a new benchmark just dropped and what it reveals is genuinely insane... AI agents can't see a quarter of the internet it tests search on live web data that changes daily so nobody can memorize the answers the winner finds 75% of what's actually out there, google's api finds 53%, and on the hardest queries even the supposed "best tool" misses almost half everyone's arguing about which model is smartest while their agents are half blind fix your search layer before your prompts