Ngokombiko we-《Business Insider》wanamuhla (mhla zingama-30), umsunguli we-Scale AI, manje osehola i-Meta Superintelligence Labs futhi enguSikhulu Esiphezulu se-AI, u-Alexandr Wang, ukholelwa ukuthi i-AI isinciphise kakhulu igebe lamandla phakathi kwezinkampani ezisafufusa namabhizinisi amakhulu, futhi akukaze kube lula kangaka ukuphonsela izikhulu zemboni inselelo.
UWang Tao ukusho emcimbini obuhlelwe yi-Y Combinator ukuthi i-AI isishintshe imithetho eyisisekelo yokuncintisana kwezamabhizinisi asafufusa. “Kulabo abanamaphupho, umbono kanye nesifiso esikhulu, abafisa ukwakha imikhiqizo evelele ukuze bazakhele ikusasa lomhlaba, lesi kungase kube yithuba elilodwa kuphela emlandweni wempucuko yonke.”
UWang Tao ukhumbule ukuthi ngesikhathi esungula inkampani yokuqeqesha i-AI ngo-2016, izinkampani ezisafufusa zazisasesimweni esibucayi kakhulu uma ziqhathaniswa nezikhulu zemboni. “Esikhathini esedlule, ukuncintisana bekufana noDavid ephonsela uGoliyathi inselelo. Izinkampani ezisafufusa bekufanele zihlakaniphe kakhulu, zithole indawo efanele yokungena emakethe, futhi zithole indlela yokuncintisana zinezinsiza ezimbalwa kakhulu kunabancintisana nabo. Njengoba ama-agent ahlakaniphile nobunye ubuchwepheshe be-AI buqhubeka nokuthuthuka, ngikholwa ukuthi isimo sesisondela kakhulu kuGoliyathi elwa noGoliyathi.”
I-Scale AI eyasungulwa nguWang Tao ingenye yezinkampani zokuqala ukungena emkhakheni wokulebula idatha. Iqasha izinkulungwane zabantu abasebenza ngezinkontileka emhlabeni wonke ukuze ihlunge futhi ihlele izimpendulo ze-AI zamalabhorethri amakhulu e-AI, iphinde ibambe iqhaza ekuqeqesheni amamodeli. NgoJuni wangonyaka odlule, i-Meta ne-Scale AI benza isivumelwano esibiza ama-dollar aseMelika ayizigidi eziyizinkulungwane ezingu-14.3 (Qaphela: ngokwesilinganiso samanje sokushintshisana, cishe ama-yuan ayizigidi eziyizinkulungwane ezingu-96.863). UWang Tao wabe eseshiya i-Scale AI wajoyina i-Meta.
UWang Tao wengeze ngokuthi, uma izinkampani ezisafufusa zisebenzisa ama-agent e-AI “ngendlela enesifiso esikhulu kunazo zonke”, akunzima ukwehlula amabhizinisi amakhulu asevele enenzuzo emakethe. Ama-agent e-AI angakwazi ukwenza ngokuzimela imisebenzi enezinyathelo eziningi, kusukela ekubhaleni ikhodi kuya ekusingatheni izinsizakalo zamakhasimende.