E47: Facebook's week from hell, Ellen Pao on sexism in Elizabeth Holmes coverage, Newsom's win, frauds & more

The prediction was that if Republicans continued to embrace the 2020 "stolen election" myth through the 2022 cycle, it would be an albatross that would "bring them down" in the 2022 midterms (i.e., cause broad Republican losses or clear underperformance).

What actually happened in November 2022:

• Republicans did win the U.S. House (albeit narrowly), gaining control from Democrats.
• Democrats held the U.S. Senate and even slightly expanded their majority (51–49 with the Georgia runoff), but this is better characterized as a mixed or split result rather than a clear, across-the-board GOP collapse.
• Historically, the president’s party typically loses more House seats in a first midterm than Democrats did under Biden; many analysts described 2022 as a Republican underperformance relative to expectations, but not as Republicans being broadly “brought down” or suffering a decisive electoral wipeout.
• Post‑election analysis (e.g., in major outlets like the New York Times, Washington Post, and various election analysts) did argue that election‑denialism and Trump‑aligned, 2020‑obsessed candidates probably hurt Republicans in a number of key races, especially statewide contests and in swing districts, but the aggregate outcome still left the GOP with control of the House. That is inconsistent with the stronger claim that the “stolen election” myth would bring them down in 2022 overall.

So while election denial likely contributed to some important Republican underperformance in specific races, the broad empirical outcome (winning the House and only narrowly failing in the Senate) does not match the stronger forecast that this issue would bring Republicans down in the 2022 midterms. On balance, that makes the prediction wrong rather than right or ambiguous.

Evidence since 2021 shows some support for Sacks’s idea that Democrats in deep‑blue jurisdictions use Trump/MAGA branding against opponents, but it does not support his stronger claim that this would become the enduring playbook, displacing local‑governance issues.

Where the prediction fits reality (partially)

• In the 2021 California recall, Gavin Newsom explicitly framed the race as a fight against “Republicans and Trump supporters” and “Trumpism,” branding the recall as the “Republican recall” and warning that “Trumpism is still on the ballot in California.” (foxnews.com)
• In New York’s 2022 governor race—an overwhelmingly Democratic state—Kathy Hochul repeatedly tied Lee Zeldin to Trump and the “MAGA agenda,” calling his events a “MAGA Republican bus tour,” even as crime and public safety were the top voter concerns. (wamc.org)
• In the 2022 Los Angeles mayor’s race, Karen Bass’s campaign and an allied group ran ads linking Rick Caruso to Trump, emphasizing his long Republican history and portraying him as Trump‑aligned in a heavily Democratic city. (foxnews.com)
• In Chicago’s 2023 mayoral runoff, Brandon Johnson repeatedly portrayed Paul Vallas as effectively a Republican backed by donors who also supported Donald Trump and by “right‑wing extremists,” and his campaign was accused (by local GOP) of pushing fake “MAGA” Vallas signs. (chicago.suntimes.com)

These examples show that anti‑Trump / anti‑MAGA framing is indeed a recurring tactic used by progressive or mainstream Democratic candidates in blue jurisdictions, sometimes against opponents whose own records are more mixed or moderate than “Trumpist” branding suggests. That aligns with part of Sacks’s forecast.

Where the prediction fails

• In the same LA mayor’s race, mainstream coverage and the candidates’ debate focused overwhelmingly on local concerns—homelessness, crime, corruption, and housing—rather than treating Trump as the central issue. Bass’s closing message emphasized solving homelessness, reducing crime, and making housing affordable, not Trump. (latimes.com) This undercuts the claim that campaigns in such cities would primarily be about Trumpism rather than local governance.
• Chicago’s 2023 mayoral campaign was dominated by crime, schools, taxes, and policing. Johnson’s attacks on Vallas’s Republican/Trump ties were one line of attack layered onto a fundamentally local, crime‑and‑schools‑centered race, not a replacement for those issues. (en.wikipedia.org)
• In San Francisco, some of the emblematic contests in this period—the 2022 school‑board recall, the 2022 recall of progressive DA Chesa Boudin, and the 2025 recall of Supervisor Joel Engardio—have been framed primarily around school reopening, competence, crime, homelessness, and quality‑of‑life concerns. Trump is largely absent from the pro‑recall messaging; instead, voters are said to be reacting to perceived local mismanagement and safety problems. (en.wikipedia.org)
• Even where Trump/MAGA rhetoric appears (e.g., Newsom’s recall, Hochul vs. Zeldin), it operates alongside, not instead of, issues like COVID policy, abortion rights, guns, crime, and cost of living. The campaigns are not simply “always campaigning against Trump” while ignoring local or policy specifics.

Assessment

• Sacks predicted that from that point forward, progressive Democrats in jurisdictions they control would always campaign against Trump/Trumpism and would rely on branding any opponent as a Trumpist “whether they are or not,” effectively making that “the playbook from now on” instead of contesting local governance.
• The record through late 2025 shows a more mixed reality: anti‑Trump/MAGA branding is an important and recurrent tool, but major state and city races in blue areas still hinge largely on, and are publicly framed around, local performance—crime, homelessness, schools, housing, and corruption.
• Because we already see numerous high‑salience elections where local‑governance themes dominate and Trump plays a minor or secondary role, the strong version of his forecast (“they’re always going to campaign against Trump and Trumpism” rather than on local issues) has not materialized.

Given that, the fairest scoring is that the prediction, as stated, is wrong: it overstates the degree to which anti‑Trumpism has displaced local governance as the primary campaign frame in progressive‑run cities and states, even though elements of the behavior he described do occur regularly.

Chamath’s prediction has two explicit time components tied to future candidate fields:

1. Effect on who ran against Newsom in his next regular election (~2022)

   • The 2021 recall was held on September 14, 2021, and Newsom survived comfortably, with about 62% voting “No” on the recall. (en.wikipedia.org)
   • Newsom then won re‑election in the regular 2022 California gubernatorial election on November 8, 2022, defeating Republican state senator Brian Dahle with about 59% of the vote. (en.wikipedia.org)
   • The 2022 primary and general fields featured Newsom as the dominant Democrat with no top‑tier Democratic challenger; however, there is no clear empirical evidence isolating the recall as the cause of which Democrats did or did not enter that race, and any comparison to a hypothetical world without the recall is inherently counterfactual.

2. Effect on which Democrats run in the next open gubernatorial cycle (~2026/2027)

   • Newsom is term‑limited and cannot run again in 2026; the 2026 California gubernatorial election is scheduled for November 3, 2026, with a top‑two primary on June 2, 2026. (en.wikipedia.org)
   • As of November 2025, the race is still ongoing with a large and fluid Democratic field (e.g., Katie Porter, Xavier Becerra, Tony Thurmond, Antonio Villaraigosa, Betty Yee and others) and several prominent figures having declined or exited (e.g., Kamala Harris and Alex Padilla declining to run; Eleni Kounalakis dropping out). (en.wikipedia.org)
   • Candidate filing deadlines for 2026 statewide races (including lieutenant governor and attorney general) run into March 2026, and the final gubernatorial ballot is not yet set; the field is still actively changing. (en.wikipedia.org)
   • No scholarly or journalistic consensus directly attributes the composition of this evolving 2026 field to the 2021 recall in a way that would let us test Chamath’s specific counterfactual claim (“compared with the candidate fields that would have emerged absent the recall”).

Because (a) the 2026 open-seat race has not yet reached its final candidate field, and (b) the claim itself is counterfactual and there is no rigorous evidence tying observed candidate choices in 2022 or 2026 directly to the recall, there isn’t enough concrete, testable information to decisively say the prediction was right or wrong at this time. Hence the status is inconclusive (too early and too counterfactual to verify).

Evidence since 2021 shows that both parts of Chamath’s prediction have materialized.

1. U.S. state AG lawsuits and related class / mass actions over teen mental health harms

• On October 24, 2023, a bipartisan coalition of 41–42 state attorneys general and D.C. filed coordinated lawsuits against Meta (Facebook and Instagram) in federal and state courts. The complaints allege that Meta knowingly designed and deployed addictive features that harm children’s and teens’ mental health, fueling a “youth mental health crisis” and violating state consumer protection laws and COPPA. (ag.state.mn.us)

• These state AG filings explicitly build on the 2021 Wall Street Journal reporting and Frances Haugen’s disclosures about internal Meta research showing Instagram worsened body image and mental health for teen girls. (apnews.com)

• Separately, hundreds (now over 2,000) private lawsuits by minors, families and school districts have been consolidated in In re Social Media Adolescent Addiction/Personal Injury Products Liability Litigation (MDL No. 3047) in the Northern District of California. Plaintiffs allege that platforms including Facebook and Instagram are deliberately designed to be addictive for children and teens and have caused depression, anxiety, self‑harm, and eating disorders. (trulaw.com)

• Individual suits against Instagram specifically claim it fueled teen girls’ addiction to the app and led to depression, anxiety and anorexia, including hospitalizations and suicide attempts, squarely matching the “eating disorders” and anxiety harms Chamath mentioned. (foxbusiness.com)

  Together, this satisfies his forecast that multiple state AGs would sue and that there would be extensive class/mass‑action style litigation over youth mental health harms allegedly caused or exacerbated by Instagram/Facebook.

2. Non‑U.S. government investigations / actions over these issues within a few years

• Under the EU’s Digital Services Act (DSA), the European Commission opened a formal investigation into Meta (Facebook and Instagram) focused on “addictive and rabbit‑hole effects” on minors, the adequacy of age‑verification, and privacy around recommender systems, explicitly citing risks to the physical and mental health of young Europeans. (voz.us)

• In 2025 the Commission went further, issuing preliminary findings that Meta is in breach of DSA obligations, including failures around mechanisms for reporting illegal and harmful content and inadequate transparency and data access related to minors’ exposure to harmful material—again tied to risks to users’ mental and physical health. (theguardian.com)

  These EU actions are regulatory proceedings by a non‑U.S. government entity, initiated within a few years of the WSJ revelations, and are explicitly grounded in concerns about youth harm and mental health on Facebook/Instagram.

Because (a) many U.S. state AGs have indeed brought major lawsuits over teen mental‑health harms linked to Instagram/Facebook, alongside extensive private and school‑district litigation, and (b) European authorities have launched formal investigations and enforcement actions focused on the platforms’ negative mental‑health effects on minors, Chamath’s prediction has been borne out.

Available evidence shows that, as of late 2025, the U.S. Food and Drug Administration has not asserted product-style regulatory authority over social media platforms (e.g., Instagram) on the basis of mental-health or eating‑disorder harms, in any way analogous to its regulation of products like Juul.

What the FDA does regulate online is: (a) how manufacturers advertise FDA‑regulated medical products on the internet and social media, and (b) its own use of social media for public communication. FDA guidance focuses on ensuring truthful, balanced risk–benefit information in drug/device promotion on social platforms, not treating the platforms or their algorithms themselves as regulated products.(fda.gov) This is a continuation of long‑standing promotional‑advertising oversight, not a new jurisdictional move targeting social media as a public‑health product.

By contrast, concern about social media’s mental‑health impact has mainly produced non‑FDA actions:

• In May 2023, U.S. Surgeon General Vivek Murthy issued an advisory on social media and youth mental health, warning of “profound risk” and urging policymakers and tech firms to protect children, but this was an advisory, not binding regulation of platforms.(integrationacademy.ahrq.gov)
• In June 2024, Murthy publicly called on Congress to require tobacco‑style warning labels on social media platforms, again highlighting harms but explicitly noting such labels would require new legislation; they have not been federally mandated to date.(apnews.com)
• Multiple state laws now attempt to regulate youth social media use and to require mental‑health warning labels (e.g., Colorado’s HB 24‑1136 and similar California measures), typically enforced by state attorneys general, not federal health regulators—and some have already been blocked in court.(en.wikipedia.org)
• In Congress, bills like the Kids Online Safety Act and Kids Off Social Media Act would impose duties on platforms or restrict minors’ use, but these proposals assign enforcement mainly to the FTC and state AGs, not the FDA, and key bills have stalled or remain pending.(en.wikipedia.org)

None of these developments amount to the FDA—or any comparable federal health regulator—reclassifying mainstream social media platforms or their algorithms as regulated health products because of mental‑health or eating‑disorder impacts. The regulatory model remains advertising oversight plus general consumer‑protection and privacy law, not FDA‑style product regulation of Instagram itself. Given that more than four years have passed since 2021 with no such assertion of authority, the normalized prediction that social media products would come under FDA‑type public‑health regulation within “the next several years” has not materialized.

This prediction concerns typical lifespans of a generation of middle‑aged adults in 2021 (people who might have been, say, 35–55 years old then). For such a cohort, we cannot know in 2025 whether it is “very likely [they] will live to [their] hundreds” because:

1. They have not yet reached the age in question. Someone who was 45 in 2021 will be about 49 in 2025, decades away from age 100. No amount of current data can yet show what fraction of that cohort will actually reach 100.
2. Available longevity data is about past and current mortality, not the realized lifespan of this future cohort. Even though life expectancy in developed countries has generally risen over the last century, it has recently stagnated or even declined slightly in places like the U.S., especially around the COVID‑19 period and opioid crisis, which complicates projections—but does not yet definitively falsify or confirm a claim about how many current middle‑aged adults will reach 100.
3. Longevity forecasts are inherently long‑term. To judge whether it is “very likely that we’re all going to live to our hundreds” for that generation, we would need to observe outcomes many decades from now (around 2050–2080). As of November 30, 2025, that evidence does not and cannot exist.

Because the prediction is about outcomes that mostly lie many decades in the future, and insufficient time has passed to observe whether a large share of that generation actually reaches 100+, the correctness of the prediction cannot yet be determined.

Available evidence suggests Jason’s prediction has largely come true.

1. Condition about no new 9/11‑scale attack
   Since the August 2021 withdrawal there have been serious terrorist incidents, including the 2025 New Orleans truck attack (15 killed), but nothing remotely on the scale of 9/11’s ~3,000 deaths or its transformative political impact. These attacks are described as among the deadliest since 9/11, not comparable to it, and they have not produced a broad push to re‑enter Afghanistan. (en.wikipedia.org)

2. Enduring support for ending the war / “thank God that’s over” sentiment
   Polls conducted during and after the withdrawal consistently show that most Americans supported the decision to pull out, even while criticizing how it was executed:

• Pew (Aug 23–29, 2021): 54% said withdrawing troops from Afghanistan was the right decision, 42% said it was wrong. (pewresearch.org)
• Marquette (Sept 7–16, 2021): 74% supported withdrawing all U.S. troops; only 26% opposed. (law.marquette.edu)
• Washington Post/ABC News poll: 77% supported the decision to withdraw all forces, across both parties, even though most disapproved of Biden’s handling. (ifpnews.com)
• Monmouth (Sept 2021): 66% approved of the decision to withdraw the U.S. presence, vs. 27% who disapproved, while opinions on Biden’s handling were much more negative. (monmouth.edu)

Follow‑up polling indicates war‑weariness and relief at being out of Afghanistan rather than a desire to re‑engage:

• Gallup (Aug 2022, one year after withdrawal) found 50% of Americans say sending troops to Afghanistan was a mistake, the highest level yet, signaling a sour retrospective view of the war itself, not nostalgia for staying. (news.gallup.com)
• A YouGov/Concerned Veterans for America poll (Aug 17–19, 2021, already after Taliban advances) found 60% supported bringing troops home, 51% wanted less U.S. military engagement abroad, only 29% favored redeploying troops to Afghanistan, and 67% said domestic issues should be prioritized over foreign policy. (cv4a.org)
• More broadly, a 2025 Reuters/Ipsos poll found 58% of Americans oppose deploying U.S. troops abroad unless facing a clear external threat, reflecting a durable reluctance to restart large ground wars. (reuters.com)

Taken together, these data points support Jason’s claim that the dominant long‑run attitude would be relief that the 20‑year war is over, even among many who think the exit was botched.

3. Fading salience of the chaotic withdrawal images
   The specific Kabul airport scenes (people clinging to planes, the Abbey Gate bombing) were intensely visible in late August 2021 but quickly receded as a top‑of‑mind concern for the general public:

• Even during the crisis, a Reuters/Ipsos poll (Aug 27–30, 2021) found only 10% of Americans named the Afghanistan war as the country’s most important problem; Afghanistan was already a low‑salience issue despite wall‑to‑wall media coverage. (ipsos.com)
• A Gallup analysis on the defense budget and Afghanistan noted that almost no Americans mentioned Afghanistan as the most important problem facing the country and that over half said they did not follow news about U.S. involvement there—evidence that, even before the final pullout, Afghanistan was “essentially out of mind” for most Americans. (news.gallup.com)
• By late 2023, AP‑NORC polling on priorities for 2024 shows foreign policy concern rising because of Ukraine, Israel‑Hamas, and China, but Afghanistan is not cited as a key issue; the top concerns are the economy and domestic matters. (apnews.com)
• Coverage of the 2022 midterms similarly emphasizes inflation, the economy, abortion, and democracy as dominant issues; CBS and other analysts explicitly predicted Afghanistan would be only “a relatively short chapter” in the campaign narrative, which proved accurate—Afghanistan was not a major voter issue. (cbsnews.com)

These patterns are what you would expect if the dramatic imagery from August 2021 had largely faded from everyday public salience: it still appears in partisan rhetoric and among veterans and activists, but it is not a sustained, central concern for most voters.

4. Countervailing evidence is limited to subgroups
   Some groups remain intensely negative about the withdrawal and continue to invoke the Kabul imagery:

• A Mission Roll Call survey of ~5,500 U.S. veterans found 73% said the withdrawal negatively affected their view of America’s legacy in the Global War on Terror. (militarytimes.com)
• Republican politicians and the Trump administration continue to highlight the “disastrous and embarrassing” withdrawal and the Abbey Gate bombing; Defense Secretary Pete Hegseth ordered yet another review in 2025. (apnews.com)
  However, Jason’s prediction was about the overall American public, not these more engaged or partisan subgroups. National polling and issue‑priority data indicate that, while these voices are loud, they do not define the majority view.

Synthesis
By late 2025:

• No 9/11‑scale attack has occurred to re‑link Afghanistan directly to mass‑casualty terrorism in the U.S.
• Majorities or pluralities consistently endorse the decision to leave Afghanistan or oppose new large troop deployments, and retrospective judgment of the war itself has turned increasingly negative—consistent with a broad sense of “good that it’s over.” (pewresearch.org)
• Afghanistan barely appears in lists of top national problems or campaign issues after 2021, signalling that the airport‑chaos imagery no longer has strong day‑to‑day salience for most Americans. (ipsos.com)

Given these points, Jason’s forecast—that the long‑term dominant public posture would be relief that the war is over and that the vivid withdrawal images would recede from public focus, absent another 9/11‑like attack—is best assessed as right.