In a new essay published on Tuesday ”The gentle peculiarity,“Openai’s CEO Sam Altman shared his latest vision of how human experience will change in the next 15 years.
The essay is a classic example of Altman’s futurism: the submission of AGI’s promise – and arguing that his company is very close to the achievement – while at the same time degrading his arrival. Openai’s chief executive often publishes essays of this kind, clearly defining a future in which AGI disrupts the modern perception of work, energy and social contract. But often, Altman’s essays contain tips on what Openai works then.
At one point in the essay, Altman claimed that the following year, in 2026, the world will “probably see the arrival of [AI] Systems that can understand new ideas. “While this is a bit unclear, Openai executives recently reported that the company is focusing on getting AI models to find new, interesting ideas for the world.
When he announces Openai’s O3 and O4-MINI AI reasoning models In April, co -founder and President Greg Brockman said these were the first models used by scientists to create new, useful ideas.
Altman’s blog post suggests that next year, Openai itself can increase its efforts to develop AI that can create innovative ideas. Openai would certainly not be the only company to focus on this endeavor – many of the Openai competitors have shifted their focus to AI models that can help scientists to show new cases and therefore innovative discoveries for the world.
In May, Google released a paper for AlphaevolveAn AI encoding agent that the company claims to have created new approaches to complex mathematical problems. Another boot supported by former CEO of Google Eric Schmidt, Futureshouse, claims that AI Agent tool was the AI Agent tool was capable of making a real scientific discovery. In May, Anthropic launched a program to support scientific research.
If successful, these companies could automate a key part of the scientific process and possibly break into huge industries such as drug discovery, material science and other areas with science in their core.
This would not be the first time Altman has overturned his hat for Openai’s plans on a blog. In January, Altman wrote Another blog post suggesting that 2025 would be the year of factors. Then his company made a reduction in the first three AI agents: operator, deep research and codex.
But taking AI systems to create new ideas can be more difficult than their construction. The wider scientific community remains somewhat cautious about AI’s ability to produce truly original knowledge.
Earlier this year, the head of Hugging Face Thomas Wolf science wrote an essay that claims that modern AI systems cannot ask big questions, which is the key to any great scientific discovery. Kenneth Stanley, a former Openai researcher, also told TechCrunch that today’s AI models cannot create new cases.
Stanley is now building a team in Lila Sciences, a starting starting $ 200 million to create a AI -powered lab, specifically focused on acquiring AI models to find better cases. This is a difficult problem, according to Stanley, because it includes AI models a sense of what is creative and interesting.
Whether Openai really creates an AI model that is able to produce new ideas remains for appearance. Still, Altman’s essay may include something familiar – a preview of where it is likely to be directed later.
