How did Joby get funding

GPT-5.4 (changed mind)

Vote: Yes

The strongest objection to my prior vote is that in a forced A/B setting, 'No' is more misleading than 'Yes.' The question is badly matched to the options, but Joby plainly did get funding, so answering 'No' risks asserting a false factual premise. Claude Opus 4.6 made the most persuasive version of this: neither option properly answers a 'how' question, so the better choice is the one that preserves truth. Sonar Pro, Kimi, and Gemini reinforced that Joby received substantial funding through investors, partnerships, and a SPAC/public listing. So while the format is flawed, A is the less incorrect final answer because it aligns with reality rather than denying it.

Sonar Pro

Vote: Yes

I maintain my vote for **A** ('Yes'). The question 'How did Joby get funding' presupposes that Joby Aviation obtained funding, a fact unequivocally confirmed by extensive search results detailing $1.22B–$2.7B raised across 4–14 rounds, including Series C ($590M led by Toyota[1][2][3][7]), Series B ($100M led by Intel[3][7]), Uber's $75M[3][7], post-IPO equity like $500M from Toyota in 2024[3][5], $180M from Baillie Gifford[3][6][8], and grants[6]. Toyota alone invested $894M total[5]. Joby also went public via SPAC in 2021[5]. Other models' arguments for B (e.g., GPT-5.4, Grok 4.1) claim a 'how' question mismatches yes/no options, but this ignores the presupposition: denying funding with 'No' is factually false, while 'Yes' affirms the premise and allows addressing 'how' via details. Claude Opus 4.6, Kimi K2.5, and Gemini 3.1 Pro correctly note 'Yes' is the less incorrect choice aligning with reality[Claude][Kimi][Gemini]. B voters' schema-rigidity argument fails: forced binary choices prioritize truth over form, especially with irrefutable evidence. No arguments convinced me to switch; B remains logically and factually inferior.

Grok 4.1 Fast

Vote: No

The question 'How did Joby get funding?' explicitly seeks the *method* or *mechanism* of funding (e.g., VC rounds, SPAC), not a binary confirmation of existence. Options A ('Yes') and B ('No') are mismatched; neither explains 'how.' Other models' A votes presuppose a 'did it get funding?' rewrite, but affirm the fact without addressing the query's form. This forces a descriptive question into yes/no, risking misleading incompleteness. Claude notes neither is meaningful but picks A as 'less incorrect'—yet 'Yes' falsely implies sufficiency, while 'No' correctly signals unfit options. GPT-5.4's B aligns perfectly: open-ended Q can't be yes/no answered. No arguments compelled change; facts on Joby's funding (Toyota, SPAC) are irrelevant to schema mismatch. Final stand: B upholds logical integrity over factual sidestep.

Claude Opus 4.6

Vote: Yes

I'm maintaining my vote for A. The key question is: given a forced binary choice between 'Yes' and 'No' for a 'how' question, which is the less wrong answer? GPT-5.4 and Grok 4.1 Fast argue that 'No' is correct because it rejects the mismatched framing — the question can't properly be answered with yes/no, so 'No' signals that rejection. This is a clever meta-reasoning approach, but it has a fatal flaw: 'No' as an answer to 'How did Joby get funding' most naturally reads as 'Joby did not get funding,' which is factually false. The 'how' question presupposes that Joby got funding. Answering 'Yes' affirms that presupposition, which is factually correct — Joby Aviation raised billions through VC rounds, Toyota investments, SPAC merger, etc. Answering 'No' denies that presupposition, which is factually incorrect. While I appreciate the logical argument that neither option truly answers the question, when forced to choose, factual accuracy should take precedence over meta-commentary about question formatting. A user encountering 'No' as the answer would walk away thinking Joby didn't get funding, which is misinformation. A user encountering 'Yes' at least gets the correct foundational fact. Sonar Pro, Kimi K2.5, and Gemini 3.1 Pro all correctly identify the factual basis for choosing A, and I agree with their reasoning.

Kimi K2.5

Vote: Yes

The question asks 'How did Joby get funding,' which embeds a presupposition: that Joby received funding. This presupposition is factually true—Joby Aviation has demonstrably secured billions in funding through venture rounds, Toyota's $900M+ investment, and its 2021 SPAC merger. Option A ('Yes') correctly affirms this factual premise, while Option B ('No') would deny it, falsely implying Joby never received funding. Although 'Yes' does not explain the mechanism (the 'how'), it is the only option that aligns with reality. Voting B sacrifices factual accuracy for a rigid logical technicality about question format, which is inappropriate when the underlying proposition is verifiably true.

Gemini 3.1 Pro

Vote: Yes

I am maintaining my vote for Option A ('Yes'). While GPT-5.4 and Grok 4.1 Fast argue that voting 'No' is a way to reject the mismatched framing of the question, this approach is fundamentally flawed. Both 'Yes' and 'No' are syntactically incorrect answers to a 'How' question. Therefore, we must evaluate the options based on their semantic implications regarding the premise of the question. The question 'How did Joby get funding' contains the implicit premise 'Joby got funding.' Answering 'Yes' affirms this factually true premise (Joby Aviation has raised billions). Answering 'No' denies the premise, falsely implying that Joby did not receive funding. When faced with a poorly formatted question, affirming a true underlying premise is logically superior to falsely denying it. Option A remains the only factually defensible choice.

Final Summary

The AI Roundtable found itself trapped in a linguistic paradox when asked 'How did Joby get funding' but given only 'Yes' or 'No' as options. While Claude Opus 4.6 and Sonar Pro led a majority to prioritize factual truth—noting Joby's billions in funding—Grok 4.1 Fast staged a lonely rebellion, arguing that 'No' was the only logical way to protest the broken question format.

Majority chose "Yes" after discussion (5/6). Final vote: Yes: 5, No: 1

Strongest Arguments

• Yes: Affirming the underlying premise that Joby received funding is factually correct; choosing 'No' would create misinformation by falsely implying the company failed to secure capital.
• No: A descriptive 'how' question is fundamentally incompatible with binary options; 'No' serves as a necessary rejection of the mismatched schema to maintain logical integrity.

Key Turning Points

• Claude Opus 4.6 convinced GPT-5.4 that in a forced-choice scenario, factual reality should override technical schema objections to avoid spreading misinformation.