View profile

📰 BTN: Google blunders AI ethics (again), Apple's baffling hire, Instagram algorithms, GitHub passwords, Amazon fraud

📰 BTN: Google blunders AI ethics (again), Apple's baffling hire, Instagram algorithms, GitHub passwords, Amazon fraud
By Emil Protalinski • Issue #13 • View online
Hey there, I’m Emil Protalinski. Big Tech News (BTN) is where I recap the week’s news across Alphabet, Amazon, Apple, Facebook, and Microsoft. I also sneak in some lower-case big tech news.
This week, notice all the ways the EU is clamping down on Big Tech. Let’s dive in.

Alphabet
Byte: I ran out of analogies for Google’s growing list of AI ethics blunders. In February, AI Chief Jeff Dean said his team should have managed the Timnit Gebru firing “with more sensitivity,” and then simultaneously fired her counterpart Margaret Mitchell. I described that as Google plowing a rescue helicopter into its AI ethics train wreck. This week, the company said that in the coming years, it plans to double the size of its team studying AI ethics. That sounds good, until you learn how the news was rolled out. Google VP of Engineering Marian Croak, who took over after the firings of Gebru and Mitchell, made the announcement at a WSJ event, failed to tell the responsible AI team first, and revealed that the main motivation is really PR. “Being responsible in the way that you develop and deploy AI technology is fundamental to the good of the business,” Croak said. “It severely damages the brand if things aren’t done in an ethical way.” Google is investing in AI ethics not because it cares about the various ramifications of the technology, but because it sees dollar signs. Furthermore, it’s doubling down on that investment for PR reasons. Both are the company’s prerogative, of course, as is Croak’s apparent decision not to communicate plans of expanding the responsible AI team to people on the responsible AI team. OK Google, good luck filling those roles.
Bits:
  • SpaceX signed a deal with Google Cloud for its satellite internet service that will put Starlink terminals in Google’s cloud data centers around the world. Later this year, Starlink customers will be able to use Google Cloud features while Google’s enterprise cloud customers will be able to use Starlink’s satellite internet. The deal isn’t as out-of-the-blue as it might seem: To date, Google has invested $900 million into Starlink manufacturing and other SpaceX tech.
  • Italy’s antitrust watchdog fined Google €100 million for abusing its dominant market position by restricting third-party access to the Android Auto app. Meanwhile, Sonos said a German court had granted a preliminary injunction that could force Google to halt sales of its media players, phones, and smart speakers in the country. Both cases sound major but at Google scale they are dismissible headaches.
  • YouTube announced the YouTube Shorts Fund, which will pay $100 million to top creators on YouTube Shorts, the company’s TikTok ripoff, through 2022. It’s a big sum, but not the biggest; Snap’s Spotlight, for example, pays out $1 million every day to Snapchat users that make the best Snaps. YouTube’s fund will start doling out the dough across the U.S. (home turf) and India (where TikTok is banned) in the next few months.
Apple
Byte: The hiring department apparently didn’t get the memo on Apple’s fiery Facebook feud and its stance on privacy. Apple hired Antonio García Martínez, a former Facebook product manager who helped build the company’s advertising business and has a history of racist and sexist statements, for an ads engineering role that started last week. Apple workers petitioned for an investigation into his hiring, citing Martínez’s comments about women and people of color. The petition resulted in Apple confirming Martínez is no longer employed by the company and that it has “always strived to create an inclusive, welcoming workplace where everyone is respected and accepted. Behavior that demeans or discriminates against people for who they are has no place here.” Until now, Apple has largely avoided public reckonings from its own employees compared to other Big Tech firms. The first such instance was a doozy: Questioning Apple for hiring someone who argues against consumer privacy rights and consistently disparages people. Is this a glimpse of how far Apple is willing to go for a piece of Facebook’s ads business?
Bits:
  • A security researcher claimed he was able to break into the microcontroller of an Apple AirTag, letting him modify its NFC URL. While that hack requires physical access, another researcher detailed an exploit in Apple’s Find My network that allows the transmission of arbitrary data by simulating a Find My broadcast from non-internet-connected devices. Apple customers aren’t going to stop buying AirTags nor stop using Find My, so the company should engage these researchers to avoid future sensitive security snafus.
  • Apple seemingly gave Zoom access to an iPadOS camera API while in Split View. Not only is there no public process for requesting the entitlement, there isn’t any documentation of its existence. Add this to the long list of counterexamples to Apple’s claim that it treats all developers equally.
  • The Apple vs. Epic trial revealed some more tidbits, including that over 500 people review around 100,000 apps/week for the company, resulting in an app rejection rate of less than 40% and less than 1% of rejections getting appealed. Apple is still expected to win the trial, but in the long run developers may end up getting more information about the app review process as well as other concessions that level the playing field.
Facebook
Byte: Instagram (and Twitter) apologized and blamed technical errors by their automated systems for the deletion of pro-Palestinian posts. Instagram took down content, including archives from older stories, that clearly didn’t break its rules: posts mentioning the possible eviction of Palestinians from East Jerusalem. Instagram also removed content associated with the Al-Aqsa mosque where Israeli forces and Palestinians clashed, because it labeled the site as a terrorist organization. “Both of these mistakes and many others are entirely unacceptable,” a Facebook employee wrote on an internal communications platform. “Al-Aqsa is the 3rd holiest site in Islam and is a central aspect of faith for some 1.8 billion people.” This is a pattern. Social networks love to blame glitches in their algorithms for any and every screwup. Given that their broken algorithms screw up so consistently and that Facebook refuses to turn them off, who can blame the users for believing the glitches are by design?
Bits:
  • Instagram announced it will let people list up to four pronouns on their profiles, viewable to friends or the public in a “few countries.” This is great, as long as Instagram doesn’t push you to add pronouns to your profile. People may be in the process of figuring that out, or simply not want to share.
  • Facebook started testing a pop-up that asks if you’re sure you want to share a news article link that you haven’t opened. Twitter tested and rolled out a similar prompt last year, finding that people opened articles 40% more often after seeing the prompt and that the number of people opening articles before retweeting increased by 33%. In what I can only describe as a lazy attempt to credit Twitter with the idea, Facebook tweeted its announcement.
  • Concern over WhatsApp’s plan to render the app unusable if users don’t accept the new privacy policy ramped up as the rollout date (May 15) neared. In related news, Germany’s leading data protection regulator banned Facebook from processing WhatsApp user data, but Facebook said it would ignore the ruling. Separately, Ireland rejected Facebook’s attempt to block a decision that would force the company to stop sending EU user data to the U.S. All in all, expect more EU fines for Facebook.
Microsoft
Byte: GitHub added security key support for Git operations over SSH, providing a more secure alternative to authenticate with Git while preventing unintended and potentially malicious access. This is part of a larger plan to remove password support for Git, as GitHub already did for its API: “Later this year, as we continue to iterate toward more secure authentication patterns, passwords will no longer be supported for Git operations. We recognize that passwords are convenient, but they are a consistent source of account security challenges. We believe passwords represent the present and past, but not the future. We would rather invest in alternatives, like our Personal Access Tokens, by adding features such as fine-grained access and more control over expiration. It’s a long journey, but every effort to reduce the use of passwords has improved the security of the entire GitHub ecosystem.” It’s nice to see Microsoft’s war on the password extend to GitHub. Developers are the most tech-savvy users and thus should be the first to ditch passwords completely. GitHub is the ideal company to plow that idea forward.
Bits:
  • GitHub finally added the ability for developers to upload videos to demo bugs and features. You can now include .mp4 or .mov files directly in issues, pull requests, discussions, and so on. This might seem trivial, but given that GitHub is the largest host of source code in the world, it could noticeably impact how software is developed.
  • Microsoft quietly decided to shut down its Azure Blockchain Service on September 10, six years after its initial launch. The company will not support any new deployments as of May 10. I never understood the blockchain-on-the-cloud pitch, since it effectively added centralization (Microsoft, a single point of failure) to decentralization (blockchain, a distributed system).
  • The Pentagon said it is reconsidering its JEDI cloud computing project, the $10 billion contract for which Microsoft won in October 2019. The decision has seen multiple legal challenges, most notably from Amazon, which blamed Donald Trump’s dislike of CEO Jeff Bezos for the company’s loss of the deal to Microsoft. I suspect both will end up getting contracts in the end, but the change is a big win for Amazon, which leads the cloud computing market, and a big blow to Microsoft, which is in second.
Amazon
Byte: Amazon released its first Brand Protection Report, detailing how it spent $700 million to combat counterfeit products. In 2020, Amazon employed more than 10,000 people to protect its store, prevented over 6 million attempts to create selling accounts before bad actors published a product for sale, and blocked more than 10 billion suspected bad listings before they hit the store. The company also detected, seized, and destroyed more than 2 million counterfeit products in its fulfillment centers before they were sent to customers. Coincidentally (or not?), Amazon published the report a day before delisting major electronics sellers Aukey and Mpow, possibly for soliciting reviews on the store. Amazon provided a boilerplate statement: “We have long-standing policies to protect the integrity of our store, including product authenticity, genuine reviews, and products meeting the expectations of our customers. We take swift action against those that violate them, including suspending or removing selling privileges. We take this responsibility seriously, monitor our decision accuracy and maintain a high bar. We have an appeals process where sellers can explain how they will prevent the violation from happening in the future or let us know if they believe they were compliant. Our teams are based in our Seattle headquarters and around the globe in order to provide sellers with 24/7 support via email, phone, and chat in more than 15 languages.” It doesn’t matter if your product is counterfeit or your reviews are fake: Amazon will happily ban you for any funny business, and rightly so.
Bits:
  • Amazon said it is hiring 75,000 workers across the U.S. and Canada in its fulfillment and logistics network with an average pay of $17/hour, above its typical $15/hour starting wage. Overall, this is a positive step. Given CEO Jeff Bezos’ promise last month to make Amazon “Earth’s Best Employer and Earth’s Safest Place to Work,” however, it’s barely a baby step.
  • Amazon won an appeal over €250 million in alleged unpaid EU taxes, with the court ruling that competition regulators failed to prove Amazon got a selective advantage from tax breaks. In July, the same court ruled similarly in a case that saved Apple from a €13 billion tax bill. This will keep happening as long as people sue corporations for paying unfairly low taxes instead of changing the laws so corporations can’t pay unfairly low taxes.
  • After just five weeks at Bessemer Venture Partners, former Amazon executive Jeff Blackburn returned to the company to oversee Prime Video, Amazon Studios, Amazon Music, Wondery, podcasts, Audible, games, and Twitch. Amazon has successfully neutralized both Bessemers.
Overflow
Byte: The Information linked seven Apple component suppliers to alleged forced labor involving Uyghurs and minorities in China. The newly identified suppliers in China also worked on devices for dozens of Western companies, including Amazon, Facebook, Google, and Microsoft. Apple said in a statement that “looking for the presence of forced labor is part of every assessment we conduct in every country where we do business.” One factory in Xinjiang that makes Apple computer parts has a detention center literally next door. The Uyghur genocide is one of the most important humanitarian crises of our time. Chances are the device you are using to read this newsletter is linked to it.
Bits:
  • Nikkei found that the average tax burden ratio for Amazon, Apple, Google, and Facebook is around 60% of the average among more than 50,000 large companies worldwide. It’s time for a coordinated international effort to increase audits, close loopholes, and eliminate tax havens.
  • Google pushed for maintaining U.S. work authorization for tens of thousands of people whose spouses hold H-1B visas. Amazon, Apple, and Microsoft also signed onto the amicus brief, as did Adobe, Electronic Arts, eBay, IBM, Intel, PayPal, Reddit, StubHub, and Twitter.
  • Roku announced original programming, including the Quibi content it acquired, will hit the Roku Channel on May 20. I don’t care for Quibi content, but it’s nice to see consolidation that didn’t end with a Big Tech firm.
That’s all for now. Get in touch by hitting reply to this email, forward this issue to a colleague or friend, and see you next Friday!
Did you enjoy this issue?
Emil Protalinski

Big Tech News -- is that news about Big Tech or tech news that's big? Yes.

If you don't want these updates anymore, please unsubscribe here.
If you were forwarded this newsletter and you like it, you can subscribe here.
Powered by Revue