Sam Altman says superintelligence could arrive in a few thousand days

Centuries ago, people made a living walking the streets at dusk to light gas lamps. Today, nobody looks back wishing they were a lamplighter.

OpenAI CEO Sam Altman published a personal essay titled "The Intelligence Age." In it, he framed our current moment around that same transition, while setting an ambitious countdown for humanity.

A countdown to superintelligence

Altman boldly predicted that we could have superintelligence in "a few thousand days." Even if it takes slightly longer, he expressed complete confidence that we will get there.

For context, a thousand days is under three years, meaning he is talking about a timeline somewhere around the next decade.

Why is he so confident? In his words, the entire breakthrough boils down to three words: deep learning worked. More importantly, it scales predictably. The more compute and data you throw at it, the better it gets at solving difficult problems.

Altman expects the progression to roll out in stages. First, autonomous personal assistants that handle real-world tasks like coordinating healthcare. Eventually, everyone will command a virtual team of AI experts, leading to massive breakthroughs in climate restoration, space colonization, and even solving all of physics.

Chips, power, and the lamplighter

However, reaching that future comes with a massive bottleneck: physical infrastructure.

Altman warned that transitioning into the Intelligence Age requires staggering amounts of compute and energy. If we fail to build out the necessary infrastructure, compute costs will stay high, turning AI into a scarce resource fought over in conflicts and reserved mostly for the wealthy.

As for the fear of AI taking everyone's jobs, Altman offered a calm perspective. While the labor market will face severe disruptions, he argues that most jobs will evolve much slower than people anticipate, and human beings will always find meaningful ways to create value.

Why it matters

For builders and creators, this essay doesn't introduce a new tool, but it lays out the roadmap OpenAI is betting its entire future on.

We think the immediate takeaway for developers is clear: Altman is telling you not to second-guess the trajectory of the foundation models. If you believe his timeline, building specialized agents and workflow automations right now is the best bet you can make.

For the broader tech industry, this essay reads like an ideological campaign for colossal infrastructure spending. OpenAI has been actively lobbying governments and investors for massive chip and energy commitments. Framing superintelligence as an inevitable outcome allows Altman to position multi-billion-dollar power plants and data centers as a moral necessity to prevent global inequality.

It is a clever narrative move, but it also carries significant risk. The entire vision hinges on the assumption that deep learning will continue to scale without hitting diminishing returns. If the scaling laws run into a wall of energy constraints or data limits, the promise of superintelligence in a few thousand days could easily end up looking like overhyped techno-optimism.