Re:Show HN: 1-Bit Bonsai, the First Commercially Viable 1-Bi...
by in-silico on Wednesday April 01, 2026 @08:50PM
(#47608679)
In theory you do lose information compared to parameters with more bits.
In practice, neural networks aren't able to store much more than 2-4 bits of useful information per parameter (regardless of the precision), so models like this are mostly getting rid of redundancy.
In practice, neural networks aren't able to store much more than 2-4 bits of useful information per parameter (regardless of the precision), so models like this are mostly getting rid of redundancy.