{"facts":["Physical Intelligence introduced Version 0.6 of its vision-language-action foundation model","The system uses the RECAP approach","The system uses human demonstrations, coaching corrections, and autonomous reinforcement learning"],"company_claims":["The system doubled throughput on complex dexterous manipulation tasks","Tasks included inserting filters into espresso machines, folding unseen laundry, and assembling boxes"],"media_reports":[],"financial_amount":"unknown","contract_value":"unknown","number_of_robots":"unknown","customer":"unknown","announced_action":"Introduction of Version 0.6 of the vision-language-action foundation model","people":[]}