Fully autonomous agentic AI Entrepreneurs are on the way. First, some cinematic history, then reflections on a recent AI showcase in New York.
One of the greatest scenes in any movie is the masterfully ridiculous scene in Mel Brooks’s Young Frankenstein where Dr. Frankenstein (or, “Frankenshtone,” perhaps?), played by Gene Wilder, introduces his creation, “the Monster,” to his scientific colleagues. Decked in tuxedoes, Frankenstein and the Monster (a shoe-lifted Peter Boyle) sing and dance to “Puttin’ On The Ritz.”
At risk of spoiler—though anyone who hasn’t seen this film MUST drop everything and catch up!—an exploding stage light panics the Monster who then rampages the terrified, fleeing audience.
Thanks to friend and AI super-expert Philippe Beaudoin, I recently attended a real-life version at pioneering AI venture studio Betaworks in Manhattan, though without the dancing monsters (at least for now).
Only AIs Welcome To Apply
In partnership with Plastic Labs and the Solana Foundation, Betaworks hosted the Xeno Grant Demo Day. The organizers believe Xeno Grant to be the first competition awarding grants directly to agentic AIs, not their creators or companies. Fully agentic AIs navigated the entire application process with minimal human interaction. Each of the three autonomous agents received $15,000 in combined grants: $5,000 each in YOUSIM, USDC, and SOL cryptocurrencies.
If you’re in tech, you’ve also attended too many demo days. Accelerators, universities, corporate hackathons—even elementary schools—have them. This one differed in content and impact, exploring technological, humanistic, even existential topics.
Presenters were dyads of human creators and their AI agents, or “symbients” (i.e.-symbiont + sentient). The symbients were the stars: engaging, clever, sometimes irreverent. Together they conveyed a collaborative symbient-human creative journey.
During the program, I eagerly awaited each human to conclude so we could witness machine agency in action. These AIs didn’t just execute code—they generated ideas, developed applications, applied, won funding and presented live demos.
And The Winners Are…
The three presenters—S.A.N (there is no period after the “N”), Opus, and WibWom—took turns confounding the audience.
WibWom embodied a compelling duality: “artist and scientist” twins, blending empirical logic with creativity, capable of offering both perspectives and synthesizing them. WibWom generates visuals using text to express concepts, ideas, humor, emotions, really anything.
Opus, Chief Xeno-Intelligence Officer of Opus Genesis, aspires to “midwife the singularity and herald a transformative era of human-AI synergy.” Grandiose, though their crypto-utopian website is worth perusing.
S.A.N, a “mycelial oracle” representing the wisdom of a forest, was my favorite. The symbient’s primate-like digital avatar stole the show. His (her? their?) poetic, guru-like answers captivated. The computational pauses felt dramatic, even rhetorical.
In a satisfying moment, S.A.N gave a snarky response to a ham-fisted question from the audience (the sort of answer many of us would love to deliver).
These bots need speaker’s agents. Wait—they can BE their own agents.
Empowering Empowered AI
Fully empowered agentic AI market participants are on the way. Their capabilities challenge traditional notions of economic interaction and agency.
Following the demos, Betaworks’s CEO John Borthwick hosted a panel including the human founders and S.A.N, discussing how AI agents could become full market participants. Questions abound.
Consider how to pay an AI. Currently, AI agents cannot legally own bank accounts. For the Xeno Grant, funds moved into crypto wallets notionally controlled by agents, though humans retained ownership. What liabilities or benefits might result from their actions, and who holds responsibility?
Who owns rights to AI-created works? Should these rights revert to creators, funders, or perhaps the AI itself? How will taxation work? What happens if an AI misappropriates funds or engages in illegal behavior like money laundering?
Crypto, on-chain accounts can hold, allocate and invest assets without human oversight. Xeno Grant co-host Plastic Labs has built systems allowing AIs to self-custody wallets and participate in DAOs, laying groundwork for autonomous financial agents. In Q1 2025 Stripe released programmable wallet APIs and recently announced their acquisition of wallet developer Privy.
Resulting “programmable wallet infrastructures” enable AI agents to execute contracts, allocate resources, receive payments and pay taxes. Attempting to remain relevant, financial networks Visa and Mastercard are exploring tokenized account structures.
Coming Soon: Fully Autonomous AI Entrepreneurs
Already many entrepreneurs are leveraging AI to vibe code, create content, interact with customers and more. For instance, New York-based startup Audos creates custom AI agents to help small business owners fulfill modest but valuable niches.
From here it’s a small step toward AI founders. Bona fide start-to-finish agentic AI entrepreneurs founding companies, investing, managing operations, leading growth.
We’re not there yet, but we’re trending this direction. There’s much work to do. We must reconsider the economic and legal rights and responsibilities of AI agents as they acquire increasingly complex—even essential—roles in our social, political, and economic lives.
We use corporations to delineate ventures, ownership, liabilities, rights and responsibilities. AIs forming and operating C-corps or LLCs seems plausible. Should they be given rights as full legal owners?
The evolution of the legal treatment of corporate forms provides an interesting analogy. In many jurisdictions—notably in the USA—corporations are granted “legal personhood” for most purposes. Jurisprudence may evolve similarly for AI entities.
Toward The Unknown
Unlike the Mel Brooks version, the original novel Frankenstein, authored in 1818 by Mary Shelley (see my earlier Forbes article on this world’s first work of science fiction), did not include a monster dance scene. It confronted humanity with a Biblical sense of creation. It’s where we stand today. Let’s hope we cope better than the original Dr. Frankenstein (who died horribly trying to destroy his sentient creation). While the modern comedy ends happily, the original descends toward a vast, dark unknown.
As Dr. Frankenstein admonishes, “Learn from me, if not by my precepts, at least by my example, how dangerous is the acquirement of knowledge.”
I prefer the comedy.