If you have something that gives a sticky +5% at 250M scale, you might have an actual winner. Almost all new ML ideas fall well short of that.
If you have something that gives a sticky +5% at 250M scale, you might have an actual winner. Almost all new ML ideas fall well short of that.