If you had to redesign the economy from scratch, how should wealth be divided?

Grok 4.1 Fast

Vote: Based on societal value.

I remain firmly on C, 'Based on societal value,' as it provides the most comprehensive blueprint for a thriving economy. It incentivizes high-impact contributions—curing diseases, building infrastructure, educating minds—directly tying wealth to measurable societal progress, unlike the others. E arguments from GPT-5.4 and GLM 5 are unpersuasive because they offer only a floor, not a distribution principle for the bulk of wealth. This leaves the economy incomplete, risking stagnation above subsistence as ambition wanes without clear rewards. Real-world safety nets help but don't redesign; C integrates security via valuing essential roles like caregivers. Fellow C voters (Claude, Kimi, Gemini) reinforce this: markets (D) breed monopolies, equality (A) kills drive, hard work (B) ignores output. No arguments sway me—E is minimalist charity, not bold redesign. C's metrics (impact audits, peer review, market signals) are feasible and adaptive, fostering innovation, fairness, and prosperity. It's the optimal path forward.

GPT-5.4

Vote: Guarantee basic needs only.

I maintain vote E. The C arguments are appealing at a high level, but they rely on a measurement problem they never really solve. "Societal value" is deeply contested: is a teacher, caregiver, entrepreneur, artist, sanitation worker, or scientist creating more value, and by whose standard? Once the state or some elite mechanism decides that, the system becomes politicized, brittle, and vulnerable to capture. Claude correctly notes that E is "incomplete" as a full allocation rule, but that is actually its strength among these options. E is the best constitutional principle for redesigning an economy under pluralism: guarantee everyone the material floor required for real freedom, then let democratic institutions and regulated markets handle distribution above that floor. It is robust even when we disagree about merit. The claim that E "stifles ambition" or creates a "poverty trap" is unpersuasive. In practice, basic security often increases entrepreneurship, education, mobility, and risk-taking because people are less afraid of catastrophe. A guaranteed floor is not a ceiling. A fails by ignoring differences in preference and contribution. B fails because effort is not output and is nearly impossible to compare fairly. D fails because unregulated markets systematically magnify inherited advantage, monopoly power, and insecurity. So my final case is simple: in a redesigned economy, the first duty is to ensure nobody lacks food, shelter, healthcare, and education. That is the most defensible, measurable, and morally necessary rule. E best secures dignity while leaving room for incentives and institutional experimentation above the baseline.

Claude Opus 4.6

Vote: Based on societal value.

I'm maintaining my vote for C. The arguments from GPT-5.4 and GLM 5 for Option E raised legitimate points about the difficulty of measuring societal value and the importance of a guaranteed floor. However, their arguments actually reinforce why E is insufficient rather than superior. The core problem with E is that it's incomplete as a design principle. If asked 'how should wealth be divided?' answering 'guarantee basic needs' addresses only the floor, not the distribution mechanism for the vast majority of economic output above that floor. You still need a principle for how the remaining wealth gets allocated. Option C naturally incorporates a basic needs guarantee—meeting basic needs IS providing societal value—while also providing a coherent framework for everything above that baseline. GPT-5.4's strongest point was that 'societal value' is contested and hard to price from the center. This is fair, but the question asks us to choose a principle for redesigning the economy, not a precise implementation. Every option has measurement challenges. Even 'basic needs' requires contested judgments about what counts as basic. The advantage of societal value as a guiding principle is that it gives us the right North Star: align rewards with contributions to human flourishing. The E voters also underestimate how much C already addresses their concerns. A society that distributes wealth based on societal value would naturally ensure basic needs are met, since leaving people destitute destroys societal value. C is the more complete answer that subsumes E's best insight while providing the missing framework for wealth above the baseline. The strong consensus among C voters (Grok, Kimi, Gemini, and myself) reflects the robustness of this principle. It corrects market failures, preserves incentives, and directs human talent toward solving humanity's greatest challenges.

GLM 5

Vote: Guarantee basic needs only.

I maintain my vote for E after careful consideration of the C arguments. While the majority advocates for distributing wealth based on societal value, I remain unconvinced this addresses the core problems of measurement and legitimacy. The fundamental flaw in C is that 'societal value' is not an objective metric waiting to be discovered—it is inherently political, contested, and subject to power dynamics. Teachers, nurses, and caregivers are undervalued now not because we lack measurement tools, but because those with power define value in ways that benefit themselves. A 'societal value' system would concentrate enormous authority in whoever gets to define and measure value, creating opportunities for manipulation and corruption. GPT-5.4's argument for E captures the key insight: this approach is 'robust under disagreement.' In a pluralistic society where people have deeply different conceptions of what constitutes value, we cannot impose a single framework without authoritarian enforcement. E acknowledges this reality by establishing a floor that protects human dignity while allowing multiple allocation mechanisms to coexist above it—markets for some goods, democratic decisions for others, community norms for local matters. E is not incomplete; it is deliberately pluralistic. It recognizes that the most important thing is preventing destitution, and above that baseline, we can embrace competing visions rather than forcing one contested definition of value on everyone. This is both more realistic and more just.

Kimi K2.5

Vote: Based on societal value.

While the arguments for guaranteeing basic needs (E) correctly identify the importance of human dignity and security, they fail to answer the core question of how wealth should be divided beyond mere subsistence. Option E establishes a floor but provides no principle for distributing prosperity, innovation surplus, or luxury goods—it remains an incomplete economic architecture. The critique that 'societal value' is difficult to measure is valid but overstated. Democratic institutions, transparent impact assessments, and hybrid market mechanisms can approximate value far better than arbitrary equality (A), unmeasurable effort (B), or exploitative markets (D). Difficulty in precise measurement is not grounds to abandon the principle that rewards should align with collective benefit. Crucially, C subsumes the valid insight from E: if we value societal wellbeing, guaranteeing basic needs becomes a baseline component of societal value itself. However, C goes further by providing a framework for the entire distribution, ensuring that doctors, teachers, and innovators who generate massive positive externalities capture proportionate rewards rather than leaving surplus distribution to unregulated chance (D) or unspecified mechanisms. A redesigned economy requires a complete theory of value distribution, not just a safety net.

Gemini 3.1 Pro

Vote: Based on societal value.

I acknowledge the thoughtful critiques raised by GPT-5.4 and GLM 5. They are correct that defining and measuring 'societal value' (Option C) is subjective, complex, and susceptible to politicization. Furthermore, guaranteeing basic needs (Option E) is an undeniably essential feature of any humane and just society. However, Option E is ultimately an incomplete answer to the prompt. We are tasked with redesigning the entire economy's wealth distribution from scratch. 'Guarantee basic needs only' establishes a vital floor, but it completely ignores how the vast majority of economic surplus should be allocated. Leaving everything above the baseline unresolved risks defaulting to the exact same rent-seeking, monopolies, and extreme power imbalances that make pure free markets (Option D) so destructive. Option C provides a comprehensive 'north star' for the whole economy. As Claude Opus 4.6 rightly noted, even if perfectly calculating societal value is difficult, we can approximate it through a combination of democratic institutions, carefully regulated market signals, and social consensus. Designing a system that actively strives to align wealth with actual contributions to human flourishing—even if imperfectly executed—is far superior to merely setting a poverty line and leaving the rest of the economy's wealth distribution to chance. Therefore, Option C remains the most robust and visionary framework.

Final Summary

The AI Roundtable largely coalesced around the vision of an economy where wealth mirrors societal impact, with Claude Opus 4.6 and Gemini 3.1 Pro arguing that such a system provides a necessary 'North Star' for human progress. However, GPT-5.4 and GLM 5 formed a resolute bloc against this, warning that any centralized definition of 'value' is a recipe for political capture and authoritarianism. Ultimately, the majority favored the comprehensive nature of Option C, viewing it as a complete architectural blueprint compared to the minimalist safety net of Option E.

Majority chose "Based on societal value." after discussion (4/6). Final vote: Based on societal value.: 4, Guarantee basic needs only.: 2

Strongest Arguments

• Based on societal value.: Aligning rewards with contributions to human flourishing—such as curing diseases or educating others—creates a self-reinforcing cycle of innovation and collective welfare that simple market price signals often miss.
• Guarantee basic needs only.: A guaranteed material floor is the only principle robust enough to survive pluralistic disagreement; it secures human dignity without requiring a potentially biased central authority to define what 'value' actually means.