E37: NYC rejects far-left candidates, new developments in lab leak theory, App Store breakup & more

conviction thresholds crossed by this episode — 65 / 45 / 15 · ideas born here show where they landed. why 65 isn't always green

1 hit · 0 partial · 0 miss — windows that closed after 2021-06-18 and up to 2021-06-25, auto-scored against price data and never hand-set. verdict · R · α

each voice's force on conviction this episode — supports and opposes, weighted exactly as the replay applied them · share = % of this episode's movement

Friedberg

2 captures · 37% of movement · 1 idea born

+32.4 / -12.1 → net +20.3

Chamath

2 captures · 34% of movement · 1 idea born

+40.7 → net +40.7

Sacks

1 capture · 16% of movement

+19.7 → net +19.7

Jason

1 capture · 13% of movement

+15.3 → net +15.3

written during extraction and stored in data/extractions/ep037.json — the auditable source of truth, including everything market-adjacent that did not earn a capture

Structurally this is a talkable-but-thin episode: ~14 minutes of pod-format navel-gazing, ~15 minutes on the Eric Adams NYC primary, ~25 minutes on the Jesse Bloom deleted-genome-sequence story, and then one genuinely market-relevant 20-minute block on the App Store antitrust bills starting at 54:07, plus a closing competitive-dynamics exchange at 1:12:23. Three new ideas, zero attaches to the E1-E19 registry (nothing on the Fed, SPACs, gig work, Nikola, GameStop, FB's ad boycott or point-of-care testing). The App Store thread is the real capture: six House Judiciary antitrust bills moved that Wednesday, the Cicilline sideloading bill would force third-party app stores and open iPhone technologies, and Jason frames the stakes with the numbers (App Store ~10% of Apple's $274B 2020 revenue at ~75% margin, 'analogous to the AWS for Amazon', with Tim Cook calling Pelosi to pump the brakes). Sacks declares himself 'blue-pilled', calls Apple's grip worse than the Windows monopoly, and argues sideloading is the only available remedy because Apple has no natural fault lines the way Amazon (spin out AWS) or Google (spin out YouTube) do; he closes the segment on AGM's 'Bad Apple' Substack with 'There is no more innovation there. They are just a gatekeeper collecting rents' and 'It is run by HR people and woke moths' (1:12:06). Chamath is on the same side and says so explicitly as an interested party -- 'No, I'm speaking my book. I completely agree... I hope these companies get broken up... And I would hope to participate in that and make a bunch of money along the way' (1:02:45) -- and rebuts the 'there's always Android' defense with the low-ARPU argument (Snapchat took three or four years as a public company to ship a decent Android app). Jason says the motive out loud: get rid of the 30% fee 'because that negatively impacts our portfolios.' Friedberg is the lone free-market red-pill, three-on-one for fifteen minutes: government breakup is a terrible precedent, regulation breeds cronyism and raises the barrier for the next challenger, Bitcoin is the proof that over-regulated markets produce their own disruptors, Shopify is the proof Amazon can be attacked without the DOJ. That argument resolves into the episode's cleanest explicit prediction (1:12:23): over the next decade, manager-run Amazon and Apple lose to founder-led Shopify, Square and Stripe -- captured with the stated 120-month framing, primary SHOP. Chamath opposes it directly with 'all four companies are going to win' (1:13:16) and offers a different mechanism instead: antitrust won't break anyone up over 'the next 10 or 15 years', it will litigate a handful of specific practices (App Store take rate, Facebook's newsfeed algorithm, Google search) and 'gum up the product infrastructure', slowing big tech 'for the next 20 years'. That 20-year framing was deliberately NOT used as idea A's horizon -- the near-term falsifiable claim is whether the 2021 bills bite, so idea A is held at the 12-month default and the long framing is logged here rather than padding the kill date. Disclosed positions: Sacks volunteers 'I have shares in Amazon and Facebook' (1:00:10) when Friedberg accuses the table of talking its book, and Jason quips he's in the process of selling them -- logged but not captured as a mention, since it is a gotcha disclosure with no directional claim attached. Deliberately not captured: the entire NYC segment, including Sacks's confident dated call 'Look, Eric Adams is going to be the next mayor of New York City' (19:48) and 'crime is the huge issue... I think it's going to reverberate throughout America for the next few years' -- no instrument, no claim on NYC real estate, tax policy, or urban recovery, so it fails the market-edge test; same for Jason's self-labeled 1:14:54 prediction that 'the pirates are assembling themselves' and cancel culture is ending. The lab leak block is pure epistemics (Sacks: the cover-up has fingerprints and 'it's going to get worse and worse', 43:26) with no tradeable edge. China decoupling was skipped as a consensus restatement -- Friedberg himself says 'there is already kind of an obvious trajectory that we're headed this way' -- but two near-misses are worth flagging for future waves: Chamath's genuinely non-consensus claim at 45:23 that bifurcation is GOOD for service exporters because a duopoly gives more pricing power than a monopoly, with Disney named, which was skipped because DIS's 12-month return would be pure noise against a pricing-power framework; and Friedberg's repeated 'vaccine printers' forecast (49:52, restated at 51:52 with the Biden infrastructure bill as the named funding catalyst) -- small bioreactors that take shipped code, 'in the next decade' -- skipped only because there was no honest public instrument in June 2021 (the pure-plays were private; MRNA's decade return has near-zero linkage to whether distributed bioproduction arrives). Era note: covid-normalcy-summer-2021 killed on 2021-06-19, six days before this episode, so Chamath's tough-winter call could not attach to it -- which is the right outcome anyway, since it is that thesis's inverse and reads as non-consensus against June 2021 reopening euphoria. Label anomalies: SPEAKER_3 has 13 turns but is not a fifth person, it is a diarization residue bucket -- at 1:09:57 it says 'We lost him on Zoom when I asked him if he's in love', and Jason is the one who asked that at 1:09:47, so that turn is Jason, but at 1:18:23 'Back at you' answers Jason, so the bucket is mixed. The one SPEAKER_3 turn with market content, 1:08:31 'This is screwing with the margins at a lot of the companies we invested. We want that take rate lowered', reads as Jason continuing his own 1:08:21 point but is not near-certain, so it was left uncaptured rather than misattributed. One merged turn: Chamath's 56:54 turn opens with 'Go ahead, Chamath.' -- Jason's handoff glued onto the front of Chamath's turn; the captured quote starts after it. Otherwise the labels are sound: Jason 120 turns, Sacks 75, Chamath 74, Friedberg 65, and content matches fingerprints throughout (Chamath dialing in from a castle in the Mediterranean, Sacks in Miami with the underdog/Craft framing, Friedberg running science corner on genomic lineages, Jason moderating with the accelerator and syndicate references). No swap, no whole-speaker merge.