A little known part of the origin story of LLMs: A lot of early Natural Language Processing research was funded by US intelligence agencies that had an interest in machine translation (and also speech recognition). This funding helped develop benchmarks, grow communities and

→ View original post on X — @andrewyng