely on Twitter’s API for their work. In a series of tweets, the company said it was “looking at new ways to continue serving this community” but didn’t elaborate. Wired previously reported the company had told some organizations API access could run as much as $42,000 a month, but that plan doesn't seem to have materialized, at least not yet.

The new details also mean that a lot of services using Twitter’s older APIs could soon stop working altogether. The company confirmed that its existing APIs, used by a vast number of developers, researchers and other services, would be deprecated within the next 30 days. “We recommend that you migrate to the new tiers as soon as possible for a smooth transition,” the company said. Though it’s unclear just how many developers will be willing to pay for stripped down versions of the APIs.

This article originally appeared on Engadget at https://www.engadget.com/twitter-announces-new-api-pricing-including-a-limited-free-tier-for-bots-005251253.html?src=rss

Twitter’s secret VIP list is the reason you see Elon Musk’s tweets so often

We now know why Twitter’s algorithm seems to recommend some users’ tweets so often. Newsletter Platformer reports that the company has a secret VIP list of a few dozen accounts “it monitors and offers increased visibility” in its recommendation algorithm. The accounts include Elon Musk, as well as a handful of other prominent Twitter users.

The revelation comes as Musk has repeatedlypromised to make Twitter’s recommendation algorithms open source. He also recently stated that soon only paid subscribers to Twitter Blue would be eligible to have their tweets displayed in the algorithmic “For You” feed. It’s not clear if that change would also affect the VIP list of users who regularly get a boost in the feed. Some of the accounts on the list, like President Joe Biden and YouTuber Mr Beast, are currently verified but don’t pay for Twitter Blue. Twitter no longer has a communications team to respond to questions.

According to Platformer, the VIP list was “originally created to monitor the engagement received by Twitter power users” as the company has tried to allay suspicions that its “shadowbans” certain users to reduce their visibility on the platform. “Our ‘algorithm’ is overly complex & not fully understood internally,” Musk recently tweeted.

But the list, which reportedly is only 35 accounts in addition to Musk, is also telling about who Musk believes should benefit from Twitter’s algorithm. Platformer didn’t publish the entire list, which includes LeBron James, New York Representative Alexandria Ocasio-Cortez, venture capitalist Marc Andreesen, Tesla fan account @teslaownerssv and conservative personality Ben Shapiro. It also includes @catturd2, an account recently described by Rolling Stone as the “Sh-tposting King of MAGA Twitter.”

It’s also not the first time that a report has surfaced about how Musk has tweaked Twitter’s algorithm to boost his own tweets. Last month, Musk said the company was fixing an issue that caused users’ For You feeds to be overwhelmed with dozens of Musk tweets. Platformer later reported that Twitter employees had changed its algorithm to favor Musk after the CEO was mad that his Super Bowl tweet didn’t receive as much engagement as Biden’s. Musk reportedly fired an engineer who suggested the lack of engagement was due to declining interest in his tweets.

This article originally appeared on Engadget at https://www.engadget.com/twitters-secret-vip-list-is-the-reason-you-see-elon-musks-tweets-so-often-181735784.html?src=rss

Google's new Classroom tools include a 'reader mode' for people with dyslexia

Google is making it easier for people with reading challenges, such as dyslexia, to be able to make out articles and text posts online. The tech giant has launched "reader mode" for Chrome, which takes a site's primary content and puts it into the sidebar to reduce clutter and distractions. Users will also be able to change the text's typeface, font size and spacing, as well as its color and background color, to find the combination that works best for them. 

Reader mode is but one of the new features and updates Google has rolled out for education users. Another new feature for Google Classroom gives educators the ability to add interactive questions to YouTube videos. That will allow students to answer them and get immediate feedback, giving teachers an insight on how well they understand the subject matter. 

Google is also giving teachers a way to share practice sets with other verified educators in their domain, so that they can expand the availability of materials their students have access to. For particularly difficult mathematical and scientific concepts, for instance, more examples mean more opportunity to better understand them. The company has released a new web player for Screencast on Chrome OS, as well, allowing users to watch casts in any browser on any platform. Plus, it has expanded language options for Screencast closed captions and for practice sets. 

Classes using Meet for online lectures will also find a new and useful feature: Hand raise gesture detection powered by AI. Apparently, when a student raises a hand in real life, the video conferencing app can now automatically activate its Hand Raise icon. In addition, two teachers can now also manage slides concurrently on Meet and co-present lectures together. Google has been growing and improving its education-related tools for years, though it has perhaps kicked things up a notch after schools shut down during the pandemic. It released a slew of updates to make virtual classrooms more usable since then, and it looks like it hasn't forgotten online-based education even though schools have mostly gone back to in-person learning. 

This article originally appeared on Engadget at https://www.engadget.com/google-new-classroom-tools-include-a-reader-mode-for-people-with-dyslexia-120046174.html?src=rss

Twitter will soon only recommend paid users in its For You feed

In a couple of weeks, all the tweets you see on the For You feed will come from Twitter Blue users. Twitter boss Elon Musk has announced that the website will only display tweets from verified accounts in For You's recommendations starting on April 15th going forward. 

Musk explained that the company is making the move to combat AI bots and that it's the "only realistic way" to prevent "advanced AI bot swarms" from taking over. He also said that Twitter is fighting a "hopeless losing battle" otherwise if it doesn't implement the change. In a follow-up tweet, Musk clarified that AI bots can pay for verification (and presumably be shown in the For You timeline), as long as they don't impersonate a human. 

Twitter's For You interface shows not just tweets from the accounts a user follows, but also content based on previous tweets they've liked or other people in their network have liked. The feature debuted on iOS back in January as an alternative to the Following tab before making its way to other platforms. Not everybody liked it, and Twitter released an update so that its web and mobile apps will default to the timeline tab users last had open after it came out. But displaying only verified users could potentially mean that those who choose to view it would see a lot more promotional tweets. 

After all, pretty soon the only verified accounts on the social media site will be Twitter Blue subscribers. Twitter previously announced that it's killing legacy checkmarks on April 1st, stripping public personalities, companies, government and non-government organizations and the like of their verified status unless they pay for it.

Starting April 15th, only verified accounts will be eligible to be in For You recommendations.

The is the only realistic way to address advanced AI bot swarms taking over. It is otherwise a hopeless losing battle.

Voting in polls will require verification for same reason.

— Elon Musk (@elonmusk) March 27, 2023

This article originally appeared on Engadget at https://www.engadget.com/twitter-will-soon-only-recommend-paid-users-in-its-for-you-feed-041101215.html?src=rss

Portions of Twitter's source code were reportedly leaked online

In yet another embarrassing development for new Twitter boss Elon Musk, court filings published Friday reveal that portions of the social media site's source code — the base programming that makes Twitter possible — have been leaked online, the New York Times reports.

Per court filings, Twitter claimed copyright infringement in an effort to have the offending code taken down from the Github collaborative programming network, where it had been posted. While the code was removed the same day, details as to how long the code had been left up were not made available, nor were the leak's scope or depth. As part of the takedown request reminiscent of Raytheon's famous -- failed -- attempt at court-sanctioned doxxing, Twitter also asked the US District Court for the Northern District of California to order Github to reveal both the identity of the user who posted the code and those who accessed and downloaded it. 

The NYT reports that, per sources within the company privy to the internal investigation into the leak, Twitter executives strongly suspect it's the work of a disgruntled employee who had left, "within the last year." Coincidentally, Elon Musk purchased Twitter last October for the eye-watering price of $44 billion and proceeded to lay-off and otherwise lose 80 percent of the company's staff, not the 75 percent that everybody feared Musk would enact in the run-up to his purchase.  

The executive who spoke with the NYT are primarily concerned that revelations gleaned from the stolen code could empower future hacking efforts, either by revealing new exploits or allowing bad actors to access Twitter user data. If the increasingly temperamental page functionality wasn't enough to send the site's user base running for the hills that the site's resurgence of scammers and white nationalists since Elon's takeover didn't already scare off, will the threat of outright hacking be the final straw for advertisers and users alike? 

This article originally appeared on Engadget at https://www.engadget.com/portions-of-twitters-source-code-have-reportedly-leaked-online-234405620.html?src=rss

Internet Archive violated publisher copyrights by lending ebooks, court rules

A federal judge has ruled against the Internet Archive in its high-profile case against a group of four US publishers led by Hachette Book Group. Per Reuters, Judge John G. Koeltl declared on Friday the nonprofit had infringed on the group’s copyrights by lending out digitally scanned copies of their books.

The lawsuit originated from the Internet Archive’s decision to launch the “National Emergency Library” during the early days of the pandemic. The program saw the organization offer more than 1.4 million free ebooks, including copyrighted works, in response to libraries worldwide closing their doors due to coronavirus lockdown measures.

Before March 2020, the Internet Archive’s Open Library program operated under what’s known as a “controlled digital lending” system, meaning there was often a waitlist to borrow a book from its collection. When the pandemic hit, the Internet Archive lifted those restrictions to make it easier for people to access reading material while stuck at home. The Copyright Alliance was quick to take issue with the effort. And in June 2020, Hachette, as well as HarperCollins, Penguin Random House and John Wiley & Sons, sued The Internet Archive, accusing the organization of enabling “willful mass copyright infringement.” That same month, the Internet Archive shuttered the National Emergency Program early.

Going into this week’s trial, the Internet Archive argued the initiative was protected by the principle of Fair Use, which allows the unlicensed use of copyrighted works under some circumstances. As The Verge notes, HathiTrust, an offshoot of the Google Books Search project, successfully used a similar argument in 2014 to fend off a legal challenge from The Authors Guild. However, Judge Koeltl rejected the Internet Archive’s stance, declaring “there is nothing transformative” about lending unauthorized copies of books. "Although [the Internet Archive] has the right to lend print books it lawfully acquired, it does not have the right to scan those books and lend the digital copies en masse," he wrote. Maria Pallante, the president and CEO of the Association of American Publishers, said the ruling “underscored the importance of authors, publishers, and creative markets in a global society."

On Saturday, the Internet Archive said it would appeal the decision. “Libraries are more than the customer service departments for corporate database products. For democracy to thrive at global scale, libraries must be able to sustain their historic role in society — owning, preserving, and lending book,” the nonprofit wrote in a blog post. “This ruling is a blow for libraries, readers, and authors and we plan to appeal it.”

This article originally appeared on Engadget at https://www.engadget.com/internet-archive-violated-publisher-copyrights-by-lending-ebooks-court-rules-164629790.html?src=rss

Twitter's bookmark counter is a new way to convince yourself tweeting is worth your time

If you take pride in people bookmarking your social media posts, Twitter is ready to feed your ego. The social network has added a bookmark counter to its iOS app. Similar to the counts for likes, quotes and retweets, you'll know just how many people saved a given tweet for later reference. The company hasn't mentioned when the feature might come to Android or the web.

Twitter stresses that bookmarks are still private. Users may know if their tweets are popular, but they won't know just who bookmarked a given post.

We love Bookmarks for saving Tweets to revisit later. Starting today on iOS, you’ll now see the total number of times a Tweet has been bookmarked on Tweet details.

— Twitter Support (@TwitterSupport) March 16, 2023

This counter isn't just useful for bragging rights, of course. If you're a creator or run a company's Twitter account, the tally could show whether or not followers find your posts useful. That, in turn, may help you refine content to reach a wider audience.

The addition comes after a handful of upgrades to Twitter, such as alerts about community misinformation notes. However, the service has lately been dominated by concerns about a string of outages and a paywall for text-based two-factor authentication. However useful the bookmark counter may be, it's not going to garner as much attention as it might have in the past.

This article originally appeared on Engadget at https://www.engadget.com/twitters-bookmark-counter-is-a-new-way-to-convince-yourself-tweeting-is-worth-your-time-210005962.html?src=rss

You can now ‘enhance’ your LinkedIn Profile with AI-written 'suggestions'

LinkedIn is the latest platform to hop on the generative AI bandwagon. The company is adding AI-powered “writing suggestions” and job descriptions to its service as it looks for new ways to infuse AI into its platform.

The writing suggestions are meant to make it easier to fill out key profile fields that LinkedIn says can otherwise feel “daunting” to complete: the “about” and “headline” sections near the top of each profile. Now, with the new “enhance” tool, LinkedIn Premium subscribers can generate descriptions based on their experience.

The company says the tool, which uses the same OpenAI models that power ChatGPT, is meant to preserve “your unique voice and style” and will draw from your job experience and skills, as well as LinkedIn’s own “insights” into what makes a good profile. In an example of a completed “about” section provided by LinkedIn, the tool generated a first-person summary of an individual’s job experience that reads almost like the beginning of a cover letter.

LinkedIn

LinkedIn also says it’s starting to test AI-written job descriptions. In those cases, hiring managers will simply need to fill out the job title, company name and a few other basic details, and LinkedIn will create a detailed draft of a relevant job description.

Of note, the company is positioning its AI writing features as more of a starting point than a final product. In both cases, LinkedIn says that users should review and edit the AI-generated text to check for accuracy. But the company says that both could be a major time saver for members who want to offload some of the more tedious writing tasks associated with LinkedIn.

These writing tools aren’t LinkedIn’s first foray into generative AI. The company also recently introduced “collaborative articles,” which rely on a combination of AI-written text as well as contributions from individual LinkedIn “experts.” Elsewhere, the company is also adding new online courses dedicated to generative AI-related topics.

This article originally appeared on Engadget at https://www.engadget.com/you-can-now-enhance-your-linkedin-profile-with-ai-written-suggestions-160054549.html?src=rss

Mozilla brings its cookie protection tool to Firefox for Android

How many times have you experienced this scenario? You go shopping online for a new shirt, speaker or other item. You click on the same one a few times and then decide against it. Suddenly every website you visit has an ad featuring that item, imploring you to reconsider. In an effort to create greater privacy online, Mozilla is now rolling out Total Cookie Protection (TCP) as the default setting on the Firefox app for Android after initially making it available for Firefox users on Windows, Mac, and Linux. 

So, what does TCP do? TCP ensures that your cookies aren't being shared across sites. Typically, third-party cookies collect information about you from across the internet to build your virtual identity. Data brokers then sell your information to businesses that will provide you with targeted ads.

Instead, with TCP, the cookies you create while browsing only belong to the site you're on. This feature limits companies from learning any information you enter or behavior you exhibit anywhere else on the internet. So, a store may know you looked for a blue laptop case in their online shop, but they won't know you also searched for size 11 shoes in another one. 

As part of today's Android update, Mozilla says it's also promoting its Firefox Relay protection to become a dedicated part of the app. Relay, which initially launched as an addon and provides users with email and phone number masks for online signups, comes with a limited free tier but requires a subscription to get the most from the service.

This article originally appeared on Engadget at https://www.engadget.com/mozilla-cookie-protection-tool-firefox-android-130003579.html?src=rss