'A genie that grants any wish': Sam Altman's promise about the next generation of AI

Sam Altman, CEO of OpenAI, once again shifted the frontier of the discourse on artificial intelligence with a deliberately mythological image: "We are close to creating a genie capable of granting any wish." He said this during an interview on the Relentless podcast, and the phrase was not a loose metaphor but rather the summary of how the company understands the next generation of its models.

The analogy is powerful because it shifts the axis of the conversation. Over the last few years, the debate revolved around capability: what a model can solve, how many exams it passes, how much code it writes without supervision. Altman proposes looking at the other side of the counter: what happens when the limitation ceases to be the machine and becomes the quality of what we ask for.

Altman insisted that this breakthrough demands enormous responsibility and that his priority is for the first "wishes" granted to broadly benefit society. His phrasing was precise: to bring the world to a point where "many more people can have many more wishes." That is to say, using the first round of extraordinary capability to expand access, not to concentrate it.

This nuance is what separates a technology of abundance from an elite technology. If the first uses are directed toward health, education, scientific productivity, and basic infrastructure, the compounding effect distributes power. If they are used to optimize closed markets, the genie ends up working for very few masters.

The creator of ChatGPT stressed that the range of possibilities is "incredibly broad and creative" and that it will be exciting to discover the real potential of this technology, always guided by human preferences and values. That last clause—guided by human values—is where all the difficult engineering is concentrated today: alignment, control, interpretability, and the traceability of decisions.

The industry already knows how to build astonishingly capable systems. What it has not yet reliably mastered is ensuring that this capability remains anchored to the intention of whoever uses it, especially when that intention is ambiguous, contradictory, or simply poorly formulated. The genie of the lamp, throughout literary tradition, is dangerous precisely for that reason: it grants what you ask for, not what you want.

Altman's most striking warning was not technical but psychological. "You start making wishes, the computer grants them, and you think: 'I didn't think it would work... Now what?' It's a strange feeling," he concluded.

That "now what?" describes a phenomenon that is already beginning to be documented among intensive users of advanced models: the disorientation caused by eliminating the friction from a task that previously defined a professional identity. When writing, programming, designing, or analyzing ceases to require effort, the question is no longer how to do it, but why and for what purpose.

For sectors that live on experience—casinos, gaming, digital entertainment—Altman's promise has an immediate reading. A system capable of generating mechanics, soundscapes, narrative, art, and economic balance on demand turns product design into a conversation, not a months-long cycle.

The symmetrical risk is also evident: when anyone can request a complete game, value ceases to lie in producing it and shifts entirely to judgment, brand, and distribution. The genie grants the wish; no one guarantees that the wish is worth it.