Ben Goertzel, widely recognized as the 'godfather of AGI,' has publicly declared his refusal to cede control of artificial general intelligence development to venture capital firms. Speaking on the 'On The Margin' podcast, Goertzel emphasized that AGI is too critical for humanity to be monopolized by a single entity. Having popularized the term 'Artificial General Intelligence' through decades of influential writing and leading the development of the Sophia robot, he now champions a divergent path from industry giants. His core conviction is that the foundational code responsible for machine thinking must remain free and open-source, a philosophy that underpins his leadership of the SingularityNET blockchain project and the broader Artificial Superintelligence Alliance.

This alliance, established in 2024, represents a strategic merger of Fetch.ai, SingularityNET, and Ocean Protocol, consolidating their respective tokens into FET, although Ocean Protocol is scheduled to withdraw by the end of 2025. In stark contrast to OpenAI and Anthropic, which have secured billions in funding while maintaining proprietary models behind closed doors, Goertzel has tethered his AGI ambitions to a decentralized network owned by users. Woofun AI notes that Goertzel identifies a critical flaw in open-source initiatives that lack infrastructure: releasing code is ineffective if the computational resources required to run it are inaccessible. He argues that true utility demands deploying the first AGI on a decentralized network controlled by tens of thousands of participants across fifty different countries.

Goertzel acknowledges the inherent risks of this approach, including the potential for bad actors to fork the system for harmful purposes, yet he accepts this trade-off based on a belief in the prevalence of good actors willing to host AI systems. He posits that an open, decentralized AI ecosystem offers a safer future than an AGI arms race dominated by a few major powers. His critique extends to companies that initially espoused open ideals before pivoting to proprietary models. Woofun AI reports that Goertzel cites the legal disputes between Elon Musk and Sam Altman as evidence of how rapidly OpenAI's original mission shifted towards hardware acquisition and closed-source development. He further criticizes Dario Amodei's closed approach at Anthropic and Elon Musk's transformation from an AI doomsday theorist in 2015 to a developer of closed systems via xAI.

While admitting that the closed path of raising venture capital and seeking acquisition is simpler, Goertzel insists the open route is viable, citing Linux and the Internet as proof that decentralized systems can generate significant economic value. Currently, his business model relies on a token economy where node operators run SingularityNET and host AI processes.

However, he anticipates a strategic pivot where the AGI code remains open-source while refined products are sold for regular currency, with the blockchain functioning as an encrypted backend. Woofun AI analysis suggests this shift will see SingularityNET introducing a paid tier for businesses and advanced users next year, offering capabilities superior to current chatbots like Claude Pro or ChatGPT Pro without pursuing loss-making mass-market retail products.

Goertzel's vision extends beyond simple chatbots to a future where small groups command fleets of AI Agents to execute complex tasks. He predicts that the next wave of success will belong to those who can effectively organize and educate their 'beneficial Agent armies.' This trajectory aligns with broader industry movements where AI Agents begin conducting transactions, making payments, and managing financial services on behalf of users. Experts like Varun Kabra of Concordium and Nitya Subramanian of Para highlight the risks of outsourcing purchasing power to software, a concern Goertzel addresses by advocating for an open network economy rather than corporate clouds.

Looking toward the timeline for human-level AGI, Goertzel projects a likely arrival by 2029, stating he would not be surprised if it occurs by 2027 or even 2030. His primary concern is not the capabilities of the machines themselves, but the societal gap that may emerge between those who understand AGI and those who do not, potentially exacerbating inequality. The viability of a crypto network competing against billion-dollar tech giants remains unproven, but Goertzel is preparing for his first major test. He plans to release the first downloadable version of the new Agent Omega Claw in a few weeks, enabling users to train personal Agents to manage their lives and generate income.