Jev reportedly retagged about 2,300 AI research papers for $0.14
A user says the job took about 83 seconds and that they accepted Jev’s classifications in all 30 tag disagreements they manually labeled.
TLDR
A user reports using Jev to retag about 2,300 AI research papers for $0.14. They say it agreed with 75% of the previous tags, which had been processed with DeepSeek V4 Flash, and identified about 579 high-confidence topic changes. The user manually labeled 30 disagreements and accepted Jev’s classifications in all 30, then reported that all changes were applied and verified in production. They say they’re more satisfied with the classifications but still see room for improvement. In a follow-up, they describe exploring more ways to use Jev for paper discovery, with evaluations ongoing.
Combined views
1.3K
1 Source, first seen 13h ago
Jev reportedly retagged about 2,300 AI research papers for $0.14
A user says the job took about 83 seconds and that they accepted Jev’s classifications in all 30 tag disagreements they manually labeled.
TLDR
A user reports using Jev to retag about 2,300 AI research papers for $0.14. They say it agreed with 75% of the previous tags, which had been processed with DeepSeek V4 Flash, and identified about 579 high-confidence topic changes. The user manually labeled 30 disagreements and accepted Jev’s classifications in all 30, then reported that all changes were applied and verified in production. They say they’re more satisfied with the classifications but still see room for improvement. In a follow-up, they describe exploring more ways to use Jev for paper discovery, with evaluations ongoing.