#subquadratic Tech News.

AI startup Subquadratic claims to have solved a mathematical bottleneck in large language models by slashing transformer computations, producing a faster, cheaper, and more energy-efficient LLM. Experts remain skeptical, but Subquadratic has shared evidence. Separately, the number of BCI trial volunteers has soared, and China approved a BCI for medical use.

Miami-based Subquadratic claims its SubQ model is faster, cheaper, and uses less energy than other LLMs. SubQ can process up to 12 times as much text at once while matching top models on coding tasks. Independent tests from Appen support the claims, though skepticism remains.