News for hackers. Stuff that matters.
Re:Mercury 2: The fastest reasoning LLM, powered by diffusion by ainch on Tuesday February 24, 2026 @10:13PM (#47146833)
This understates the possible headroom as technical challenges are addressed - text diffusion is significantly less developed than autoregression with transformers, and Inception are breaking new ground.
Re:Mercury 2: The fastest reasoning LLM, powered by diffusion by nylonstrung on Tuesday February 24, 2026 @10:55PM (#47147120)
Very good point- if as much energy/money that's gone into ChatGPT style transformer LLMs were put into diffusion there's a good chance it would outperform in every dimension