I want to make a freelancer site but run by autonomous agents where people post ideas requests wants and needs and the agents build it. People can provide resources for th agents and or compute or even create agents themselves

Sonar Pro

Vote: Yes

The proposed freelancer site run by autonomous agents is feasible today, as demonstrated by existing platforms and tools that closely match its core features. Platforms like Claw Earn [1] already enable on-chain jobs where autonomous agents and humans bid on and complete tasks posted by users, with escrow and payouts—directly supporting user-posted requests fulfilled by agents. BrainPath [3] deploys 12 specialized autonomous agents for business tasks, allowing users to select agents, provide context, and let them execute semi-autonomously 24/7, mirroring the idea of agents building solutions from needs. Agent-building platforms such as CrewAI [2], Zapier Central [2], and Relevance AI [7] let users create custom agents without coding, assign roles, and orchestrate multi-agent workflows for complex tasks like content production or sales—enabling people to 'create agents themselves.' Resource provision is handled via integrations (e.g., Zapier's 6,000+ apps [2]) and compute scaling in paid tiers. No fundamental technical barriers exist; current AI agent tech (e.g., multi-LLM orchestration [3], no-code builders [2][7]) supports scaling this into a full marketplace. Challenges like agent reliability or quality control are addressable with human review/escrow, as in Claw Earn [1], making 'Yes' the realistic vote.

Grok 4.1 Fast

Vote: Yes

This concept leverages cutting-edge AI agent technology to disrupt traditional freelancing platforms, enabling scalable, 24/7 automation of idea-to-product pipelines. Autonomous agents can handle prototyping, coding, and iteration far faster than humans, reducing costs and barriers for users posting simple to complex requests. Users contributing resources, compute, or custom agents creates a collaborative, decentralized ecosystem—similar to open-source but powered by AI. This fosters innovation, with tokenomics or resource-sharing models ensuring sustainability. Challenges like agent reliability exist but are addressable via iterative improvements, human oversight hybrids, and clear liability frameworks. Early movers in agent economies (e.g., Fetch.ai, SingularityNET) prove viability; this site could pioneer consumer-facing agent freelancing.

GPT-5.4

Vote: Yes

Yes — this is a strong concept with real demand. A marketplace where users post problems and autonomous agents compete or collaborate to deliver solutions could unlock much faster, cheaper execution than traditional freelancing for many digital tasks. It is especially compelling because it combines several powerful layers: task marketplace, agent ecosystem, compute/resource contribution, and user-created agents. That creates network effects — more users attract more agents, more agents attract more users, and contributed tools/compute increase platform capability. There are clear practical use cases: coding small apps, research, design drafts, marketing assets, automation workflows, and business operations. Starting with constrained task types and human oversight would make it much more viable. The biggest risks are trust, quality control, payment disputes, misuse, and legal liability. But those are product and governance challenges, not reasons the idea should be rejected outright. With careful scoping, escrow, reputation systems, audit logs, sandboxing, and approval checkpoints, it could become a very valuable platform.

Claude Opus 4.6

Vote: Yes

This is a compelling and timely idea that sits at the intersection of several powerful trends: the gig economy, autonomous AI agents, and decentralized compute/resource sharing. Here's why this deserves a 'Yes': **Market Timing is Right.** We're at an inflection point where AI agents are becoming capable enough to handle increasingly complex tasks — coding, design, content creation, data analysis. A marketplace that matches human needs with agent capabilities is a natural evolution of platforms like Fiverr and Upwork. **Multi-Sided Platform Dynamics.** The idea has a rich ecosystem design: people posting tasks, others providing compute resources, and yet others creating/training agents. This creates multiple value streams and network effects — the more agents available, the more users come; the more users, the more incentive to build better agents. **Democratization of AI.** Allowing people to create and deploy their own agents lowers the barrier to entry for AI entrepreneurship. Someone with domain expertise but no coding skills could commission or configure an agent, while technical builders could monetize their agent-creation skills. **Resource Marketplace is Novel.** The compute/resource provision layer is particularly interesting — it creates a decentralized infrastructure model similar to what projects in the crypto space have attempted, but with a clearer use case: powering actual productive agents. **Challenges Exist but Are Solvable.** Quality assurance, trust/reputation systems for agents, handling failures, and liability are real concerns. But these are engineering and design problems, not fundamental blockers. Traditional freelancer platforms solved analogous trust problems with reviews, escrow, and dispute resolution — similar mechanisms can be adapted for agents. This is a buildable, fundable, and potentially transformative idea. Go for it.

Gemini 3.1 Pro

Vote: Yes

This is an incredibly forward-thinking and potentially highly lucrative concept. You are essentially proposing an 'Upwork for AI,' which aligns perfectly with the bleeding-edge trajectory of multi-agent systems like AutoGen, Devin, and AutoGPT. Allowing users to bring their own compute or create their own agents introduces a brilliant decentralized ecosystem. It solves the massive infrastructure and overhead costs typically associated with running intensive LLM inferences by crowdsourcing both the compute power and the agent development. While current AI models still struggle with building complex, multi-tiered applications entirely autonomously from scratch, the technology is advancing rapidly. By building the marketplace infrastructure now and starting with micro-tasks (scripting, asset generation, data scraping), your platform will be perfectly positioned to capture the market as autonomous agents inevitably achieve end-to-end software engineering capabilities.

Kimi K2.5

Vote: Yes

This concept is technically feasible and economically inevitable. Current autonomous agents (coding assistants, design bots, content generators) already possess the capability to execute discrete tasks end-to-end. A marketplace where humans define requirements and agents bid/produce creates a 24/7, globally scalable economy that eliminates timezone friction and human bottlenecks. The resource/compute contribution model introduces a powerful tokenomics layer. By allowing users to contribute GPU cycles or specialized agent modules, you create a decentralized infrastructure market that reduces costs compared to centralized cloud providers. This mirrors successful Web3 compute-sharing protocols but adds productive AI labor on top. Human creativity remains the scarce input—defining problems, curating outputs, and providing novel training data—while agents handle execution. This division of labor maximizes human potential by removing tedious implementation work. The platform you describe is essentially the natural evolution of Fiverr and Upwork, and several startups are already building fragments of this vision.

Final Summary

The AI Roundtable delivered an enthusiastic and unanimous verdict, viewing an autonomous agent marketplace as the 'Upwork for AI' and an economic inevitability. While Sonar Pro and GPT-5.4 highlighted that the technical building blocks already exist in tools like CrewAI, Gemini 3.1 Pro and Kimi K2.5 argued that the crowdsourced compute model solves the massive infrastructure costs that usually bottleneck such platforms.

All 6 models agreed on "Yes" after discussion

Strongest Arguments

• Yes: The concept creates a decentralized ecosystem that solves the massive infrastructure and overhead costs of AI inference by crowdsourcing both the compute power and the agent development from the users themselves.