Community
Voyage AI argues simple classifiers on embeddings beat Jev decision models
A Voyage AI write-up claims embedding inputs with an open or cheap embedding model and fitting a scikit-learn model on labeled data was cheaper and about 37% more accurate than Jev in its tests. The author discloses working at Voyage.
Every outlet that ran it 1 article
r/MachineLearningCommunityLead
Old-school classifiers are still generally better than Jev [P]