ℏεsam

A provocative claim that AI peaked in January 2025, citing fine-tuning momentum, reinforcement learning growth, and DeepSeek R1's release as the high point, dismissing subsequent developments as hype and inferior imitations.

peak ai was jan 2025 when fine-tuning was hot, rl was taking off, and deepseek r1 was released. everything after has been hype and cheap temu knock offs.

<p>peak ai was jan 2025 when fine-tuning was hot, rl was taking off, and deepseek r1 was released. everything after has been hype and cheap temu knock offs.</p>