Urgent.News

What's breaking now, across thousands of outlets.

AI

Are we witnessing an intelligence explosion?

From Luis Garicano and Alex Imas: Hence AI capabilities improved 1.91 times faster after October 2025, but R&D spending grew 6.19 times and the technical workforce 3.10 times. So measured capability gains per dollar fell by approximately 69%, and measured gains per technical employee-year decreased by 38%. These figures are accounting; nothing causal can be […] The post Are we witnessing an…

In recent years, the capabilities of artificial intelligence have shown remarkable improvement. According to research by Luis Garicano and Alex Imas, AI's progress has increased 1.91 times faster since October 2025. However, the spending on research and development has exploded by 6.19 times, and the technical workforce has expanded by 3.10 times. Consequently, the efficiency of gains per dollar has decreased significantly by 69%, and gains per technical employee-year have fallen by 38%.

While the accounting analysis provided by the research cannot definitively pinpoint the cause, it highlights two potential explanations. One possibility is congestion, where the increasing budget leads to diminishing returns due to organizational challenges in rapidly scaling spending. The other explanation is that each advancement becomes more complex than the previous one, implying that the rate of improvement might require even larger input increases to maintain the same level of progress.

The findings align with the perspective of economist David Bloom and his coauthors, who suggest that ideas in AI become increasingly difficult to discover, necessitating larger input increases to sustain the rate of improvement. Nevertheless, the notion of an intelligence explosion remains unproven. The article titled "Are we witnessing an intelligence explosion?" was published on the website Marginal REVOLUTION.

Written by urgent.news from Marginal Revolution's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at marginalrevolution.com →

More in AI

More from Friday 13 November →