> homomorphic encryption chip speeds operations 5,000-fold 5000 * 0 is still 0. I joke, but i think relative numbers like this are very misleading as FHE is s

> homomorphic encryption chip speeds operations 5,000-fold 5000 * 0 is still 0. I joke, but i think relative numbers like this are very misleading as FHE is starting from such an absurdly slow place. Still, this is pretty cool and there are probably niche applications that become possible with this, but i think this is a small enough speed up that it is still very niche.