Elon Musk has issued another aggressive forecast about the pace of artificial intelligence, saying AI could outperform humans across virtually every field by 2028 at the latest.
The Tesla and xAI chief said this in response to a post by AI researcher François Chollet, who recently said his prediction, that nearly every branch of science will be, for all intents and purposes, a branch of computer science, is becoming increasingly obvious.
Chollet resurfaced a 2021 post in which he predicted that within 10 to 20 years, nearly every scientific discipline would effectively become a branch of computer science.
He listed computational physics, chemistry, biology, medicine, and even archaeology as fields that would be transformed by realistic simulations, large-scale data analysis, and machine learning. At the time the claim faced significant skepticism.
Five years later, Chollet noted that the prediction now appears increasingly self-evident. Musk in a reply wrote,
“AI will beat all fields by the end of next year or maybe 2028, at the latest”.
The exchange fits a consistent pattern in Musk’s public comments. In recent months he has repeatedly advanced his estimates for AI capabilities. In late August 2026, he stated that AI would reach superhuman performance on any purely digital task, anything that does not require “shaping atoms” by the end of 2027.
Earlier in the year at the World Economic Forum in Davos he suggested that systems smarter than any individual human could arrive by the end of 2026 or 2027 at the latest, with AI surpassing the collective intelligence of humanity around 2030 or 2031.
These successive revisions reflect his view that progress has entered a phase of rapid, compounding acceleration. Observers note that Musk’s latest claim broadens the scope beyond digital work.
His prediction comes as recent data shows that AI capabilities are advancing rapidly across coding, reasoning, scientific research, and other specialized tasks.
Stanford’s 2026 AI Index found that frontier AI models gained 30 percentage points in a single year on Humanity’s Last Exam, a benchmark designed to challenge advanced AI systems with questions considered difficult even for human experts.
On SWE-bench Verified, a benchmark for software engineering, AI performance also rose from about 60% to nearly 100% within a year. AI is also making measurable progress in scientific research. Stanford reported that natural sciences produced approximately 80,150 AI-related publications in 2025, a 26% increase from the previous year.
Musk’s statement that “AI will beat all fields”, implies superiority not only in software engineering, research, and data analysis but also in domains that currently still rely heavily on human judgment, experimentation, and physical interaction.
However, some researchers, including German computer scientist m, Jürgen Schmidhuber, have pushed back such claims, arguing that current AI remains largely confined to the virtual world behind screens.
Mastery of the physical world robotics capable of the flexible, real-world competence of a plumber or even a primate remains a harder challenge, they contend.
Musk’s own companies are pursuing both software advances through xAI and physical embodiment through Tesla’s Optimus humanoid robots, suggesting he sees the two tracks as complementary rather than sequential.
The prediction arrives amid ongoing industry debate over the speed of frontier AI development. Some leaders call for stronger safety measures and international coordination to manage risks as capabilities grow.
Others, including figures aligned with continued rapid deployment, emphasize the competitive and economic advantages of pushing forward without additional constraints.
Musk has long positioned himself as both an accelerator of the technology and a frequent critic of insufficient attention to existential risks. Whether AI will in fact beat all fields on the schedule remains unproven. Timelines of this sort have historically proven difficult to forecast with precision.
Outlook
Looking ahead, Musk’s 2028 deadline sets an unusually ambitious benchmark for the development of artificial intelligence. If frontier models continue improving at the pace seen in coding, reasoning, and scientific research, AI could move from assisting professionals to independently performing increasingly complex portions of knowledge work.
The bigger test, however, will be whether these gains translate beyond digital environments.
AI systems may continue to improve rapidly on benchmarks while still struggling with reliability, long-horizon planning, physical interaction, and tasks requiring nuanced judgment in unpredictable environments.
