💸 IBridgewater ithi iBig Tech ingachitha cishe amaRandi ayizigidi ezingama-650 kwingqalasizinda ye-AI ngo-2026 ↗
I-Bridgewater empeleni iphakamisa ifulegi eliphuzi: ukwanda kwezindleko ze-AI kukhula kakhulu kangangokuthi kungase kungalawuleki. Leli nothi libeka utshalomali oluhlanganisiwe lwe-Alphabet, i-Amazon, i-Meta, kanye ne-Microsoft ku-$650B, kusukela esibalweni esincane kakhulu ngonyaka odlule. (Reuters)
Into ethakazelisayo - akusikho nje ukuthi "ngicela ama-GPU engeziwe." Yimiphumela eqondile: ingcindezi ekubuyiselweni kwemali, ukuthembela emalini yangaphandle, kanye nengozi yokuthi okunye kwalokhu kusetshenziswa kwemali akuhumushi inzuzo ngokushesha ngokwanele. Ukuchuma okusakhula... kodwa okunemiphetho ebukhali, noma kubonakala kanjalo. (Reuters)
🧑💼 I-OpenAI ibiza abeluleki ukuthi basebenze kanzima ebhizinisini labo ↗
I-OpenAI ithambekele kakhulu esigabeni "sokwenza kube ngokoqobo emsebenzini" - isebenzisana nezinkampani ezinkulu zokubonisana ukusiza izinkampani ezinkulu ukuthi zidlulele ngale kwezinhlelo zokuhlola kanye nokuhlola. Kungumdlalo wezinkampani kakhulu, kodwa empeleni, yilapho imali eningi ikhona. (TechCrunch)
Ithoni lapha ayiyona "idemo epholile" kodwa "uhlelo lokuqalisa, ukuthengwa kwempahla, ukubusa, ukuqeqeshwa, yonke i-sandwich yamaphepha." Uma uke wabuka inhlangano enkulu izama ukwamukela ubuchwepheshe obusha, uyazi ukuthi kungani iletha abantu abadala. (TechCrunch)
🧾 I-OpenAI iqinisa ubudlelwano nezinkampani ezinkulu ezinikeza izeluleko ukuze kuthuthukiswe i-AI yebhizinisi ngale kokuhlola ↗
Isinyathelo esifanayo esiyinhloko, imininingwane eyengeziwe: I-OpenAI ihlela ubudlelwano obujulile nezikhulu ezibonisana ukuze kusheshiswe ukwamukelwa kwamabhizinisi futhi kudlule isigaba esithi "sazama emnyangweni owodwa". Lona umsebenzi odingekayo ukuze kutholakale futhi kugcinwe ama-akhawunti amakhulu ezinkampani. (Reuters)
Kukhona futhi indaba yokucindezela ecashile ngezansi: uma uzoba yipulatifomu yebhizinisi ezenzakalelayo, udinga uhlelo olusebenzayo olungakusebenzisa ngezinga elikhulu, hhayi nje imodeli enhle. Ukufakwa kwamapayipi okungathandeki kubalulekile, kuyacasula. (Reuters)
🕵️♀️ Amathuluzi esithombe se-AI kumele alandele imithetho yobumfihlo, kusho izingqapheli ↗
Abalawuli bobumfihlo babuyisela ukukhiqizwa kwezithombe kanye nemiphumela efana nobuso ngaphansi kokukhanya - empeleni: uma uhlelo lwakho lungakwazi ukukhafula abantu abangokoqobo, izibopho zokuvikela idatha zisasebenza. Cha "kodwa kuyisigqoko somlingo esenziwe ngokwenziwa". (I-Register)
Lokhu okusebenzayo kuzwakala sengathi kucindezelwa kakhulu ukuthobela imithetho kubahlinzeki - ikakhulukazi mayelana nedatha yokuqeqesha, izingozi zokufana ezitholakalayo, kanye nendlela imikhiqizo esetshenziswa ngayo. Kungenye yalezo zindawo lapho ubuchwepheshe buhamba khona ngokushesha futhi imithetho ilandela khona... bese ngokuzumayo igijima. (Irejista)
🛡️ I-NVIDIA iletha ukuphepha kwe-cyber okuqhutshwa yi-AI kwingqalasizinda ebalulekile yomhlaba ↗
I-Nvidia iphakamisa indawo eyengeziwe yokuvikela nge-AI, ihlose ukusetshenziswa kwezokuphepha kwe-inthanethi okuhlobene nengqalasizinda ebalulekile. Umyalezo ucacile impela: njengoba izinhlelo zixhunywa kakhulu - futhi zisizwa nge-AI kakhulu - indawo yokuhlasela iba yinkimbinkimbi kakhulu, ngakho-ke izivikelo nazo kufanele zikhuphuke. (Igumbi Lezindaba le-NVIDIA)
Futhi iNvidia iyaqhubeka nokwelula ngale kokuthi “sithengisa ama-chips” iye kokuthi “siyindaba yesikhulumi,” okuyinto… enesifiso esikhulu, kodwa hhayi engahleliwe. Ukuphepha kungenye yezindawo ezimbalwa lapho ukusetshenziswa kwe-AI kungavunyelwa khona ngokushesha ngoba ukwesaba kuyisithako esinamandla sesabelomali. (Igumbi Lezindaba le-NVIDIA)
🚰 I-Breakingviews: I-Big Tech izoqeda ingxenye ethile yengozi yamanzi ye-AI ↗
Lokhu kufana nokugeza okubandayo: izikhungo zedatha ezintsha zingasebenzisa amanzi kahle, kodwa inkinga enkulu ukuthi kuphi - amaqoqo avame ukuhlala ezindaweni ezivele zibhekene nokucindezeleka kwamanzi. Ngakho-ke ukuzuza kokusebenza kahle kuyasiza, kodwa akususi isithiyo esiyisisekelo. (Reuters)
Impikiswano ngokuyisisekelo ithi “ukulungiswa kobuchwepheshe akusona isisombululo esiphelele.” Uma ingqalasizinda ye-AI iqhubeka nokukhula, iphenduka inkinga yezinsizakusebenza zendawo kanye nendaba yokusungula izinto ezintsha emhlabeni wonke - njengokuzama ukusebenzisa ipayipi lomlilo empompini yengadi. (Reuters)
Imibuzo Evame Ukubuzwa
Ithini isexwayiso seBridgewater ngokusetshenziswa kwezimali kwengqalasizinda ye-AI ngo-2026?
I-Bridgewater iveza ukuthi ukwanda kwe-AI capex kungase kukhule ngokwanele ukudala izinkinga ze-second order, hhayi nje ukusheshisa inqubekela phambili yemodeli. Lokhu kuphawula ukuthi i-Alphabet, i-Amazon, i-Meta, ne-Microsoft kutshalomali oluhlanganisiwe lwengqalasizinda ye-AI olucishe lube yi-$650B ngo-2026. Isixwayiso ukuthi isilinganiso singakhulisa ubungozi uma imbuyiselo ibambezeleka, ukuxhaswa ngezimali kuqina, noma isidingo sihluleka ukuhambisana nokwakhiwa.
Ukusetshenziswa kwemali enkulu yengqalasizinda ye-AI kungathinta kanjani ukuthengwa kabusha, izinzuzo, kanye nokubuyiselwa kwemali?
Uma izinkampani zinyusa ukusetshenziswa kwemali kwengqalasizinda ye-AI, zivame ukuba nokugeleza kwemali okuncane kwamahhala okutholakalayo kwembuyiselo yabanikazi bamasheya njengokuthenga kabusha kanye nezabelo. Iphuzu le-Bridgewater ukuthi leli zinga lokusebenzisa imali lingacindezela ukubuyiselwa kwemali futhi landise ukuthembela emalini yangaphandle. Uma amaphrojekthi ethatha isikhathi eside ukuguqulela ekubeni inzuzo, abatshalizimali bangase bazwele kakhulu ngezikhathi ezithile, imingcele, kanye nokuqagela kokukhokha.
Kungani ukutshalwa kwezimali kwezinye ingqalasizinda ye-AI kungase kungabi nemiphumela emihle ngokushesha?
Ukuthenga i-computing eyengeziwe akufani nokuthola inzuzo eyengeziwe kuyo. Uma izinkampani zakha amandla ngaphambi kwemali engenayo ecacile, engalinganiswa, igebe phakathi kokusetshenziswa kwemali kanye nokukhokha lingakhula. Ingozi egqanyisiwe isikhathi: ukukhuphuka kwemali kungahlala kukhuphuka, kodwa kube nemiphetho ebukhali uma ukwenza imali kungahambisani nesivinini. Emijikelezweni eminingi, inkinga akuyona ukunyamalala kwesidingo - kungukubuyiselwa kwemali okufika sekwephuzile kunalokho obekulindelwe.
Ukucindezela kwe-OpenAI nezinkampani zokubonisana kusiza kanjani amabhizinisi ukuthi adlulele ngale kwezindiza?
Inhloso ukuguqula izivivinyo "ezinhle zokubonisa" zibe ukuthunyelwa okusinda ekuthengweni, ekubuseni, ekuqeqeshweni, kanye nasekusebenzeni kwansuku zonke. Izinkampani zokubonisana zisiza izinhlangano ezinkulu ukuthi zilinganise izinhlelo zokwethula, zivumelanise ababambiqhaza, futhi ziphathe ushintsho kuyo yonke iminyango. I-Reuters kanye ne-TechCrunch zombili ziyibeka njengemisipha ye-ecosystem: ukuze ibe yipulatifomu yebhizinisi ezenzakalelayo, ukusetshenziswa kwayo ngezinga elifanele kubaluleke kakhulu njengemodeli uqobo.
Basho ukuthini abaqaphi bobumfihlo uma bethi amathuluzi esithombe se-AI asawela ngaphansi kwemithetho yobumfihlo?
Abalawuli babonisa ukuthi "ukwenziwa" akususi ngokuzenzakalelayo izibopho zokuvikela idatha uma imiphumela ibukeka njengabantu bangempela. Ukukhathazeka okusebenzayo kufaka phakathi ukuqeqeshwa kwedatha evelayo, izingozi ezizungeze ukufana okubonakalayo, kanye nendlela amathuluzi esithombe asetshenziswa ngayo emikhiqizweni. Okufanele kufundwe yingcindezi eyengeziwe yokuthobela imithetho kubahlinzeki nakubasebenzisi, ikakhulukazi lapho ubuso obungokoqobo noma imiphumela efana nomuntu ingabangela izinkinga zobumfihlo kanye nemvume.
Kungani izingozi zamanzi zesikhungo sedatha ziba yingxenye yengxoxo ye-AI?
Ngisho noma izikhungo zedatha ezintsha zithuthukisa ukusebenza kahle kwamanzi, isithiyo esikhulu singaba indawo. Impikiswano ye-Reuters Breakingviews iwukuthi amaqoqo avame ukuphelela ezifundeni esezivele zibhekene nokucindezeleka kwamanzi, okuguqula ukukhula kwe-AI kube inkinga yezinsizakusebenza zendawo. Ukusebenza kahle kuyasiza, kodwa kungase kungavimbi umthelela wokwakha ngezinga elikhulu ezindaweni ezingalungile. Ukukhetha indawo kungabalulekile njengokwenza ngcono ubuchwepheshe.