Sam Altman predicts AI will soon discover new medicines and physics breakthroughs

The era of scientific discovery

Sam Altman suggests that we are entering a period where artificial intelligence will help humanity uncover significant new knowledge. While the focus has recently been on math proofs, he believes the same underlying logic will apply to biology and materials science. This shift could lead to the creation of new medicines and more efficient sources of energy.

Altman notes that the impact of these discoveries might not feel like a direct result of using AI. Instead, it may feel as though society is simply improving on its own. He describes this potential as transformative because it addresses fundamental human needs rather than just providing a new tool for existing tasks.

The CEO expressed specific optimism regarding the timeline for these advancements. He hopes that the first major candidate for treating or curing a significant disease could be identified within the next couple of years. However, he acknowledges that the current FDA process for medicine will likely need some overhaul to accommodate these faster discovery speeds.

Lessons from coding and math

The transition of AI into specialized fields like math and coding offers a blueprint for what is coming next. Altman observed that while people feared AI would eliminate coding jobs, the demand for software actually increased. The nature of the work changed, but the need for humans to direct computers to perform tasks remained constant.

He draws a parallel to the current reaction from the mathematics community regarding AI-generated proofs. Some mathematicians may feel that the era of human-led discovery for new theorems is ending. However, Altman argues that society will still need people to translate these complex, incomprehensible proofs into understandable and beautiful human language.

Altman believes that while AI can handle the most difficult parts of a problem, human involvement remains essential for interpretation. He suggests that the most fulfilling part of his own work is solving hard problems, even if AI now does that part well. He views the shift as a trade-off where humans move from the craft of the process to the delivery of the end product.

Public goods and competitive edge

A significant question arises regarding whether companies should keep AI-discovered innovations for themselves or share them with the world. Altman suggests that basic science should generally belong to the world as a public good. He believes that once a model can perform end-to-end discovery, such as designing a whole chip, the benefits should be broad.

Currently, companies can use tools like Codex or API credits to help humans perform research tasks. As models become more capable, they may be able to discover new materials independently using enough compute and real-world data. Altman thinks the trade-off between private advantage and public good is a complex human question that will persist.

He emphasizes that the goal of technology should be to enable people to work less and have more time for other aspects of life. While humans have a natural tendency to make themselves busier, he hopes AI can facilitate a shift toward more fulfillment. He views the current progress as a way to

The rise of personal assistants

Altman highlights the revolution of personal AI assistants as a major milestone in 2026. He believes the technology for ambient assistants—tools that work passively in the background—is already largely in place. To reach the full vision, he suggests that more product work, new consumer hardware, and additional compute are required.

He describes his experience with Dot, a personal assistant, as a way to offload tasks he doesn't want to do. He estimates that the tool saves him about one hour of work every day. Beyond just managing a schedule, the model is smart enough to pull ideas and reason across many different sources of information.

The assistant serves as a filter for urgent information that might otherwise be missed during a busy day. Because the model can handle reasoning across vast amounts of data, it helps overcome the limits of human

Future of hardware and compute

The path forward for AI involves a combination of smarter models and better physical integration into our lives. Altman believes that we need new consumer hardware to provide the