"The Takeoff Has Started": What the Accelerationists Actually Said, in Their Own Words

Headlines about AI leaders often turn long essays into one dramatic sentence. This post goes back to what five of them actually said, with the date and the context, and then gives you a method for reading predictions without being carried away or dismissing them.

What they said

Sam Altman, OpenAI, 10 June 2025, "The Gentle Singularity": "We are past the event horizon; the takeoff has started. Humanity is close to building digital superintelligence." He also writes that so far it is much less weird than it seems like it should be. Robots are not yet walking the streets.

Dario Amodei, Anthropic, January 2026, "The Adolescence of Technology": Humanity is about to be handed almost unimaginable power, and it is deeply unclear whether our social, political and technological systems have the maturity to wield it. In the same essay he repeats his earlier prediction that AI could displace half of all entry-level white-collar jobs in the next one to five years.

Elon Musk, March 2024, reported by Fox Business: AI will probably be smarter than any single human next year, and by 2029 smarter than all humans combined.

Mark Zuckerberg, Meta, 30 July 2025, "Personal Superintelligence": Meta's AI systems are showing glimpses of improving themselves, slowly but undeniably, and developing superintelligence is now in sight.

Demis Hassabis, Google DeepMind, August 2025: AGI could arrive within five to ten years and bring change "10 times bigger than the Industrial Revolution, and maybe 10 times faster" (as PC Gamer reported).

How I treat these quotes

1. They run the companies. A prediction of rapid progress also helps raise money and attract talent. That does not make it wrong, but it is a reason to read carefully.
2. Timelines slip. Musk's 2024 prediction for 2025 has not happened. Treat the dates as sentiment, not schedules.
3. The direction is still useful. Even the sceptics accept that capabilities are improving quickly, and that is enough to justify preparing.
4. Read the original. Quotes lose their conditions. Amodei's essay is mostly about risks and how to handle them, which a headline may not show.

What you can do

• Pick two of these sources and read them in full, not a summary.
• Write down one claim you agree with and one you doubt, and why.
• Watch for the pattern. The same people who call this a historic breakthrough also sell the products and raise the money. Neither fact cancels the other.

The strongest counterweight to this enthusiasm comes from people who also build the technology but are less sure about speed. See Ask Ten Leaders, Get Ten Timelines and The Decade of Agents, Not the Year.

Keep reading

• Human-Level AI: When? Ask Ten Leaders, Get Ten Timelines
• "It's the Decade of Agents, Not the Year": The Realists and Skeptics of AI
• Is AI a Bubble? Both Sides Have a Point

See the complete AI Era series

---