An Atlantic essay arguing that AI "does not have to be built this way" set off a familiar fight on X over whether large language models are wasteful hype or the only approach that supporters say has actually worked at scale.
The Atlantic's "Generative AI Is an Engineering Disaster" set off a sharp argument on X after critics seized on what they described as its claim that the industry got "stuck with LLMs" for reasons closer to marketing and engineering fashion than technical necessity. In a widely shared post on X, AI researcher Aidan Clark argued the piece misses the central reason labs moved toward large language models: "they are the only thing we've discovered which works."
Combined views
21.5K
14 posts, first seen 11h ago
