When I write the Weekly Roundup, I go back through the previous week’s news, perusing mainstream media, tech blogs, and a patchwork of custom RSS feeds I’ve developed over the last year. This week felt like I was catching up on 6-months worth of news over the span of a few days. I was shocked to see the ban-unban deluge of TikTok news only happened 6 days ago.
Benjamin Wittes, a writer I have long read and respected, wrote a compelling piece on how to deal with the news being served up via firehose lately. I’ll sprinkle in his 7 Principles throughout this post, with a slight reordering.
Principle #1: Slow Down. “The news actually doesn’t care if you follow it.”
Principle #2: Feel Free to Ignore Important News. “Grant yourself this indulgence. If you follow everything, you follow nothing well. Allow yourself to specialize. You’ll be better at the stuff you do follow.”
Principle #3: Let Others Do The Work For You.
Principle #4: Choose Trusted Sources.
In the spirit of abiding with #3 & #4, here is your Weekly Roundup:
TikTok, the ever-present headline
I’m unsure if the dust will ever settle on the TikTok issue. Still, amid the multitude of apocalyptic headlines over a video-sharing app, there were excellent pieces on different aspects of the controversy.
Trump delayed the enforcement of the TikTok Ban, and it is unclear if 1) he can legally do that and 2) if not, what liability will TikTok and companies supporting TikTok face?
“Unfortunately for the companies, established legal doctrine suggests that they are making a remarkably risky bet on both the scope and durability of [Trump’s] executive non-enforcement promises.”
“The non-enforcement promise offers minimal security. As discussed above, courts rarely treat such promises as binding, even when defendants face serious consequences from relying on them. Trump could change his mind at any time or selectively enforce against companies that fall from political favor, and a future administration, taking advantage of the five-year statute of limitations, would almost certainly be free to pursue violations regardless of Trump’s stance.”
How People Actually Make a Living on TikTok. This story draws on various users’ experiences and business models and the effect a ban would have. “I’ve built the business to accommodate the sales from TikTok Shop and TikTok in general,” Burns, [one interviewee], said. “Without that, it’s going to suck.” It is a warning for anyone who wants to center their entire livelihood on an application.
TikTok owner asks Chinese staff in Singapore to pay taxes to Beijing. This is an expected story for those who doubt the lip service TikTok is paying to American policymakers about being independent from the PRC.
American TikTokers Get a Taste of Chinese Censorship as They Rush to RedNote. “It took only about a day for Chinese censors to crack down on posts from Americans who had flooded onto the Chinese social-media app Xiaohongshu.”
TikTok users allege censorship, altered algorithms after Trump saved platform. Users complain that their algorithms “felt different” after the platform came back online. However, this may be a perception or part of the technical problem of rebooting a massively used social media app.
Principle #5: Don’t Get Your News From Social Media
I’m going to quote Wittes’ entire Principle #5 because it’s a principle I’ve written about a ton, and I feel very strongly that being able to choose content and free speech are inseparable.
I have nothing against social media. I like it for makeup videos and projections against Russian embassies and finding out which of my friends’ brothers-in-law just died and other such stuff. But it’s a bad way to get news.
The reason is that while it appears to put control of the news in your hands (you decide whom to follow, after all), it actually does something else: it lets you decide to whom to subcontract your sense of what is important. And it asks an algorithm to decide for you what is “for you” or what you should “discover” or what is popular among your friends.
When you look at the New York Times home page, you are making a decision that you’re going to let the editors of the New York Times—for the next block of time, at least—prioritize the universe for you. When you go on your Bluesky or X or Facebook feeds, you’re deciding to give that task to a random assortment of people you selected because they were something like you or related to you or interested in things that interest you. The overwhelming odds are that you have chosen a group of people who are (a) less expert at organizing the universe for you than the editors of the New York Times or the Wall Street Journal and (b) likely to magnify your prior opinions and anxieties. So if you get your news principally by scrolling, you are likely to speed things up, be less cautious about information, and be more anxious about it.
Just say no.
Or, failing that, just say no more of the time.
If you need any evidence that algorithms are curating your life for you in a way no other media has been able to before, here are some stories from just the past week.
“When I try to unfollow these accounts, the apps won’t let me. What’s up with that?
This is where it’s not you, it’s Meta.
The company said it “may take some time for follow and unfollow requests to go through”1 as the account transitions occur. It is possible that the company is receiving such a high volume of unfollow requests during the transition that it is running into errors processing them all.
Meta claims it will be sorted out soon, but declined to go into detail on why it was happening.”
TikTok Says It’s Not Censoring ‘Free Palestine’ Comments. Users See Something Different. This would be a stark departure from the algorithm under Biden’s Administration, where there was evidence that Pro-Palestinian content was being promoted and solid evidence that the algorithm is not so innocent.
It cannot be ignored, however, that the information feed on all social media platforms, not just TikTok, is curated to increase user engagement. TikTok, in particular, has a history of putting its thumb on the algorithmic scale of content recommendation. Pro-Palestinian hashtags far outnumber pro-Israel hashtags on the platform. Black Lives Matter content was sunk in 2020, and a “Stop the Ban” phone-banking campaign was boosted just last week, which (deliberately or not) targeted school children. Shadow banning is a common practice across the entire industry.
“A key driver behind the success of Chinese apps is that they have integrated e-commerce into their platforms, blending entertainment and networking with sales to monetize their famously addictive algorithms, according to Chinese social media experts and marketing firms. The platforms also continually surface fresh, engaging content, not only fueling binge-watching but also exposing users to new trends and shopping opportunities.”
If algorithms were random, would our problems be solved? Doubtful because computers can’t be random. Computers follow a set of instructions, so randomness means, by definition, the computer isn’t working. Spotify is an excellent example of trying a “random” algorithm.
Spotify’s first iteration of its shuffle feature was dictated by a decades-old algorithm that generated unbiased randomness from a finite sequence of elements. Breathtakingly efficient, the Fisher-Yates shuffle, developed by Sir Ronald Aylmer Fisher and Frank Yates in 1938, was employed by Spotify to dismantle user playlists and reassemble them into new, unpredictable orders.
It turned out, however, that users didn’t really want random; they wanted the appearance of random. Users didn’t like the first algorithm, so Spotify went in a different direction to create a smoother, less random experience for users. Their true method is unknown.
Principle #7: Don’t Be Afraid of Primary Sources
I parsed through most of the initial barrage of Executive Orders (EO) last week. Below, you will find the ones most applicable to technology. I refrained from too much analysis because many of the new EOs are being challenged in court (some of them are actually quite hollow and toothless), and usually, the first step a judge will take is to issue an injunction temporarily halting the order in question.
“Removing Barriers to American Leadership in Artificial Intelligence.” This EO ordered a new review to determine a new policy due to the previously revoked Executive Order 14110 of October 30, 2023 (Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence).
“Strengthening American Leadership in Digital Financial Technology.” This act essentially takes the leash off of crypto assets and development.
“Regulatory Freeze Pending Review” pauses all regulations for 60 days.
“Application of Protecting Americans from Foreign Adversary Controlled Applications Act to TikTok.” Citing the inconvenience of a law passed by Congress, signed by the former President, and upheld by the Supreme Court, President Trump instructed the AG not to take any action for 75 days so the new Administration can “determine the appropriate course forward in an orderly way that protects national security while avoiding an abrupt shutdown of a communications platform used by millions of Americans.”
Security
There have been a lot of commentaries regarding the efficacy of sanctions and DeepSeek. However, most don’t fully appreciate what sanctions are designed to do and how they work. ChinaTalk breaks down what the headlines miss:
The new Administration’s deportation plan will have plenty of technological help. The previous Administration has amassed an arsenal of surveillance technology that should bring a level of efficiency and precision to the deportations—however dystopic that may be.
A review of nearly 15,000 contracts shows that two agencies — Immigration and Customs Enforcement, and Citizen and Immigration Services — have spent $7.8 billion on immigration technologies from 263 companies since 2020.
The contracts, most of which were initiated under the Biden administration, included ones for tools that can rapidly prove family relationships with a DNA test to check whether, say, an adult migrant crossing the border with a minor are related. (Families are often treated differently from individuals.) Other systems compare biometrics against criminal records, alert agents to changes in address, follow cars with license plate readers, and rip and analyze data from phones, hard drives and cars.
The decision, issued Tuesday night by the eastern district of New York Judge LaShann DeArcy Hall, comes in the case of Agron Hasbajrami, a U.S. resident who was arrested in 2011 and initially pleaded guilty to attempting to provide material support to a terrorist organization. Hasbajrami appealed his case after learning that federal agents had acquired some of the evidence against him through a warrantless search of databases containing communications intercepted under the Foreign Intelligence Surveillance Act (FISA).
TSA’s facial recognition tech is highly accurate, review says. “The biometric technologies used at some U.S. airports to verify the identities of travelers are more than 99% accurate, the Department of Homeland Security said in a report released on Jan. 17.”
In other news
Two European countries have different views on digital IDs. The UK is seeing pushback by privacy groups on a digital ID scheme. Poland, on the other hand, while not throwing caution to the wind, seems to have more or less embraced the digital system:2
Wojciech Klicki, a privacy lawyer with the Panoptykon Foundation, a Polish rights NGO, stressed that such services needed to be designed to comply with strict “privacy-by-design, privacy-by-default” principles, but was not too concerned about their use. “We are primarily talking about the government reusing the data it already has or gets, just more efficiently,” he said.
California Is Investigating OpenAI’s Conversion To a For-Profit Company. “The state’s attorney general asked the company how it plans to transfer assets out of its charitable nonprofit.”
In motion to dismiss, chatbot platform Character AI claims it is protected by the First Amendment. The company is the Defendant in a lawsuit brought by the parents of a teenager the chatbot persuaded to commit suicide.
A South Carolina Company is courting buyers, possibly in the tech center, to purchase old nuclear reactors to power their data centers.
“Cancel Culture insurance” is being offered to celebrities and other at-risk public personalities.
Amazon has failed to make inroads in Argentina because of the country’s tariff system.
Principle #6: Use The Information You Are Taking In
“It is good citizenship to be informed. But the ideal citizen is not one who voraciously inhales news and does nothing with it except fret.”
I put #6 last instead of #7 because I felt it was a great note to end on. I don’t have any definitive advice for those wanting to use the information they are taking in, but writing your representative is an easy start. If you are looking for an emerging tech issue to provide content for your first letters, here are a couple:
Push to support Congressional (not Judicial or Executive) action on internet issues.
The failure of Congress to act as technology grows ever more powerful can be seen as an invitation for the executive and judicial branches of government in the U.S., as well as regulators in Europe and elsewhere, to fill the void. That leaves the tech industry, its business customers and consumers to navigate unresolved policy questions, a patchwork of state regulations and international rules, and federal agency policies vulnerable to court decisions.
Save your bureaucrats. There absolutely needs to be deep-state reformation, but many of them do excellent work that you hardly ever see and, unfortunately, may not realize how much you rely on their work until after they are gone. The Fifth Risk is a short read that gives you an inside look at agencies that will almost certainly be gutted in the coming months.
I’m suspicious, given that I don’t think I’ve ever had a problem instantly unfollowing any accounts.
I’ve seen this system in action in Poland and got some fellow travelers out of a pinch when they didn’t have other IDs.