Sam Altman claims humanity has entered the singularity

In a blog post titled 'The Gentle Singularity,' published June 10, 2025, Sam Altman declared that humanity has crossed the AI 'event horizon': 'We are past the event horizon; the takeoff has started. Humanity is close to building digital superintelligence.' ^[1]He frames this as a gradual unfolding rather than an abrupt rupture: 'This is how the singularity goes: wonders become routine, and then table stakes.' ^[2]Altman lays out a rough year-by-year timeline - 2025 for AI agents doing real cognitive work, 2026 for systems capable of surfacing novel insights, and 2027 for robots performing real-world tasks. ^[1]The economic core of his argument is that the cost of intelligence should eventually converge toward the cost of electricity, pointing to AI compute prices falling roughly 10x every twelve months - including an 87 percent price cut for OpenAI's o3-pro API relative to o1-pro. ^[4]He also cites scientists using AI reporting being 'two or three times more productive.' ^[4]

Not everyone accepts that the term applies. A Forbes analysis characterized Altman's claim of having passed the event horizon as an act of 'immense boldness and brashness,' noting disagreement across the AI field over whether current large language models indicate genuine readiness for AGI or ASI, and that the essay blurs the distinction between the two. ^[2]That skepticism was echoed in wider public reaction, where the most common technical objection was definitional - that the singularity classically requires recursive self-improvement and fundamental unpredictability, neither of which current AI systems demonstrate, with one widely-shared framing calling Altman's usage 'mathematically incoherent.' A middle-ground reading - that AI is accelerating unusually fast but has not yet crossed that threshold - also featured prominently in that discussion.

Altman does not deny that disruption is coming. He acknowledges 'there will be very hard parts like whole classes of jobs going away,' arguing the offset is that 'the world will be getting so much richer so quickly that we'll be able to seriously entertain new policy ideas we never could before.' ^[3]Critics argue the 'gentle' framing glosses over harder questions. Writer Emile P. Torres points to Altman's own earlier statements - including a past worst-case scenario framed as 'lights out for all of us' - arguing they reveal an acceptance of extinction risk that the softened language now obscures: 'Altman is a pro-extinctionist, though he's generally careful not to publicly advertise this.' ^[7]Analyst Zvi Mowshowitz went further, calling Altman's approach to alignment insufficient: 'I'm sorry, but that's contradictory, doesn't address the hard questions, isn't an answer.' ^[5]Public reaction added a distributional angle largely absent from Altman's own framing - recasting the singularity not as broad-based abundance but as a mechanism that concentrates wealth upward, with the most engaged critiques describing AI as a force 'sucking society dry of its wealth' rather than sharing gains widely.

The timing of Altman's essay has drawn its own scrutiny. AEI analyst James Pethokoukis frames it as a strategically-timed counternarrative to more alarmist messaging from Anthropic CEO Dario Amodei, who has warned more bluntly about AI-driven job losses - with commentator Ethan Mollick noting that both CEOs are now making 'very bold, very testable predictions.' ^[3]The essay also arrives against the backdrop of OpenAI's rapidly growing business, reportedly crossing $25 billion in annualized revenue by March 2026, alongside Altman's own forecast that productivity could reach 10 times 2020 levels by 2030. ^[8]Some observers go further, arguing the optimistic framing carries financial stakes of its own: coverage of the essay's market ripple effects warns that AI-token and AI-adoption valuations have already priced in aggressive timelines, leaving them exposed to sharp corrections if predicted milestones slip - a risk compounded by the fact that current systems still hallucinate and require heavy human oversight, undercutting claims that superintelligence is already arriving. ^[6]