artificial We’re Learning Backwards: LLMs build intelligence in reverse, and the Scaling Hypothesis is bounded /u/preyneyv April 12, 2026 April 12, 2026 submitted by /u/preyneyv [link] [comments]