For the Republic
← Command Center / 📄 Article / 2026-03-15

Buying Both Doors

Draft Complete — Pending Author Review

Fact Check

7/10

Fact Check Report

Summary

The draft is built on a strong factual foundation. The core architecture of the argument -- dual-PAC structures, ad misdirection, the Fairshake blueprint, the Bores case study -- is well-sourced and independently verifiable. However, there are several claims that are factually wrong, misleading, or materially oversimplified. Two red flags require correction before publication. Several yellow flags need tightening. The piece's biggest vulnerability is not fabrication but compression: in squeezing complex stories into single sentences, the draft occasionally distorts what actually happened.

  • RED flags: 2
  • YELLOW flags: 7
  • BLUE flags: 3

Findings

RED Flags

"AIPAC spent $14.5 million to defeat Summer Lee in 2024 and failed."

  • Location in script: "Proof of Concept" section, paragraph on spending effectiveness
  • Issue: This is factually wrong on multiple levels. AIPAC/UDP did NOT spend $14.5 million against Summer Lee in 2024. In fact, AIPAC essentially sat out her 2024 race entirely, pulling back over concerns about the challenger's low profile and Lee's strong polling. Lee won her 2024 primary 61-31 without significant AIPAC opposition. The $14.5 million figure appears to be the amount AIPAC spent to defeat Jamaal Bowman, not Summer Lee -- and Bowman lost, so AIPAC succeeded in that case. In 2022, AIPAC spent roughly $4 million against Lee (not $14.5 million), and she barely survived that challenge by about 1,000 votes.
  • Evidence: CNN, Mondoweiss, The Intercept, WESA, and Haaretz all report AIPAC backed off Summer Lee's 2024 race. ABC News/538 confirms UDP went "four for five" in 2024 Democratic primaries with the only loss being Joanna Weiss in California. OpenSecrets and Common Dreams confirm the $14.5 million figure is associated with AIPAC spending against Bowman.
  • Recommended fix: Replace with an accurate counter-example. Options: (1) Use the Joanna Weiss loss in CA-47, where UDP spent money and its candidate still lost. (2) Use Summer Lee 2022, where AIPAC spent ~$4M and she narrowly survived by 1,000 votes. (3) Simply note that AIPAC's win rate in contested primaries is high but not perfect (UDP went 4 for 5 in 2024). The broader point -- that money doesn't mechanically determine outcomes -- is correct, but the specific example used to support it is wrong in every particular.

Hochul "crossed out the entire text" of the RAISE Act and "replaced it with industry-preferred language"

  • Location in script: "Architecture of Both Doors" section, third paragraph ("the harvest")
  • Issue: This is a materially misleading compression of events. What Hochul did was propose a redline that crossed out the entire legislative text and substituted California SB 53 language -- but this was her opening negotiating position, not the final outcome. After a public standoff, Hochul and legislative leaders reached a compromise. Hochul signed the RAISE Act on December 19, 2025. The final version was weaker than the legislature's original but stronger than Hochul's initial rewrite -- it included some provisions the legislature fought to keep, including a 72-hour critical safety incident reporting requirement. As drafted, the sentence implies Hochul unilaterally gutted the law and that was the end of it. In reality, she signed a negotiated version of the RAISE Act into law. The draft's framing turns a contested negotiation into a fait accompli.
  • Evidence: American Prospect (Dec 11, 2025) reported the redline. But TechCrunch (Dec 20, 2025), Axios (Dec 19, 2025), and City & State NY (Dec 2025) all report the subsequent negotiation and signing of a compromise version. The Governor's own press release calls it "nation-leading legislation."
  • Recommended fix: Rewrite to accurately reflect the sequence: Hochul initially proposed replacing the entire bill text with weaker California-style language after receiving industry fundraising support from Ron Conway and others. She ultimately signed a compromise version that was substantially weakened from the legislature's original -- but not the wholesale gutting her initial redline proposed. The point about industry influence on the process still stands; the specific characterization of the outcome needs correction.

YELLOW Flags

"$62 million is flowing into four Illinois Democratic primaries from AI, crypto, and AIPAC PACs"

  • Location in script: Opening section, second paragraph
  • Issue: The $62 million figure is the TOTAL spending (direct campaign contributions plus outside spending) in four Chicago-area Democratic primaries, per WBEZ. The draft attributes this entire amount to "AI, crypto, and AIPAC PACs," but only $26.9 million of that $62 million comes from crypto, AI, and pro-Israel groups. The remaining ~$35 million includes direct campaign fundraising and other outside spending. The sentence as written implies the full $62 million is industry PAC money, which overstates the figure by roughly 2.3x.
  • Context: WBEZ reports: "Nearly $62 million has poured into Chicago's four most competitive Democratic congressional primaries -- a combination of direct campaign contributions and outside spending." And separately: "portions from the crypto, AI and pro-Israel groups total $26.9 million."
  • Recommended fix: Either use the $26.9 million figure for industry PAC money, or clarify that $62 million is total spending in these races, of which $26.9 million comes from crypto, AI, and AIPAC PACs.

"over $100 million in bilateral primary spending in 2024, ads about everything except Israel, a 70% win rate in contested Democratic primaries" (AIPAC)

  • Location in script: "The Real Election" section
  • Issue: The 70% win rate figure is sourced to source-12 (which cites UDP intervening in "ten separate Democratic primaries" and winning seven). This appears to be a characterization of UDP's 2022 performance, not 2024. For 2024, ABC News/538 reports UDP went "four for five" (80%) in contested Democratic primaries where it spent independently. The "$100 million" and "ads about everything except Israel" claims check out, but the 70% win rate figure needs verification of which cycle it refers to, and the most commonly cited 2024 figure is actually better for the argument (80%, or 4/5). Separately, the phrase "bilateral primary spending" is somewhat misleading for AIPAC since AIPAC PAC endorses candidates in both parties but UDP's contested independent expenditures in primaries have been overwhelmingly focused on Democratic races.
  • Recommended fix: Either specify the election cycle for the 70% figure and source it precisely, or use the 2024 UDP figure of 4/5 (80%). The argument is actually strengthened by the higher number.

"the operative who ran Fairshake's Democratic strategy... He's now a top official at Leading the Future"

  • Location in script: "Proof of Concept" section, paragraph on Vlasto
  • Issue: "Top official" slightly overstates Vlasto's role at Fairshake and may overstate his role at Leading the Future. At Fairshake, Vlasto served as a spokesman/spokesperson and adviser, not someone who "ran" its Democratic strategy. At Leading the Future, he is described as a "co-strategist" alongside Republican operative Zac Moffatt. The core point -- that the same person bridges both operations -- is accurate and well-documented. But "ran Fairshake's Democratic strategy" implies a more senior operational role than "advised" or "served as spokesman."
  • Context: NOTUS describes Vlasto as having "previously helped advise Fairshake." Multiple outlets call him a "spokesman" for Fairshake. At Leading the Future, he is a "co-strategist" or the group is described as "run by Zac Moffatt and Josh Vlasto."
  • Recommended fix: Change "ran Fairshake's Democratic strategy" to "helped advise Fairshake's Democratic strategy" or "served as Fairshake's Democratic-side strategist." Change "top official" to "co-strategist" or "co-leader." The connection is real; the characterization just needs to match the sourcing.

"Sherrod Brown -- progressive banking committee chair, top opponent of cryptocurrency"

  • Location in script: "Proof of Concept" section
  • Issue: The characterization of Brown as "top opponent of cryptocurrency" is accurate but should be noted that this is the crypto industry's framing, not a neutral descriptor. Fairshake's Josh Vlasto used this exact phrase. Brown was Banking Committee chairman and a vocal crypto skeptic, but calling him a "top opponent" adopts the industry's own language for why they targeted him. For a piece about how industries frame their opponents, this is worth noting.
  • Context: Fairshake's own statement: "Sherrod Brown was a top opponent of cryptocurrency and thanks to our efforts, he will be leaving the Senate." Brown's actual positions involved skepticism and hearings on consumer harms and illicit use, typical of financial regulators.
  • Recommended fix: This is minor. Consider attributing the phrase ("what the industry called a 'top opponent of cryptocurrency'") or using a more neutral description ("the banking committee chair who had blocked crypto-friendly legislation").

"69-80% of Americans want more AI regulation"

  • Location in script: "Architecture of Both Doors" section
  • Issue: The range is sourced to two different polls measuring different things. The 69% figure (Quinnipiac, April 2025) measures the share who think government is "not doing enough" to regulate AI. The 80% figure (Gallup/SCSP, April-May 2025) measures support for maintaining government rules for AI safety even if it slows development. These are related but distinct questions. Combining them into a single range ("69-80%") implies they're measuring the same thing on a single scale, which they aren't.
  • Context: Both polls are from credible sources and both point in the same direction. The underlying claim -- that large majorities of Americans support more AI regulation -- is well-supported.
  • Recommended fix: Either pick one figure and cite its specific question, or note that "polls consistently show 69-80% of Americans support greater AI regulation, depending on how the question is asked."

"Fairshake... spent roughly $290 million in the 2024 cycle"

  • Location in script: "Proof of Concept" section
  • Issue: The $290 million figure requires precision. According to OpenSecrets, Fairshake PAC raised $260 million in the 2023-2024 cycle. The "$290 million" figure appears to include spending across all three entities (Fairshake + Defend American Jobs + Protect Progress) and may include some 2026 cycle fundraising overlap. Citation Needed and Wikipedia use "$290 million" for the combined network. This is close enough for editorial purposes but may be on the high end of available estimates.
  • Context: The figure is consistent with source-05 and multiple outlets. The variation depends on whether you count all three entities, what time window, and whether you count raised vs. spent.
  • Recommended fix: Add "across three aligned PACs" or "through its network of three aligned super PACs" to make clear this is the combined figure, not a single entity's spending.

"The spending ratio is roughly 6:1 anti-regulation"

  • Location in script: "Proof of Concept" section, final paragraph
  • Issue: This ratio is calculated by comparing Leading the Future ($125M) to Anthropic's Public First ($20M), yielding approximately 6:1. But the comparison excludes Meta's $65M on the anti-regulation side (which would push the ratio higher) and does not account for other pro-regulation spending beyond Anthropic. The ratio is defensible as a rough figure but the exact calculation depends on what you include. If you include Meta, the ratio is closer to 9:1 or 10:1. If Public First hits its stated $50-75M fundraising target, the ratio narrows. The "6:1" figure is the most conservative version of the anti-regulation advantage.
  • Context: Source-04 cites the 6:1 ratio based on the Leading the Future vs. Public First comparison.
  • Recommended fix: This is fine as stated if you note it refers specifically to the Leading the Future vs. Public First comparison. The ratio is conservative. No urgent change needed, but adding "at minimum" would be more accurate.

BLUE -- Verification Needed

"One startup, PerceptIn, budgeted $10,000 for compliance and spent $344,000 before shutting down"

  • Location in script: "Architecture of Both Doors" section, counterargument paragraph
  • Issue: This claim is verified through a Fortune article (Jan 30, 2026) and a Harvard Kennedy School research paper (arXiv: 2301.13454). The Fortune piece states PerceptIn "budgeted $10,000 for compliance" and the "actual bill exceeded $344,000 per deployment project." However, the draft says PerceptIn spent $344,000 "before shutting down," which slightly conflates the per-deployment compliance cost with the company's total spending and demise. The writer's notes flag this for verification against the original Harvard Kennedy School research. The host should verify: (1) whether "$344,000" was per-deployment or total compliance spend, and (2) the causal chain between compliance costs and the company's shutdown.
  • Note: The Fortune article confirms both the $10K/$344K figures and that PerceptIn "went out of business," so the core claim holds. The nuance is whether the $344K was the total or per-project figure.

"Over 1,200 state-level AI bills were introduced in 2025"

  • Location in script: "Architecture of Both Doors" section, counterargument paragraph
  • Issue: MultiState.ai's AI Legislation Tracker confirms 1,208 AI-related bills were introduced across all 50 states in 2025, with 145 enacted into law. The figure checks out. However, "AI-related" is defined broadly and includes autonomous vehicles, facial recognition, deepfake restrictions, task forces, budget items, and more -- not just the kind of comprehensive AI regulation the piece is discussing. This context might matter for the counterargument's force.
  • Note: The figure is accurate as stated. Host should decide whether the broad definition matters for the argument.

Brad Carson quote attribution and role at Public First

  • Location in script: "Architecture of Both Doors" section
  • Issue: The draft attributes the quote "We know AI isn't the first thing on every voter's mind when they go to the polls" to Brad Carson and describes him as the person "who runs the pro-regulation Public First PAC." Carson is co-founder of Public First Action (along with Chris Stewart, R-UT). The quote is verified through NBC News. Describing him as the person who "runs" the PAC is a reasonable simplification of "co-founder and co-leader." The quote is accurate. The only wrinkle: the draft calls it "Public First PAC" while the formal name is "Public First Action." Minor, but worth noting.
  • Note: Verified. The formal name is "Public First Action," not "Public First PAC."

Sources Consulted

Clean Claims

The following major factual claims in the script checked out and can be relied on:

  • Leading the Future's dual-PAC structure: $5M to Think Big, $5M to American Mission. Verified via FEC filings and NOTUS.
  • Meta's $65M bilateral PAC structure: Forge the Future (R) and Making Our Tomorrow (D). Verified via NYT, Common Dreams, Benzinga.
  • Anthropic's $20M to Public First Action: Verified via CNBC, Anthropic corporate blog, Axios.
  • Fairshake's ~$290M combined spending in 2024 cycle: Verified via multiple sources. The three-PAC structure (Fairshake/Defend American Jobs/Protect Progress) is accurately described.
  • Sherrod Brown defeat with $40M in crypto spending: Verified. Brown lost to Moreno; Defend American Jobs spent $40M+.
  • GENIUS Act stablecoin legislation passed: Verified. Signed by Trump on July 18, 2025.
  • Josh Vlasto as the human link between Fairshake and Leading the Future: Verified. Former Schumer press secretary, Fairshake adviser, now Leading the Future co-strategist.
  • AI PAC ads avoid mentioning AI: Verified by NBC News investigation. Both sides run ads about immigration, healthcare, Trump -- never AI.
  • Brad Carson quote: "We know AI isn't the first thing on every voter's mind when they go to the polls." Verified via NBC News.
  • Andreessen Horowitz dormant commerce clause argument (September 2025) becoming the backbone of Trump's December 2025 AI executive order: Verified via NPR, Center for American Progress, Harvard Law Review, and the a16z publication itself.
  • Trump executive order creating "AI Litigation Task Force": Verified via White House text and multiple legal analyses.
  • Alex Bores background: Former Palantir employee, left 2019, New York assemblymember, co-authored RAISE Act, Time 100 AI 2025. All verified.
  • Think Big spending $1.8M+ against Bores: Verified via TechCrunch (which uses the $1.8M figure in its headline) and NBC News ($1.5M+ from NBC, with additional spending documented since).
  • Joe Lonsdale / Palantir irony: Palantir co-founder's money funding attacks on former Palantir employee over his Palantir work. Verified via Fast Company, Yahoo News/CNBC.
  • Ron Conway held fundraisers for Hochul while opposing RAISE Act: Verified via American Prospect and City & State NY.
  • Citizens United: 163x increase in billionaire spending since 2010: Verified via Roosevelt Institute.
  • Over 80% of 2024 billionaire spending used channels prohibited before Citizens United: Verified via Roosevelt Institute.
  • Over 1,200 state-level AI bills introduced in 2025: Verified (1,208 per MultiState.ai).
  • AIPAC spent over $100 million in 2024 elections: Verified via Common Dreams, Sludge, OpenSecrets.