Every reasonable adult with even limited exposure to TikTok knows its political content skews to the far, Clown World left. This is how the topic of the app’s ban became politicized, after all, and why left wing journalists so often defend the right of an authoritarian prison state to sell its spyware unencumbered into US markets. But separate from the endlessly fascinating, furry-adjacent “frog/frogs pronoun” discourse, few topics on the Chinese spy app are so wildly divorced from reality as Israel. With over half of Zoomers using TikTok as their primary search engine, most of us saw the next bit coming, but the results are no less difficult to stomach: a recent Harvard/Harris survey found 51% of young people consider Hamas’s 10/7 mass murder justified.

If there were a terrorist attack in Times Square tomorrow, how many scantily-clad TikTokers would be dancing in defense of it by Friday?

We tend to blame this generational divide on a couple of convenient villains: first, our impressionable young minds are being shaped by an evil algorithm. Second, let’s talk about the spy app’s mysterious cabal of “content moderators” (CCP agents). But while I do think the algos are fucked, and believe the spy app should be banned, my sense is our deranged, divergent youth perspective is just what happens in a global information market.

In a long thread worth reading here, Jeff Morris Jr. describes the total, overwhelming victory of the anti-Israel perspective on TikTok, and credits the coordination of anti-Israel bot farms with shaping an illusion of popularity for the ‘Palestinian cause.’ In an attention economy, runs his argument, the perception of popularity provokes further anti-Israel content. A few media cycles into the flywheel, and your daughter’s shouting “Allahu Akbar” over Thanksgiving dinner (itself a celebration of genocide, she separately insists).

I think Jeff’s partly right. Just as is true in the real world, a sense of popularity surrounding any topic online attracts more interest in a subject. People are social creatures, and most of us don’t think about politics. Our goal is largely just to get along, which is why we generally adopt consensus positions — or rather, what we perceive to be consensus. This essential human quality is how basically all marketing works: convince the masses the masses already want what you’re selling, and rake in your Coca-Cola billions. But the flood of anti-Israel people we’re seeing online, and not only on TikTok but all over X, are not, for the most part, bots. They’re people, and I think we are probably all underestimating how many of them are foreign. Not agents. Not intentionally malicious. Just regular ass shitposting assholes from Bangladesh or whatever. We are probably also underestimating their impact.

There are around 240 million people living in Pakistan, half of whom speak English well enough to hop online and call you racist if you don’t want Sharia law (a real thing that happened to me this week). After Americans, the second largest population on TikTok is comprised of something like 100 million Indonesians, a majority of whom are Muslim, around 10 percent of whom speak English. Reception to the perspectives of Indonesians and Pakistanis, as with reception to the perspective of Americans, informs each platform’s algorithm. In all of this, a primary weight is placed behind majority opinion: what most people want, most people get. Given the psychology of humans, this in turn shapes culture.

We tend not to double click on who is saying what, or why. We just see likes, shares, comments, assume these are our peers, and an understanding of consensus forms.

The steel man here is this social dynamic works in both directions. A Paris-based terrorist sympathizer might perhaps be swayed by one of our kindly, dough-faced American teens, provided they generate enough support. Is that frog/frogs pronoun girl changing hearts and minds on the topic of crazy queer identity in Islamabad? I do think it’s possible. But if I were Jewish, representing a tiny .2% of the global population in a world increasingly shaped by majority opinion, would I find solace in the fact?

Probably not.

---

The Fifth Estate

Notable Industry Trends

Last week, we wrote a bit about the “yay Hamas” blacklists (jump down to the industry trends for my take on the ‘cancel culture’ of it all). This week, the battle of the blacklists continues. A list of Israel-defending tech investors has made the rounds, billed as a way for Palestinians to avoid the taint of apartheid, I guess. Basically every top venture firm was included. Bad news for all the Gazan founders (an enormous population, TechCrunch insists).

Last week, Biden issued an executive order on “Safe, Secure, and Trustworthy Artificial Intelligence,” promising to protect citizens from a wide range of negative AI outcomes including everything from the creation of “dangerous biological materials” to discriminatory landlords. Henceforth, companies developing models posing “a serious risk to national security, national economic security, or national public health and safety” have to notify the government when training the model, and share the results of red-team safety tests before release. The Department of Homeland Security will establish an AI Safety and Security Board as an enforcement entity (federal agents investigating nerdy Bay Area twinks, is how I’m reading this), and the Department of Commerce will “develop guidance for content authentication and watermarking to clearly label AI-generated content.” (Axios) (White House)

Fear not, they also released a Blueprint for an AI Bill of Rights.

I feel safer already!

Warren’s clown car war on crypto rages on: after the Wall Street Journal falsely claimed Hamas and other terror groups received hundreds of millions of dollars in crypto — an overstatement “by ~200x” — the nation’s favorite Native American Senator sent a letter to the Biden administration ‘demanding action’ (she wants it all banned). Following the story’s unraveling, Warren has of course doubled down. Elsewhere, an unshowered-looking John Fetterman repeated the inaccurate claims (video here). In response, Nic Carter is now offering a bounty for the 21 best “open source analyses of the facts surrounding the WSJ/Warren letter.”

Biden’s request for additional funding and rising tensions abroad have heated up the defense tech space. Anduril is on track to secure $625 million in new government contracts this year while Shield AI told investors it expected to grow more than 90%. This market signal is encouraging for investors in the space, along with the success of tech companies like Palantir winning contracts against incumbents.

---

Industry Links

Broad Tech

• Monday, at an impromptu event held 72 hours before his fourth-quarter earning call, Tim Apple unveiled a MacBook Pro and iMac model using his company’s new in-house M3 chip. A Hail Mary response to Apple’s 24.2% YoY decline in Mac sales? Hard to say (but definitely yes). (Apple)

• WeWork is planning to file for bankruptcy as early as next week. The company was valued at $47 billion in 2019. (WSJ)

• BP ordered $100 million of Tesla ultra-fast chargers as part of their plan to invest $1 billion in US EV charging. “As early as 2024, Tesla chargers will be installed” at gas stations including ampm, Amoco, and more. 

• Manifold unveiled a prediction market for matchmaking, in which you can bet on which of your single friends might happily couple off with each other for at least six months. Get your money, Yenta.

• 23&Me will give drugmaker GSK access to consumer DNA in a $20 million deal. (Bloomberg)

• “Google Pixel’s ad campaign is destroying humanity” is the headline of a Tuesday Washington Post opinion piece. In the fourth paragraph, with no hint of irony, the author argues: “It’s the main selling point here that bothers me — that the Pixel’s advanced artificial intelligence tools enable you to creatively edit your photos as never before.” Remember folks, if it’s good it’s bad. (WaPo)

• GM reached a tentative deal with UAW that could end the six-week strike at Detroit automakers. Next, union members have to ratify the agreement. The deal includes 25% pay increases over the four-and-a-half year term of the agreement, among other things. No four-day work week, alas. (CNBC)

Human resources:

• In a recap of 2023, Crunchbase is reporting more than 178,771 layoffs for U.S. tech companies; layoffs.fyi reports even more, at 246,214.

• LegalZoom intends to lay off a 100-plus person sales team in December, closing the division down.

• 59% of recent AI job posts are for roles in the Bay Area, according to compensation platform Comprehensive. (SF Chronicle)

• The All-In boys are looking for a CEO who must “understand the value of close friendship” and “living one’s best life.” They hope to unite their community with events, hot takes, and products. Could be you (provided you are a woman, per Jason’s request). Morning Brew founder Austin Rief threw his hat in the ring, and was promptly, publicly, rejected. (Call me, Austin, I’m tired.)

• Anduril hired Tesla and Toyota alum Keith Flynn to lead manufacturing. (Anduril)

• Fred Ehrsam, co-founder of crypto investment firm Paradigm, has stepped down as managing partner. (The Information)

$$$:

• Google will invest $2 billion in OpenAI competitor Anthropic after it secures a lead investor for the round they’re currently raising. (The Information)

• Four months after raising $113 million from Lightspeed Partners, European AI firm Mistral wants to raise another $300 million. (The Information)

• Shield AI, a defense firm building autonomous drones and other military tech, raised $200 million from Thomas Tull’s U.S. Innovative Technology Fund, Riot Ventures, and ARK, among others. (Bloomberg)

• Analytics company Databricks acquired enterprise data startup Arcion for $100 million. (Crunchbase)

• Traba, a light industry staffing marketplace, announced a $22 million funding round led by Founders Fund. The investment is the first from the firm’s eighth fund, signaling a focus on business metrics while most funds triple down on AI. (@mike_sheb)

• The British Ministry of Defence awarded Anduril a 31-month contract worth £17 million to develop future capabilities for the Royal Air Force and Strategic Command. (Anduril)

• Monday, fueled by concerns over inflation and geopolitical instability, and interest in potential ETF approval, Bitcoin nearly broke $35,000. May, 2022 was the last time BTC was priced higher than $35,000. (Bloomberg)

• Flexport is in talks to buy the technology assets of failed freight startup Convoy. (The Information)

• Meta revenue beat estimates and rose 23% to $34.15 billion in the third quarter. Profit has more than doubled from last year due to a resurgence of digital ad sales, and a 7% drop in expenses. (NYT)

• Technology “journalist” Casey Newton, who promised to leave Twitter almost one full year ago, once again insists Twitter is dead, Threads thriving

• (He did so on Twitter)

• Tesla shares dropped 20% since the company’s October 18th third quarter earnings call, against a 3.6% drop in the S&P, with analysts citing decreasing demand for EVs, high interest rates, and various issues with battery manufacturing. (Bloomberg)

Litigation and regulation:

• Autism Capital, the nation’s foremost SBF reporter, has called SBF’s courtroom testimony “disastrous,” and expects the trial will soon conclude. 

• The provenance of this hunky SBF courtroom portrait is “sketchy” at best, given longtime courtroom artist Jane Rosenberg — whose distinct style is much different — is assigned to the trial, and also that it is ‘too good to be true,’ but that hasn’t kept the image from going viral on X. (For what it’s worth, the three AI image detectors we ran the portrait through say it’s likely AI.) (@westcoastbill)

• The Google antitrust trial recently revealed the company paid Apple, Samsung, and Firefox a combined $26 billion+ to be the default search option in 2021, representing 58% of Google’s “traffic acquisition costs.” (The Information)

• A U.S. District Judge has dismissed parts of an AI copyright infringement lawsuit, throwing out charges against Midjourney and DeviantArt. A claim for direct infringement against Stability AI will proceed, however. (Hollywood Reporter)

• Thursday, two days after the California DMV ordered a halt of operations in the state in response to a human-caused hit and run incident, Cruise paused all driverless operations nationwide. (NYT)

• Hydrogen and electric semi-truck maker Nikola has successfully clawed back $165 million from its founder Trevor Milton. Nikola alleged the founder made fraudulent claims about the company’s technology that resulted in an SEC investigation, which blew up the company’s stock. (Forbes)

• Biden attack dog / FTC Chair Lina Khan will have a fireside chat with Y Combinator President and CEO Garry Tan at their SF office this Friday, Puck News writer Teddy Schleifer says.

AI:

• The AI-generated, never-ending Seinfeld Twitch stream (“Nothing, Forever” (Pirate Wires explainer here)) is down from thousands of viewers it had in February 2022 to just a few. There are only two AI-generated characters left in the show, neither of whom “say anything anymore and just stand still in complete silence…There is a strange orange man that patrols the house silently. He disappears for minutes at a time only to reappear sitting on a couch.” (@AnimeSerbia)

• Demonic?

• ChatGPT is multimodal for real this time — a new release gives users access to all GPT-4 features without having to switch between them for different tasks. (Search Engine Journal)

• Publishers are giving readers the option of having a conversation with books rather than reading them. They are hoping to target consumers who one might describe as “unable to shut the fuck up for once.” 

• OpenAI’s new 486,600-square-foot sublease from Uber is the biggest SF lease signed in the last five years, and one of the largest office deals in the city’s history. (SF Chronicle)

• We are so back

• (Or, AI is very popular right now)

• (Okay OpenAI in particular)

“Friends,” “allies”:

• The U.K’s AI summit kicks off today, with Kamala and Elon in attendance. Sunak, our British tech-bro-in-residence, said last week, “The U.K.’s answer is not to rush to regulate” (WaPo)

• The UN (remember them?) announced a 39-member advisory body on AI. Unclear why American tech executives and politicians continue to care. (Reuters)

• Citing “evolving European regulations,” Meta will charge European users a monthly subscription fee to not see ads in their Facebook and Instagram feeds. (The Information)

Trade war:

• Nvidia may have to cancel over $5 billion in orders headed to China, after the U.S. government sent them a letter last week explaining that export restrictions set to go into effect in mid-November are actually “effective immediately.” (WSJ)

• “If we don’t do anything, China’s going to get ahead of us,” Senate Majority leader Chuck Schumer said at a Washington Post event last week, referring to AI. (The Hill)

• TikTok is testing an AI feature that identifies products in unrelated videos, and encourages users to buy them. (Insider)

• TikTok remains, for some reason, not banned.

Godspeed, and see you next week.

-Solana

---

This letter was put together with a great deal of help from Brandon Gorrell and Sarah Bengard.