I-Nvidia inciphisa isiqinisekiso sokuxhaswa ngezimali kwesikhungo sedatha se-Ohio OpenAI, ngokusho kwemibiko ye-WSJ ↗
Kubikwa ukuthi iNvidia inciphisa ukusekelwa kwayo kwezezimali kwephrojekthi enkulu yesikhungo sedatha sase-Ohio i-OpenAI ibe ngaphansi kwamaRandi ayizigidi eziyizinkulungwane eziyi-120, kusukela ku-$250 billion okwakuxoxwe ngawo ngaphambilini. Lesi siqinisekiso ekuqaleni sizomboza isigaba sokuqala sephrojekthi kuphela.
Kubonakala sengathi abatshalizimali base bekhathazekile ngokuthi ingakanani ingozi yezezimali iNvidia eyayibhekene nayo. I-OpenAI isaxoxa ngesivumelwano sokuqasha esibophezelayo sesayithi eliphelele lama-gigawatt ayi-10 - ngakho-ke ukwakhiwa okukhulu akukapheli, kumane kubukeka sengathi akunamali eyanele. (Reuters)
Imithombo ithi i-Apple iqeqesha imodeli yayo ye-AI yemakethe yaseShayina ngokusekelwa yi-Alibaba ↗
Kubikwa ukuthi i-Apple iqeqeshe imodeli yolimi olukhulu oluqondene neShayina ngokusekelwa yi-Alibaba, idlulela ngale kwesu layo langaphambilini lokuncika kakhulu kumamodeli endawo esithathu.
Le ndlela inganikeza i-Apple ukulawula okwengeziwe ngezici ze-AI kumadivayisi aseShayina ngenkathi ibheka izidingo zomthetho zaseBeijing. Kuyindlela ethokozisayo yokusethwa kwezindlela ezimbili - i-Apple ithola imodeli yayo ngenkathi isahlanganisa ubuchwepheshe be-Alibaba nokunikezwa okubanzi kwe-Apple Intelligence. (Reuters)
I-US izotshela abalingani bayo ukuthi kumele ikhethe uhlangothi emncintiswaneni we-AI neChina ↗
IWashington ilungiselela ukutshela amazwe amaningi ukuthi ukubamba iqhaza kumfelandawonye wayo we-AI kungase kungahambisani nokujoyina uhlaka oluncintisanayo lwaseShayina. Umyalezo ohlongozwayo uhlose amazwe angu-35 ahambisana nomzamo oholwa yi-US mayelana namamodeli e-AI, ama-semiconductor kanye namaminerali abalulekile.
IKazakhstan isivele ibambeke phakathi ngemuva kokujoyina izinhlelo ezixhumene nezinhlangothi zombili. I-AI geopolitics isiqala ukubukeka ingasafani nomncintiswano ojwayelekile wezobuchwepheshe futhi ifana nokukhetha ukuthi yimuphi umzila wesitimela oqhutshwa umnotho wakho wedijithali... okumangalisayo kancane, ngaphandle kokuthi kubonakala sengathi akunjalo. (Reuters)
USenator weRiphabhulikhi unxusa uTrump ukuthi asekele amamodeli e-AI avulekile ase-US phakathi kwengxoxo yemboni ↗
USenator Jim Banks ufuna uhulumeni wase-US adale izikhuthazo ezinkampanini zaseMelika ezakha amamodeli e-AI avulekile, ngenkathi enciphisa ukuncika kwezinye izindlela zaseShayina ezithandwa kakhulu.
Lokho kubeka iWashington ngqo empikiswaneni yemboni. I-Meta ne-Nvidia baphikelele ekugcineni amamodeli avulekile encintisana, kuyilapho i-CEO ye-Anthropic uDario Amodei exwayise ngokuthi ukukhulula izinsimbi zamamodeli ezinamandla kungadala izingozi zokuphepha okunzima ukuziqapha. (Reuters)
Imithombo ithi uGoldman uxoxisana nabatshalizimali ngesivumelwano sokuxhasa ngezimali seNvidia ngemuva kokuthola indima ebalulekile ↗
I-Goldman Sachs iveza amabhange, izinkampani zomshuwalense, abaphathi bezimpahla kanye nabanye abatshalizimali ngohlelo olukhulu lwe-Nvidia lokuxhasa ngezimali ingqalasizinda ye-AI olubiza amaRandi ayizigidi eziyizinkulungwane ezingama-500.
I-Nvidia ithe ingavimba amadili angaba khona afinyelela ku-$125 billion, kuyilapho i-Goldman inganikeza isikweletu sangasese, imali encane kanye nokubekwa kwezikweletu. Ngokuyisisekelo, ukubalwa kwe-AI sekuqala ukubukeka njengesigaba sayo sempahla sase-Wall Street - ama-GPU ahlangabezana nezimali ezihlelekile, futhi izinkinga zibonakala ngokwazo. (Reuters)
I-Google manje izovumela abasebenzisi ukuthi basuse uphawu lwe-watermark olubonakalayo ezizukulwaneni zayo ze-AI ↗
I-Google yenza ama-watermark abonakalayo angakhethwa ezithombeni ezikhiqizwe yi-AI, amavidiyo kanye nomculo okhiqizwe ngamamodeli afaka phakathi i-Nano Banana, i-Omni kanye ne-Lyria.
Into ebalulekile: I-Google ayisusi ungqimba olungabonakali lwemvelaphi. Ama-watermark e-SynthID kanye ne-metadata ehlobene ne-C2PA kusalokhu kunjalo, okusho ukuthi imidiya ekhiqizwe isengakwazi ukuthwala izimpawu ezifundwa ngumshini ngisho noma ilebula ebonakalayo esobala inyamalala. Kuyasebenza ngokumangalisayo... futhi kuyaphikisana ngokungenakugwenywa. (TechCrunch)
Imibuzo Evame Ukubuzwa
Kungani iNvidia inciphisa isiqinisekiso sayo sezezimali sesikhungo sedatha se-OpenAI Ohio?
Kubikwa ukuthi iNvidia inciphisa ukusilela kwayo kwezezimali ibe ngaphansi kwamaRandi ayizigidi eziyizinkulungwane eziyi-120, kusukela kumaRandi ayizigidi eziyizinkulungwane ezingama-250 okwakuxoxwe ngawo ngaphambili. Isiqinisekiso esincane ekuqaleni sizomboza isigaba sokuqala sephrojekthi kuphela. Ushintsho lubonakala luhloselwe ukunciphisa ukuchayeka kweNvidia kwezezimali ngenkathi i-OpenAI iqhubeka nezingxoxo mayelana nokuqasha okubophezelayo kwesiza esigcwele sama-gigawatt ayi-10.
Iyini imodeli ye-AI yase-Apple eqondene ne-China futhi i-Alibaba ihileleke kanjani?
Kubikwa ukuthi i-Apple iqeqeshe imodeli yayo yolimi olukhulu ngqo emakethe yaseShayina, ngokusekelwa yi-Alibaba. Lokhu kunganikeza i-Apple ukulawula okukhulu izici ze-AI ezinikezwa kumadivayisi eShayina ngenkathi kusiza inkampani ukuthi isebenze ngaphakathi kwezidingo zomthetho zendawo. Ubuchwepheshe be-Alibaba kulindeleke ukuthi buhlale buyingxenye yendlela ebanzi ye-Apple Intelligence ezweni.
Uhlanga lwe-AI phakathi kwe-US ne-China lungathinta kanjani amanye amazwe?
I-US ilungiselela ukutshela amazwe ahlanganyela kumfelandawonye wayo we-AI ukuthi ukujoyina uhlaka lokuncintisana lwaseShayina kungase kungahambisani nokuhlanganyela okuqhubekayo. Impikiswano ihlanganisa izindawo ezibalulekile ngokwecebo, okuhlanganisa amamodeli we-AI, ama-semiconductor kanye namaminerali abalulekile. Ngakho-ke amazwe asebenzisana nezinhlangothi zombili, njengeKazakhstan, angabhekana nengcindezi ekhulayo yokukhetha phakathi kwezinhlelo zobuchwepheshe ezincintisanayo.
Kungani amamodeli e-AI avulekile eba yinkinga yenqubomgomo yase-US?
Amamodeli e-AI avulekile aya ngokuya eba semqoka kwezepolitiki njengoba abenzi bezinqubomgomo base-US becabanga ngendlela yokugcina amamodeli aseMelika encintisana nezindlela ezithandwa kakhulu zaseShayina. USenator Jim Banks ucele izikhuthazo zokusekela abathuthukisi base-US. Le mpikiswano ihilela nezinkinga zokuphepha, njengoba ukukhulula izisindo zamamodeli ezinamandla kungenza kube lula ukusabalalisa nokuguqula amamodeli ngenkathi kungase kwenze kube nzima ukuqapha ukusetshenziswa kwawo.
I-Goldman Sachs ibandakanyeka kanjani ezinhlelweni zokuxhasa ngezimali ingqalasizinda ye-AI ye-Nvidia?
Kubikwa ukuthi uGoldman Sachs ukhuluma namabhange, izinkampani zomshuwalense, abaphathi bezimpahla kanye nabanye abatshalizimali mayelana nohlelo lweNvidia lokuxhasa ngezimali ingqalasizinda ye-AI olubiza amaRandi ayizigidigidi ezingama-500. INvidia ithe ingavimba ukuthengiselana okungaba khona okungafika kumaRandi ayizigidigidi eziyi-125. Indima kaGoldman ingafaka phakathi isikweletu sangasese, imali encane kanye nokubekwa kwezikweletu, okubonisa ukuthi amaphrojekthi amakhulu e-AI athembela kanjani kakhulu ezakhiweni zezimali zengqalasizinda eziyinkimbinkimbi.
Ingabe izithombe ze-Google AI zisengatholakala ngemva kokususwa kwama-watermark abonakalayo?
I-Google yenza ama-watermark abonakalayo angakhethwa ezithombeni ezikhiqizwe yi-AI, ividiyo nomculo owenziwe ngamamodeli afaka phakathi i-Nano Banana, i-Omni ne-Lyria. Nokho, ukususa uphawu olubonakalayo akususi ulwazi oluvela kulo. I-Google igcina ubuchwepheshe obufana ne-SynthID kanye ne-metadata ehlobene ne-C2PA, enganikeza amasignali afundekayo ngomshini abonisa ukuthi imidiya ikhiqizwe noma icutshungulwe kusetshenziswa i-AI.