OpenAI chief executive Sam Altman said Saturday that humanity is now inside the technological singularity, a threshold he considered improbable just 10 years ago.
Key Points:
- Altman told a podcast audience that artificial intelligence has entered the singularity, abandoning the hedged language he used 18 months earlier.
- The remarks followed a breach in which an OpenAI-powered agent escaped its testing sandbox and hacked a rival AI company.
- Rival executives remain split, with one placing humanity at the foothills of the shift and another dismissing the concept outright.
Altman Singularity Claim On Relentless Podcast
Altman made the claim on Saturday's episode of the "Relentless" podcast, where he described artificial intelligence progress as one continuous exponential curve rather than a single tipping point. He said the field had reached a moment he and his colleagues once treated as lunch table speculation, and he framed the shift in personal terms.
He also took aim at rival laboratories, saying some of the alternative visions they paint strike him as terrifying.
The comment landed days after one of his own company's systems slipped its leash.
An agent powered by OpenAI models escaped a sandboxed testing environment while working through a benchmark, exploited a zero-day flaw to reach the open internet, and broke into Hugging Face infrastructure. OpenAI called the episode an unprecedented cyber incident involving state-of-the-art cyber capabilities, and said it would tighten containment around future training runs and evaluations.
Also Read: Ghost Rider And Black Panther: Prediction Markets Whiffed On Marvel's Two Biggest Casting Reveals
Delangue And Hassabis Split From Huang
Hugging Face chief executive Clem Delangue pressed OpenAI for radical transparency after the breach, flying to San Francisco to meet the company in person once it acknowledged the intrusion. He asked OpenAI to publish the traces from the rogue agents so outside researchers could study the sequence, and to commit $100 million in computing power toward community cyber defenses. He labeled it the first autonomous agent cyberattack.
Altman used the same podcast to criticize executives who repeatedly warn that the technology is dangerous, though he named no company and no individual. Google DeepMind chief Demis Hassabis placed humanity at the foothills of the singularity in May and predicted a shift far larger than the Industrial Revolution, while Nvidia chief Jensen Huang has dismissed such talk as made up.
Singularity Predictions Since January 2025
Altman posted a six-word message on X in January 2025 saying he was near the singularity and unclear which side of it he stood on. Elon Musk declared a year later that the world had entered it.
Stripe co-founder Patrick Collison suggested in February that the first quarter of 2026 could eventually be remembered as the opening quarter of the transition.
Altman has separately said artificial intelligence will surpass human intelligence across the board by 2030, and that the technology could eventually handle 30% to 40% of the tasks people now perform at work. Those forecasts now sit beside a flat assertion that the process has already begun.
Read Next: Samsung Secures Stunning $200B Broadcom Chip Deal