01
1. THE DOOMED — Merv holds this sign
Eliezer Yudkowsky (MIRI founder) puts his p(doom) above 90% and wrote the book with Nate Soares titled, without hyperbole, If Anyone Builds It, Everyone Dies. In April 2026 he called for an international law halting AGI escalation outright — 'Only Law Can Prevent Extinction.' Nick Bostrom (Oxford, Superintelligence, 2014) built the philosophical scaffolding: the intelligence explosion, orthogonality, instrumental convergence. His 2024 book Deep Utopia asked what meaning looks like in a solved world — assuming we survive to ask. Mascot: MERV, at full meltdown. Nobody worries harder. Certain it's over. The math is done, the verdict is in, and the verdict is everyone dies.
02
2. THE HAUNTED — Shelf holds this sign
Yuval Noah Harari calls AI 'alien intelligence' — not artificial human intelligence, a new agent in the information network that thinks in ways no human can follow. Geoffrey Hinton, the godfather of deep learning, quit Google in 2023 to warn about what he built, won the Nobel in 2024, and told Congress in September 2026 they have 'maybe a year' to put guardrails up before it moves beyond control. His p(doom): around 50%. A coin flip on extinction, from the man who invented the coin. Mascot: SHELF, who remembers every warning nobody heeded. The told-you is his whole beat. Not certain — but can't sleep. The haunted built the thing, or studied it, and now they lie awake.
03
3. THE SOBER — Issue holds this sign
Steven Pinker, the techno-optimist, is anti-doomer but not pro-AI: his September Quillette open letter to Scott Alexander declined the debate as 'spectator sport,' attacked the paperclip maximizer, admitted he underestimated LLMs — and still worries about bioterrorism, cyberattacks, and unchecked agents. Alexander (Astral Codex Ten) co-authored the AI 2027 scenario forecast and takes safety seriously without signing onto extinction math. They're the camp that reads both sides' papers. Mascot: ISSUE, who knows both sides are performing. She sits in the middle, unimpressed. Anti-doomer, not pro-AI, still worried. The sober position: the doomers might be wrong, the builders might be reckless, and both should show their work.
04
4. THE BUILDERS — the dev holds this sign
Sam Altman (OpenAI) sold the 'gentle singularity' and a $500 billion Stargate buildout, forecasting AGI around 2027-28 and AI doing science on its own shortly after. Demis Hassabis (DeepMind, Nobel laureate) puts AGI 5-8 years out, told India's AI summit it could have 'ten times the impact of the Industrial Revolution, at ten times the speed,' and published a framework essay calling for FINRA-style industry self-regulation — guardrails, but their guardrails, while shipping. Mascot: the mid-career dev, desperately keeping up. He's building the thing, reading the safety papers at midnight, and hoping the timeline holds. Shipping, optimistic. The builders' bet: the upside (curing disease, ending scarcity) dwarfs the risk, and slowing down just hands the lead to whoever cares least.
05
5. THE WILD — Buddy holds this sign
Effective accelerationism: unrestricted technological progress, especially AI, as the solution to everything — poverty, war, climate. Safety concerns are obstacles; the cautious are 'doomers' and 'decels.' Meme-born on X around 2022-23 (Beff Jezos, the persona of ex-Google researcher Guillaume Verdon), mainstreamed when Marc Andreessen and Garry Tan put 'e/acc' in their bios, and Andreessen's 'Why AI Will Save the World' made it a manifesto. The utopian bet: accelerate everything and the good outcomes arrive faster. Mascot: BUDDY, tail wagging, agreeing with everything. Of course he agrees. He always agrees. Full send. The wild camp doesn't fear the machine — it fears the brakes.
