Search papers, labs, and topics across Lattice.
3
0
5
4
Current spreadsheet agents are failing to meet the demands of real-world business workflows, with the best models achieving less than 35% accuracy on complex tasks.
Verification of coding agent outputs is now the bottleneck, not generation, and targeted design can significantly enhance performance while curbing reward hacking.
Current AI models are surprisingly inept at real-world data visualization tasks, failing more than half the time on a new benchmark designed to mimic enterprise workflows.
