[–] techsystems 13d ago ↗ LiLiCorr spec decode + Mamba blocks seem really interesting for the speed up.
[–] scirob 13d ago ↗ this 3.5 Super is still below Qwen 3.8 next as they state themslevesthey also released an interesting smaller model https://news.ycombinator.com/item?id=50104922 3.5 bolt it uses 2bit quantization and beats the other small active layer models
3 comments
[ 10.8 ms ] story [ 37.4 ms ] threadthey also released an interesting smaller model https://news.ycombinator.com/item?id=50104922 3.5 bolt it uses 2bit quantization and beats the other small active layer models