Elon Musk says he’ll open-source Twitter’s algorithm. Here’s why that would change everything

There would be three major effects of open-sourcing Twitter’s algorithm. The first would be political: Allowing outsiders to poke around the code should settle any questions about alleged ideological bias within the company’s algorithm itself. 

That’s not quite the same issue as the historical liberal bias that was supposedly exposed in the “Twitter Files” episode, in which Musk granted a few journalists access to Twitter’s pre-Musk internal communications and let them craft a selective account of what went down in terms of human content-moderation decisions. But with the algorithm, everyone would get equal access, so it could prove more genuinely illuminating.

Secondly, if Musk really does mean “open source” in the traditional sense, Twitter’s algorithm would become available for modification and redistribution, by anyone. I still find it hard to believe he’s preparing to let people basically copy Twitter, but let’s see what the fine print says. (Open-source Twitter rival Mastodon must be watching carefully.)

The third effect would be the most far-reaching: The world would get to see what happens when Big Tech opens up one of its black boxes.

This could go horribly wrong. It’s possible that, when shown how Twitter’s systems make decisions, miscreants could swiftly learn how to game the system and inundate Twitter with (even more) spam and hate speech. That’s certainly the kind of outcome Big Tech has over the years cited as the main reason for keeping its black boxes closed.

However, as open-source advocates have been arguing for just as long, open code can make for better safety, because a host of outside experts get to pore over it and spot weaknesses that internal staffers missed. 

If it turns out that the benefits of opening up an algorithm like Twitter’s outweigh the risks, there may be major implications for the future of artificial intelligence regulation. 

In Europe, an Artificial Intelligence Act is being considered that would ensure that “high-risk A.I. systems shall be designed and developed in such a way to ensure that their operation is sufficiently transparent to enable users to interpret the system’s output and use it appropriately.” That would require a level of algorithmic transparency that Big Tech for now heavily resists.

If Musk makes good on his promise, that could be the first sign of the dam breaking. Everyone should be watching to see what happens next.

Want to send thoughts or suggestions to Data Sheet? Drop me a line here.

David Meyer

Data Sheet’s daily news section was written and curated by Andrea Guzman. 

NEWSWORTHY

Meta prepares for more layoffs. In November, Meta cut about 13% of its workforce. Now, rolling layoffs potentially affecting thousands of workers are expected at Facebook’s parent company in the coming months. The company plans to put some managers into lower-level roles while giving other managers larger teams as part of a reorganization, the Washington Post reports. Along with the job cuts, some projects may also be killed.  

Tesla makes up with California. Elon Musk and California Gov. Gavin Newsom said that Tesla would open a “global engineering headquarters” in Palo Alto. The move comes a couple years after Tesla moved its corporate headquarters from California to Austin. Shortly after the joint livestreamed announcement, during which the two said Tesla will occupy the former headquarters of Hewlett Packard, Tesla posted job openings for engineers.  

DOJ’s antitrust lawyers home in on Google Maps. Justice Department officials are meeting with Google’s competitors and customers in an effort to scope out information for a possible lawsuit challenging Google Maps’ control of mapping and location data. Citing unnamed sources, Politico reports that a lawsuit targeting Google Maps could be filed this year. It would become the third antitrust lawsuit against Google—with one filed in 2020 that targets search and another filed last month focused on its online ad business.  

Generative A.I. could boost chipmakers’ revenue. Chipmakers are expecting to cash in on the A.I. craze, with Intel, Nvidia, and Advanced Micro Devices creating CPUs, graphics chips, and other hardware tailored for A.I. computing. Analysts estimate the companies could see tens of billions of dollars in additional annual sales, the Wall Street Journal reports.

FOOD FOR THOUGHT

You can ask for a better severance package. More than 24,000 tech workers have been laid off this month alone, according to layoffs.fyi. And while some employers have offered generous severance packages, people can ask for even more as they head out the door, including additional weeks of pay, receiving a bonus sooner than scheduled, or having any noncompete agreements voided.

From the article: 

The most important thing to do once you learn about a job loss is to take a breath and not sign anything right away, says Crowder. While your emotions are running wild, there’s no reason you need to sign a severance agreement right when your employer hands it to you—that only benefits them.

Instead, ask for a day or two to review everything. Or longer: By federal law, workers over 40 get 21 days to decide whether or not to sign a severance agreement, plus another seven days to revoke it, says Peter J. Glennon, a New York employment and business litigation attorney. Other workers are given a “reasonable” time to consider.

IN CASE YOU MISSED IT

NLRB to Elon and other tech CEOs: No, you can’t gag workers in exchange for severance, by Christiaan Hetzner

Judge rules Dapper Labs can’t dodge lawsuit on whether Top Shot NFTs are securities, by Marco Quiroz-Gutierrez

Unstoppable Domains pledges patent nonaggression pact across expanded Web3 Domain Alliance, by Jeff John Roberts

Google wants returning staff to share desks with a ‘partner’ to maximize office space—but they’ve got to agree on ‘tidiness expectations,’ by Eleanor Pringle

No, you can’t copyright images made by A.I., says the U.S. Copyright Office, by Prarthana Prakash

Nvidia CEO says A.I. has hit an ‘inflection point’ as he praises ‘through the roof’ use of firm’s artificial intelligence services, by Nicholas Gordon

BEFORE YOU GO

Billionaire birthdays. At just 23 years old, Mark Zuckerberg became a billionaire. While Facebook may have helped that, his zodiac sign—Taurus—may have played a role. Using available birth date information for the top 300 billionaires on the Forbes Billionaires 2022 list, U.K. lender Cashfloat ranked the zodiac signs based on their representation among the world’s wealthiest people. Libra and Pisces led the list. Zuckerberg and other Tauruses were ranked third. The least likely group to be billionaires are Capricorns.