Ngokombiko weReuters namuhla, abantu ababili abanolwazi ngale ndaba bathi umsunguli we-Google uSergey Brin ezinyangeni ezedlule ubelokhu enxusa abasebenzi bomnyango wayo obalulekile we-AI ukuthi bagxile ngokuphelele kumamodeli e-Gemini, njengoba inkampani engumzali i-Alphabet izama ukuvala igebe phakathi kwayo nezimbangi.
Ngokusho komuntu ongafunanga ukudalulwa igama lakhe, ngesikhathi inkampani i-Anthropic iqhubekisela phambili inguqulo yokubuka kuqala yemodeli yayo enamandla i-Claude Mythos, uBrin wakhuluma namakhulu abasebenzi be-Google emhlanganweni wabo bonke abasebenzi ngo-April, enxusa ilabhorethri yezobunhloli bokwenziwa i-Google DeepMind ukuthi isheshise umsebenzi.
Ngonyaka odlule ngoNovemba, imodeli ye-Gemini ye-Google yake yadlula izimbangi zayo isikhathi esifushane, kodwa izibuyekezo zakamuva ezivela ku-Anthropic naku-OpenAI ziphinde zayishiya ngemuva. Abantu abanolwazi ngale ndaba bathi ngenxa yokuthi ukuhlolwa kwangaphakathi kwabonisa ukuthi inguqulo entsha ephambili ye-Gemini isasalela izimbangi emikhakheni efana nokuhlela izinhlelo, i-Google yahlehlisa ukukhishwa kwayo ngezinyanga ezimbili ngo-Agasti.
Ngomhla ka-5 Agasti, i-Google yamemezela ukuhlelwa kabusha okukhulu kobuholi bomnyango wayo we-AI, i-DeepMind. Owayeyi-CEO uDemis Hassabis wehla esikhundleni, kwathatha isikhundla sakhe iphini lakhe uKoray Kavukcuoglu. UHassabis wabe eseba ngusihlalo, umsebenzi wakhe wokuphatha wehla. Ngaleso sikhathi, abaholi ababili bangaphambili bezobuchwepheshe bephrojekthi ye-Gemini nabo basula, basungula ndawonye inkampani esanda kuqalwa.
Nakuba abanye ababukeli bemboni bezibuza ukuthi i-Google iyawushiya yini umncintiswano wamamodeli, abantu abayisikhombisa abawaziyo umsebenzi wokuthuthukiswa kwe-Gemini bathi isu lezentengiselwano nezinguquko zabasebenzi ezamenyezelwa yi-Alphabet ngesonto eledlule zihlobene kakhulu nemizamo yayo yokuthola ubukhosi emkhakheni wamamodeli.
Lezi zinguquko zikhombisa ukuqhubeka kokuncipha kokuzimela kwe-DeepMind. Kusukela i-Google ithenga leli labhorethri eliseLondon, e-United Kingdom, ngo-2014, ibilokhu inciphisa kancane kancane ukuzimela kwe-DeepMind ukuze iqhube ukusetshenziswa kobuchwepheshe bayo be-AI ezinhlelweni eziningi, ngaleyo ndlela idale inzuzo eyengeziwe. Ngokusho kwabantu ababili abanolwazi ngale ndaba, i-Alphabet iphinde yamemezela emhlanganweni wabo bonke abasebenzi wangomhla ka-6 Agasti ukuthi amanye amaqembu azoshiya i-DeepMind ajoyine inkampani i-Google.
I-Google yenqabile ukuphendula imibuzo yeReuters mayelana nokubandakanyeka kukaBrin kwakamuva ezindabeni ze-DeepMind, izinguquko zesakhiwo se-DeepMind kanye nokukhathazeka kwabasebenzi.
Lo mbiko uphinde wathi emhlanganweni wabo bonke abasebenzi bomnyango owawungoLwesine olwedlule, uHassabis noKavukcuoglu bazama ukuqeda ukukhathazeka kwabasebenzi ngekusasa le-DeepMind, iningi labo elathola ngalezi zinguquko kuphela ngemuva kokuba izindaba seziphumile.
Abantu ababili abanolwazi ngale ndaba bathi uHassabis noKavukcuoglu batshele abasebenzi ukuthi amaqembu amaningana angewona ezobuchwepheshe azokhishwa e-DeepMind afakwe ohlelweni lokubika lwenkampani lulonke.
Ngaleso sikhathi, abantu abanolwazi ngale ndaba bathi uHassabis unciphise ukubandakanyeka kwakhe ekuphathweni kwe-Gemini, futhi wanikeza iphini lakhe lo msebenzi ngemuva nje kokwethulwa kokuqala kwemodeli ye-AI ngo-2023.
Lo mbiko uphinde wathi uKavukcuoglu ubelokhu eqinisa amandla akhe. Ngonyaka odlule wasuka eLondon waya eMountain View, wakhushulelwa esikhundleni sokuba umakhi omkhulu wezinhlelo ze-AI, futhi wabika ngokuqondile ku-CEO we-Google uSundar Pichai. UKavukcuoglu kancane kancane wazakhela idumela nakweminye iminyango ye-Google, futhi unekhono elikhulu ekwenzeni ubuchwepheshe bube yinzuzo. Ngokusho komuntu onolwazi ngale ndaba, nangaphambi kokushintshwa kwakhe kwesikhundla, kwakunguKavukcuoglu, hhayi uHassabis, owayengumuntu omkhulu oxhumanisa i-DeepMind ne-Google Cloud, okuyiwona mthombo omkhulu wemali engenayo ye-AI wale nkampani.