E54: Spread trading big tech, capital allocation, Zillow's misfire, Progressives suffer losses

Using closing prices on Nov 5, 2021 (the last trading day before the podcast’s Nov 6, 2021 release) and recent prices in late November 2025, and comparing returns on an equal‑weighted basis:

1. Starting prices (Nov 5, 2021, closes)

   • MSFT: $324.57
   • GOOGL (Alphabet A): $147.83
   • AAPL: $148.17
   • AMZN: $175.95
   • META (then Facebook): $339.04
   • NFLX: $64.57
     These values come from StatMuse’s daily price tables for November 2021 for each ticker. (statmuse.com)

2. Recent prices (late Nov 2025)
   The same StatMuse pages report latest prices around late November 2025 of approximately:

   • MSFT: $492.08
   • GOOGL: $320.10
   • AAPL: $278.86
   • AMZN: $211.70
   • META: $633.72
   • NFLX: $107.57 (statmuse.com)

3. Approximate total price returns (split‑adjusted, dividends ignored)

   • MSFT: +51.6%

   • GOOGL: +116.5%
     → Average for the long side (MSFT & GOOGL): ~+84%.

   • AAPL: +88.2%

   • AMZN: +20.3%

   • META: +86.9%

   • NFLX: +66.6%
     → Average for the equal‑weighted short basket (AAPL, AMZN, META, NFLX): ~+65–66%.

   These percentages are computed directly from the StatMuse start and latest prices above.

4. Did the spread trade work?
   Interpreting Chamath’s idea in the normalized way you gave — that “MSFT and GOOG will outperform the equal‑weighted basket of AAPL, META, AMZN, NFLX” on a return basis — the data show:

   • Long side average return ≈ +84%
   • Short basket average return ≈ +65–66%
   • Relative outperformance ≈ 18–19 percentage points in favor of MSFT + GOOGL over ~4 years.

   In a standard, dollar‑neutral spread (equal notional long MSFT+GOOGL vs equal‑weighted short of the four others), that return differential translates into a positive cumulative P&L for the trade over this period.

Because MSFT+GOOGL have indeed outperformed the specified equal‑weighted basket since the prediction, the podcast forecast — that this spread would make money on a relative basis "over the coming years" — has come true so far, hence the classification of “right.”

Using an illustrative version of the trade—long Microsoft (MSFT) and Alphabet Class A (GOOGL) and short Apple (AAPL), Meta (META), Amazon (AMZN), and Netflix (NFLX), sized dollar‑neutral from the podcast date (Nov 5 2021 close) to late Nov 2025—here’s what the data show:

• Entry prices on Nov 5 2021 (split‑adjusted)

  • MSFT ≈ $324.57
  • GOOGL ≈ $147.81
  • AAPL ≈ $148.28
  • META (then Facebook) ≈ $339.04
  • AMZN ≈ $175.95
  • NFLX ≈ $64.57
    These are from StatMuse-style historical pages listing daily closes for each ticker in Nov 2021. (statmuse.com)

• Current prices (Nov 28 2025) from live market data:

  • MSFT ≈ $492.01
  • GOOGL ≈ $320.18
  • AAPL ≈ $278.85
  • META ≈ $647.95
  • AMZN ≈ $233.22
  • NFLX ≈ $107.58
  • SPY (S&P 500 ETF) ≈ $683.39

• Approximate price returns, Nov 5 2021 → Nov 28 2025 (ignoring dividends):

  • MSFT: +51.6%
  • GOOGL: +116.6%
  • AAPL: +88.1%
  • META: +91.1%
  • AMZN: +32.5%
  • NFLX: +66.6%
  • SPY: +45.9%

From these:

• Long basket (MSFT + GOOGL) average return ≈ +84%.
• Short basket (AAPL + META + AMZN + NFLX) average return ≈ +70%.
  → In a dollar‑neutral long/short, you gain roughly the difference, about +14–15% total over four years (~3–4% annualized) from the relative spread.
• The overall market (SPY) is up about +46% over the same period, starting near 468.5 in early Nov 2021 and ending around 683.4. (reddit.com)

Why this is ambiguous rather than clearly right or wrong:

1. The trade is modestly profitable but not a home run. The long basket has indeed outperformed the short basket by ~14–15 percentage points, so the basic idea of some positive spread has worked, but only mildly.
2. It has not obviously beaten a simple market long. SPY’s +46% absolute gain over the same window is far larger than the roughly +14–15% you’d earn from the market‑neutral spread, so in straightforward absolute‑return terms the spread did not “outperform the overall market.”
3. Risk and exact construction are under‑specified. Chamath spoke generically about “figuring out which ones you want to be long [and] short” in mega‑cap tech; he did not lock in this exact basket or weighting scheme, nor define what “outperform” means on a risk‑adjusted basis. Other plausible baskets (e.g., including Nvidia on the long side) could have looked very different. (fr.advfn.com)
4. The horizon isn’t fully elapsed. He gave a 4–5 year window; we’re a bit past 4 years from Nov 2021, but not yet at 5 years, so the final outcome could still change meaningfully.

Given these points, the evidence so far neither clearly confirms nor clearly falsifies the normalized prediction. The specific long/short described has been somewhat profitable on a relative basis but has not clearly delivered the strong, market‑beating, volatility‑neutral payoff the statement implies, and there is still time left in the upper end of the forecast window. Hence the label “ambiguous.”

Evidence from 2022–2025 shows Notion accelerating rather than suffering “persistent, significant growth headwinds” after Microsoft’s competitive move (Loop) and broader bundling strategy.

Key facts:

• Rapid, sustained revenue growth: Multiple sources estimate Notion’s revenue at about $67M in 2022, $250M in 2023, and $400M in 2024, implying several hundred percent growth in two years. (taptwicedigital.com) In September 2025, Notion reported over $500M in annualized revenue/ARR, confirming that growth has continued at scale, not stalled. (cnbc.com) This is inconsistent with the idea of material, persistent growth headwinds.

• Strong mid‑market and enterprise penetration: Reports state that over 50% of Fortune 500 companies have teams using Notion, and that growth is increasingly driven by large corporate clients. Named enterprise customers include Kaiser Permanente, Mitsubishi Heavy Industries, Nvidia, and Volvo Cars, and Notion’s COO says the company is “racing to keep up with enterprise demand.” (cnbc.com) This contradicts the claim that Microsoft’s distribution/bundling would materially constrain Notion’s ability to scale into mid‑market and enterprise.

• Healthy valuation for a private SaaS company: Notion’s last disclosed funding round in 2021 valued it at $10B, and there is no evidence of a down‑round or major valuation impairment; the company has not needed new primary capital since. (cnbc.com) With ~$500M ARR by 2025, that implies a roughly 20× ARR private valuation, which is broadly in line with, not obviously below, what high‑growth productivity SaaS companies might command in public markets during favorable conditions. This does not support the prediction that competitive pressure would prevent Notion from achieving a “full, standalone” type of valuation.

• Microsoft competition did not stop adoption: Microsoft announced Loop, a Notion‑like app integrated into Microsoft 365, in late 2021; it became available to organizations in 2023. (cnbc.com) Yet the same period (2023–2025) is when Notion’s revenue and enterprise usage surged, aided by rapid adoption of Notion’s AI features. (cnbc.com) The data shows coexistence and strong Notion growth, not clear evidence that Microsoft’s 80%-as-good bundled product has significantly capped Notion’s trajectory.

• Time horizon has elapsed: The prediction was made in November 2021 and explicitly concerned “the coming years.” We now have roughly four years of post‑prediction data (through 2025), enough to judge whether major, persistent headwinds emerged; they did not.

Because Notion has grown quickly, expanded meaningfully into mid‑market and enterprise, and maintained a robust valuation profile despite Microsoft’s entry, the prediction that Microsoft’s bundled competition would materially constrain Notion’s growth and scaling has not borne out. The most appropriate verdict is “wrong.”

Public filings and contemporaneous news show that Zillow did in fact reduce its workforce by roughly 25%, and that the cuts were driven primarily by shutting down its iBuyer/home‑flipping arm, Zillow Offers.

On November 2–3, 2021, Zillow announced it would wind down Zillow Offers and that this wind‑down was expected to result in approximately a 25% reduction of the company’s workforce.

• A Seattle Times/Anchorage Daily News report notes that closing the house‑flipping division “will include a reduction of Zillow’s workforce by approximately 25%.” (adn.com)
• Zillow’s own SEC filing states that the November 2, 2021 decision to wind down Zillow Offers “is expected to be complete during the second half of 2022 and result in approximately a 25% reduction of the Company’s workforce.” (sec.gov)

Multiple outlets describe this as about 2,000 employees, or ~25% of staff, tied specifically to the shutdown of Zillow Offers, i.e., the iBuyer/home‑flipping business. GeekWire reports that shutting down Zillow Offers would lead to layoffs of “approximately 25%, or about 2,000 people,” and later notes that continuing layoffs were “part of the company’s decision to end its home buying business.” (geekwire.com) RealTrends likewise describes the 2,000‑employee layoff “last November” as happening “after shuttering its iBuyer operation.” (realtrends.com) A later Wall Street Journal piece summarizes that the workforce was down 28% in 2022, largely due to these layoffs tied to winding down the algorithm‑based home‑flipping business. (wsj.com)

Relative to the November 2021 recording:

• Magnitude: The reduction was indeed about 25% of total headcount.
• Timing: The layoffs were announced immediately (early November 2021) and carried out over the following several quarters/months, which fits the prediction of cuts happening “over the next few months.”
• Primary target: The cuts were largely due to and focused on the Zillow Offers iBuyer/home‑flipping operations.

Because the real‑world outcome matches the prediction on all key points—~25% workforce reduction, driven primarily by the iBuyer unit, and executed in the months following November 2021—the prediction is best classified as right.

Republicans did win control of the U.S. House in the November 2022 midterms, but their net seat gain was only 9 seats, not the roughly 44–51 seat pickup that Sacks cited and implicitly endorsed.

In 2020, Republicans held 213 seats and Democrats 222. After the 2022 elections, Republicans held 222 seats to Democrats’ 213, for a GOP net gain of +9 seats—just four seats over the 218 needed for a majority. (en.wikipedia.org)

So while the direction (Republicans win the House) was correct, the core quantitative claim in the normalized prediction—gains on the order of 44–51 seats—was off by more than 30 seats. Because that magnitude was central to the prediction (a big "red wave" versus the very narrow majority that actually occurred), the prediction is best judged as wrong.

Chamath’s statement was explicitly conditional: if Youngkin had a successful four-year term as Virginia governor and then ran for president, he would be able to ‘crush’ the race. As of November 30, 2025, Youngkin’s first and only term is still in progress and scheduled to end in January 2026, when he will be succeeded by Abigail Spanberger as governor of Virginia. (apnews.com) He did not run for president in the 2024 cycle and repeatedly stated he would not launch a 2024 bid, focusing instead on Virginia and its legislative elections. (fox5dc.com) Current reporting frames him only as a potential or speculative contender for 2028, not a declared candidate. (foxnews.com) Because the prerequisite (him actually running for president) has not occurred and no presidential race featuring Youngkin has taken place yet, there is no outcome to compare against the prediction. It is therefore too early to judge the prediction as right or wrong.

So far, the record is mixed, and the decade is only about halfway over.

Key federal outcomes so far:

• 2020 cycle / start of the decade (before and just after the prediction): Joe Biden (Democrat) won the 2020 presidential election, and Democrats held the House with a 222–213 majority going into 2021. (en.wikipedia.org) In the 117th Congress (2021–2023), Democrats controlled both chambers (via the vice president’s tiebreak in a 50–50 Senate) and the presidency, i.e., a Democratic federal trifecta. (en.wikipedia.org)

• 2022 midterms: Republicans won control of the House (about 222–213), while Democrats slightly expanded their effective Senate majority to 51–49 (including independents caucusing with them). (en.wikipedia.org) This produced a split Congress under a Democratic president.

• 2024 elections: Donald Trump, the Republican nominee, won the 2024 presidential election. (en.wikipedia.org) Republicans also held the House with a very narrow majority (around 220 seats) and flipped the Senate to a 53–47 Republican majority. (en.wikipedia.org) From January 2025 onward, Republicans therefore have a unified hold on the presidency, House, and Senate.

Interpretation vs. the prediction:

• The prediction was that the 2020s as a whole would be a "Republican decade" in the sense of Republicans generally dominating federal elections across 2021–2030.
• The early 2020s (2021–2022) were clearly dominated by Democrats at the federal level. (en.wikipedia.org) The mid‑2020s (2025 onward) currently favor Republicans, who now control the presidency and both chambers of Congress. (en.wikipedia.org)
• Crucially, there are still major federal elections to come in 2026, 2028, and 2030, which will determine whether Republicans truly dominate the balance of the decade or whether control swings back toward Democrats.

Because the time period the prediction is about (roughly the entire 2020s) has not yet concluded, and the pattern so far is mixed—Democratic dominance early, Republican dominance as of 2025—the correctness of “this could be a Republican decade” cannot yet be determined. The fairest status is therefore **“inconclusive (too early).”

The prediction was that after the November 2021 elections, Republican politicians and media would rapidly stop alleging that U.S. elections are rigged or stolen and would largely drop 2020-election-fraud rhetoric, "overnight" and within months (by mid‑2022).

By mid‑2022, that clearly had not happened:

• Donald Trump was still explicitly claiming that the 2020 election was "rigged and stolen". On June 9, 2022, he posted on Truth Social that January 6 was about an election that was "Rigged and Stolen," and PolitiFact noted he had shown "no signs of backing down" from that claim. (politifact.com) On June 19, 2022, he again described the 2020 election as "RIGGED & STOLLEN" in a public statement, as documented by Snopes. (snopes.com) Trump is the leading Republican figure; his continued rhetoric alone contradicts the prediction.
• High-profile Republican candidates in 2022 were still centering campaigns on false stolen‑election claims. For example, Arizona GOP gubernatorial candidate Kari Lake repeatedly called the 2020 election "stolen" and "corrupt" and used these claims prominently in a June 2022 primary debate. (en.wikipedia.org)
• Systematic analyses of 2022 midterm candidates show election‑denial rhetoric remained widespread well past mid‑2022. CBS News found that, as of early November 2022, 308 of 597 Republican federal and statewide candidates had raised unfounded doubts about the 2020 election, including many who explicitly said it was stolen or rigged. (cbsnews.com) The Washington Post similarly reported nearly 300 GOP "election deniers" on the 2022 ballot, noting that the movement arising from Trump’s effort to overturn 2020 was, in many respects, even stronger two years later. (washingtonpost.com) Academic work from Stanford also characterizes 2022 as a cycle in which a large share of Republican statewide candidates questioned or denied the 2020 results. (gsb.stanford.edu)

Because prominent Republican politicians and a substantial share of GOP candidates were still actively using 2020 stolen‑election rhetoric well into and beyond mid‑2022, the forecast that such claims would quickly stop and effectively disappear was not borne out.

The prediction was that, after the November 2021 elections, the bipartisan federal infrastructure bill would "sail through" Congress without serious further obstruction from progressives.

Key facts:

• The bill in question is the Infrastructure Investment and Jobs Act (IIJA), also known as the bipartisan infrastructure bill or Bipartisan Infrastructure Law.
• It had already passed the Senate on August 10, 2021, but was being held up in the House in a standoff between moderates and progressives, with progressives tying it to the larger Build Back Better reconciliation bill. (en.wikipedia.org)
• After the off‑year elections on November 2, 2021, the House brought the IIJA to a final vote on November 5, 2021, just three days later. The bill passed 228–206, with 13 Republicans joining most Democrats. Only six progressive Democrats (“the Squad” and allies) voted no and did not succeed in blocking the bill. (theguardian.com)
• Once the House agreed to the Senate version on November 5, the bill’s congressional journey was complete; President Biden signed it into law on November 15, 2021. There was no additional congressional obstruction from progressives after that point, since no further votes on the IIJA itself were required. (en.wikipedia.org)

Interpretation relative to the prediction:

• “Sail through” / “pass quickly” after the election: The last needed congressional action (House passage of the Senate bill) happened almost immediately after the November 2 elections—on November 5—after months of prior delay. That matches the prediction that, in the wake of the elections, it would move quickly.
• “Without serious further obstruction from progressives”: Progressives had been the main source of obstruction before the elections by refusing to pass the infrastructure bill without simultaneous action on Build Back Better. After the elections, they allowed the infrastructure bill to come to a vote and did not use procedural tactics to stop it. A small group of six progressives voted "no," but the bill passed comfortably with bipartisan support, and their opposition did not create any new or effective obstruction. (commondreams.org)

Because the bill did indeed pass very quickly after the elections and there was no successful or prolonged new obstruction from progressives in that post‑election window, the prediction is best characterized as right.

Democrats did not lose anywhere near 40–50 House seats in the 2022 midterms. In the 2020 election they held 222 seats; after the 2022 elections they held 213, a net loss of 9 seats, not 40–50. (en.wikipedia.org)

Although Sacks framed this as conditional on Biden continuing to let Bernie Sanders and Elizabeth Warren "dominate the agenda," we can still evaluate the measurable part of the prediction (the size of the House losses). Regardless of how one judges that policy condition, the forecasted outcome (40–50 lost seats) clearly did not occur, so the prediction is wrong.