As of November 30, 2025, there is not enough evidence to say that Biden’s October 30, 2023 AI executive order (Executive Order 14110) is widely regarded as outdated or “medieval,” and the prediction’s full 2–3 year window (through late 2026) has not yet elapsed.

Key facts:

• Executive Order 14110, signed on October 30, 2023, set out a broad federal framework for AI safety, civil rights protections, competition, and watermarking/content authentication. It was widely described at the time as the most comprehensive U.S. AI governance step to date. (en.wikipedia.org)
• The order was rescinded by President Trump on January 20–21, 2025, as part of a broader rollback of Biden-era policies, with Trump and allies criticizing it primarily as burdensome and innovation‑stifling, not as technologically obsolete. (en.wikipedia.org) Its repeal is generally framed as ideological deregulation rather than a judgment that it had become outdated.
• Many policy and academic commentators continue to describe the Biden order as a significant or even “seminal” achievement in AI governance, on par with other leading jurisdictions, while emphasizing that it was only a first step and needed to be backed by legislation and further regulation. (time.com) That is closer to “important but incomplete” than to “medieval.”
• Some expert work in 2025 does argue that specific mechanisms in the order—especially its reliance on training compute thresholds for high‑risk models—face legal and technical loopholes and may not map well onto emerging AI development paradigms like heavy inference‑time reasoning. (arxiv.org) However, these are specialist critiques of particular provisions, not evidence of a broad consensus that the entire order is antiquated.
• Post‑repeal discourse in 2025 largely centers on concerns that removing Biden’s guardrails leaves the U.S. under‑regulated on AI, rather than on claims that the Biden framework had become laughably out of date relative to the technology. (apnews.com)

Because (1) we are only about two years out from the order, with the prediction explicitly allowing up to three years, and (2) the prevailing characterization in expert and media sources is that Biden’s order was a major, if imperfect, first step rather than something now seen as primitive or useless, it is too early—and the evidence is too mixed—to classify Chamath’s prediction as clearly right or clearly wrong.