Week 3/23 

​

⚖ TikTok on trial

​

TikTok CEO Shou Zi Chew appeared in front of the House Energy and Commerce Committee today – and in fact, is still taking questions from legislators as I write this. From the outset Chew was hit by intense criticism, not least from the Republican Committee Chair Cathy McMorris Rogers: in early remarks Rogers described the app as “a weapon by the Chinese Communist Party to spy on you and manipulate what you see,” and stayed at a similar level of rhetoric throughout.

Chew has, of course, tried to reassure lawmakers that TikTok is free from interference from the Chinese government, but it’s hard to know what to make of these claims. Though banning the app seems like a drastic step, opinions are divided among many smart tech commentators as to the level of security threat that TikTok really represents, and what should be done about it. 

Writing in the New York Times, veteran technology reporter (and my former boss) Julia Angwin argues that we should not ban TikTok, on the grounds that most of the surveillance allegations made against the Chinese video app could be levelled against just about any large tech company. Meanwhile, influential technology and economics blogger Noah Smith thinks we should absolutely ban TikTok, and points out – correctly, it should be said – the asymmetry in the fact that China already aggressively blocks the most popular US social media apps, and has done for a long time. 

​

EFF Director of Cybersecurity Eva Galperin argues that many people who object to TikTok don’t really care about privacy, they care about the fact that China has built a dominant social media platform. Though I’m a big proponent of what Galperin suggests instead of a ban, it’s also worth taking seriously some of the additional concerns around TikTok, like Beijing's ability to use it to promote narratives that further China's foreign policy aims. (Of course, one could argue that US social media companies are a tool of American foreign policy....but that’s always going to be acceptable to the US government, in a way that an app under the influence of a rival superpower is not.)

Ultimately, an all-out ban seems like a bad outcome here. A sale to a US company would be preferable – but few have the cash on hand to buy it, and those that do may be wary of running into antitrust suits.

In other news:

​

🧠 A telehealth startup shared its users’ mental health assessments from advertisers

This story comes via Dan Adler: telehealth startup Cerebral disclosed to its users that it shared the private health information of more than 3 million patients with advertisers. Seems bad! Almost like we need a comprehensive privacy law regulating data brokers.

​

📛 Meta sued in Kenya for human rights violations

Emily Tseng sent in this coverage of a lawsuit filed in Kenya, which alleges that content moderators working for Meta not only had to deal with viewing toxic and traumatizing content all day, but also suffered discrimination, poor working conditions, and unlawful termination.

​

🤐 Google’s low key chatbot release

Also, this week Google released its long-awaited chatbot, “Bard.” With different timing this could have been big news, but after first ChatGPT, then Microsoft’s delightfully unhinged BingGPT (a.k.a. Sydney), Google’s offering has felt boring to some commentators. 

​

Also, the NYT published an article claiming that conservatives are trying to build their own chatbot, but it seems to be based mostly on the work of an academic researcher from New Zealand studying bias in LLM’s? Never let the facts get in the way of a good headline I guess.

 

Week 3/16 

​

🏦🔥 Silicon Valley Bank goes up in flames.

​

Silicon Valley Bank was the darling of the West Coast tech scene. Around this time last week, you could have walked into a branch, deposited some cash, taken out a loan, or done whatever else it is that people do in banks. Now, pretty much in the blink of an eye, it's extinct — the largest bank to collapse since Washington Mutual in 2008.

The rapid failure of SVB, which has now been taken over by the federal government, came as a total surprise to most people. Because this is America, why people think it actually happened depends heavily on political ideology and…well…willingness to actually learn things, so competing explanations have occurred.

Here are some different interpretations of the problem with SVB:

 

• It managed investment risk poorly

 

In late February, the Financial Times wrote an article pointing out that SVB was facing scrutiny over an investment decision it had made at the peak of the tech boom. Without getting too deep into the weeds, the bank had poured a lot of consumer deposits into assets that had depreciated in value — but as long as it didn't sell the assets, it never had to declare the loss. This meant that on paper, the bank's market capitalization appeared higher than it actually was, as illustrated in this chart from the FT's Tabby Kinder.

On Wednesday of last week, a combination of economic conditions meant that SVB was finally forced to sell some of these assets, and properly disclose that it had made a huge loss. Immediately certain well-connected people in the startup world started to get spooked — notably mega-investor / surveillance capitalist overlord Peter Thiel, whose Founders Fund VC firm advised all companies in his investment portfolio to pull their money out of SVB on Thursday. As news of that call leaked out, more and more SVB customers rushed to pull their money out, leading to a classic bank run that wiped out SVB by Friday. (If you do want to get into the weeds of things, you can read more about this in Matt Levine's consistently fantastic Money Stuff newsletter.)

 

• It escaped regulation (which was already weakened)

 

A decade and a half ago when we had another crushing financial crisis, the people involved in regulating the finance industry realized that things had to change. So they put in place a bunch of regulations aimed at preventing the exact same thing from happening again. Unfortunately, the Trump administration had a fairly contemptuous attitude towards regulation across the board, so a whole bunch of the Dodd-Frank regulations passed in 2010 were rolled back in 2018 (with the support of some Democrats, it should be said). More to the point, SVB spent a lot of money specifically lobbying against these regulations, by arguing that exactly the kind of collapse which happened, would never in fact happen.

 

• It was too "woke"

 

Despite the fact that SVB catered to a wealthy, elite clientele which — as mentioned above — had enough political cloud to avoid the kind of regulation that would have stopped this, the fact it had a DEI program was enough for right wing commentators to brand it "one of the most woke banks." This discourse even made it as far as the Wall Street Journal, where opinion columnist Andy Kessler wrote the vomit-inducing line, "I’m not saying 12 white men would have avoided this mess, but the company may have been distracted by diversity demands."

So: risky, deregulated, or woke — pick your fighter. Regardless of which you believe, it's also worth reading this piece in Slate by Edward Ongweso Jr. on the broader systemic failures of the VC industry

 

In other news:

​

😰 More layoffs at Meta. The company announced that more than 10,000 employees would lose their jobs in the next two months, after 11,000 people were cut at the end of last year.

🏕️ Elon Musk is trying to build a company town in Texas, which is one of the most miserable things I can imagine.

📖 Gigi Sohn might write a book about her FCC confirmation battle, referenced in last week's installment of TNT. Spill that tea Gigi!

​

Week 3/9 

​

☎️ Telcos push out FCC nominee

This week Gigi Sohn, Biden’s nominee for the vacant seat on the Federal Communications Commission, withdrew her candidacy, citing “cruel attacks” from right wing media and smear campaigns from cable and media industry lobbyists. It’s a major victory for the telecom companies, who have fought ferociously against Sohn's nomination for the past year and a half. In the meantime the FCC’s four remaining commissioners are deadlocked 2-2 along partisan lines, meaning the Biden administration will find it hard to push through much-needed internet provider reforms, and our broadband will continue to suck.

 

📽️ Support for a TikTok ban still growing

In other DC news, a bipartisan group of US senators proposed legislation that would empower the government to ban TikTok and potentially other apps too. CNN writes: “The legislation, called the Restricting the Emergence of Security Threats that Risk Information and Communications Technology (RESTRICT) Act, does not target TikTok specifically for a ban. But it aims to give the US government new powers, up to and including a ban, against foreign-linked producers of electronics or software that the Commerce Department deems to be a national security risk.” The White House has apparently endorsed the bill, marking a change of direction from the administration on this issue.

 

🐦 FTC looks deeper into Twitter

The FTC is now seeking testimony from Twitter owner and disability awareness advocate Elon Musk, according to the New York Times. In a case that started long before Musk’s ownership, the FTC was given oversight of Twitter’s privacy and security practices after a series of data breaches exposed vast amounts of user data. The trade regulator has now raised concerns that Musk’s drastic cuts to staffing make the company incapable of meeting its obligations under the supervision decree.

 

🗣️ Scammers leverage voice cloning

A number of stories recently have highlighted the dangers of AI voice cloning technology, as in this Washington Post report. Although voice cloning has been around for a while, the amount of audio needed to generate a convincing fake has decreased significantly, while the sophistication of the clone voices created has increased. 

 

✴️ Chomsky weighs in on ChatGPT

In a New York Times guest essay, the creator of the theory of universal grammar gives his thoughts on LLMs. On a factual level Chomsky and his co-authors aren't breaking much new ground here, but the range of references — Borges, Sherlock Holmes, Newton's laws of motion, "the banality of evil" — makes this more entertaining than run-of-the-mill AI criticism.

 

🕵️ WIRED goes inside the ‘suspicion machine’

This week WIRED launched a new series exploring algorithms used to flag welfare fraud. Worth a read for any scholars of algorithmic bias — and also illuminates the way that private companies are increasingly brought in to design the systems that administer public benefits.