Burry Sounds Alarm on $1 Trillion AI Spending Bubble

Noted investor Michael Burry has expressed concerns over the estimated $1 trillion being spent on artificial intelligence (AI), likening it to a 19th-century bubble. He cited an 1880 newspaper case study, 'Thought Without Language,' which observed a deaf-mute child who developed complex existential inquiries before acquiring language. This, he argues, suggests that reasoning can precede language and that large language models are built backward.

Burry claims that these models attempt to scale language into reason, resulting in primitive simulations prone to hallucination. He asserts that language should be the output of understanding, not its core engine. In contrast, 'System 2' architectures from Meta Platforms (META) and Alphabet Inc. (GOOGL) focus on reasoning-oriented approaches.

Burry draws a parallel between this AI hardware frenzy and San Francisco's 'bonanza speculators' who triggered absurd market valuations in the late 19th century. He warns against the 'parameter trap,' characterized by supply-side gluttony and massive capital misallocation, leading to a directionless waste of computing resources.

Burry points to Google's AlphaGeometry and Meta's Coconut as examples of reasoning-oriented AI research that emphasize compression and logic over brute-force text prediction. These approaches align with the 1880 case study's revelation that true comprehension transcends language entirely.