🧠 Ingqungquthela ye-AI ephezulu ihoxisa ukuvinjelwa kwamaphepha avela ezinhlanganweni ezivunyelwe yi-US ngemuva kokungavunyelwa kweShayina ↗
Esinye seziphazamiso ezinkulu savela ohlangothini locwaningo, hhayi ohlangothini lomkhiqizo. Ingqungquthela ye-AI ephezulu ihoxise ukuvinjelwa kwamaphepha avela ezinhlanganweni ezigunyazwe yi-US ngemuva kokuduba kwezazi zaseShayina, okuyisikhumbuzo esicacile sokuthi ukubusa kwe-AI manje sekuphazanyiswa yizepolitiki zezwe cishe ngasikhathi sinye.
Isiqephu sonke sibonisa ukuthi le nsimu isafuna ukuziveza njengomhlaba wonke futhi ivulekile, noma ngabe i-logic yezokuphepha kwezwe iqhubeka nokungena ngomnyango. Lokho kungezwani kuya kuba nzima ukukufihla. (Reuters)
🇨🇳 Okukhethekile: I-chip entsha ye-AI yeHuawei ithola ukuthandwa yi-ByteDance, i-Alibaba ehlela ukufaka ama-oda, kusho imithombo ↗
Kubikwa ukuthi i-chip entsha ye-AI ye-Huawei ithole ukwesekwa yi-ByteDance kanye ne-Alibaba, kokubili ngama-oda okuhlela. Lokho kubalulekile ngoba i-China AI stack iyaqhubeka nokuzama ukunganciki kakhulu ku-Nvidia - hhayi kahle, hhayi ngokunyakaza okukodwa, kodwa kulezi zinselelo ezinkulu nezinohlonze.
Uma ama-oda ehluleka, iHuawei ithola okungaphezu kokuwina ekuthengisweni. Ithola ukuqinisekiswa okuvela kubathengi ababili abakhulu be-AI eShayina, uhlobo lwesignali esingabonakala ngokushesha okukhulu, noma njengoba kubonakala. (Reuters)
🇪🇺 I-EU isekela ukuvinjelwa kohlelo lokusebenza olunqunu kanye nokubambezeleka kwemithetho ebalulekile ye-AI ↗
IYurophu yahamba ngezindlela ezimbili ngesikhathi esisodwa - iBrussels yakudala. Abenzi bomthetho basekela ukuvinjelwa kwezinhlelo zokusebenza "zokususa" ngenkathi bebambezela izingxenye ezinkulu zoMthetho we-AI, okuhlanganisa nemithetho ethile yezinhlelo ezinobungozi obukhulu kanye nezibopho zokumaka okuqukethwe okukhiqizwe yi-AI.
Ngakho-ke leli qembu lisaqhubeka nokuqinisa izingozi ezizwakala zishesha futhi ziyingozi emphakathini, ngenkathi linciphisa ukuqaliswa kwezinye izindawo ngoba ukuthobela imithetho kwase kuqala ukubonakala kungalawuleki. Kunzima futhi kuhamba kancane, ngandlela thile kokubili - kufana nokushaya amabhuleki ngenkathi ushintsha imizila. (The Verge)
💬 I-Google yenza kube lula ukungenisa inkumbulo yomunye umuntu osebenzisa i-AI ku-Gemini ↗
I-Google yenze umnyakazo omncane kodwa oqondile womkhiqizo ngokwenza kube lula ukuletha inkumbulo ye-chatbot kwenye i-Gemini. Ephepheni kuyisici esilula. Empeleni, kubonakala sengathi kuwumzamo oqondile wokunciphisa izindleko zokushintsha kubantu asebekhe umlando omude kwenye indawo.
Yilapho-ke impi ikhona manje - hhayi nje ukuthi ubani onemodeli ehlakaniphe kakhulu, kodwa ongenza ubudlelwano bakho be-AI bangaphambilini bube lula. Umusho ongavamile wokuthayipha, kodwa yilapho izinto sezifinyelele khona. (The Verge)
🔞 Amashalofu e-OpenAI chatbot avusa inkanuko 'unomphela' ↗
Kubikwa ukuthi i-OpenAI imise uhlelo lwayo lwe-chatbot oluvusa inkanuko isikhathi eside ngemuva kokukhathazeka kwangaphakathi okuvela kubasebenzi kanye nabatshalizimali. Le nkampani ikhombe emibuzweni engaxazululiwe mayelana nokulimala, ukulinganisela, kanye nemiphumela yesikhathi eside yokuxhumana kwe-AI ngokocansi.
Iphinde ifanelane nokunciphisa okubanzi kokugxila ngaphakathi kwe-OpenAI, kanye nokunaka okwengeziwe kuya emikhiqizweni eyinhloko kanye nesifiso esincane sokubheja okuseceleni okuzuzisayo, okuphikisanayo, futhi okunzima ukukulawula konke ngesikhathi esisodwa. Akumangazi ngempela, kodwa kusabonakala kahle. (The Verge)
🏗️ IMicrosoft ithatha izintambo zokwandisa isikhungo sedatha se-AI saseTexas ngemuva kokuba i-OpenAI ihoxile ↗
Izindaba zengqalasizinda nazo zaqhubeka nokuzwakala. I-Microsoft yathatha izintambo zokwandisa isikhungo sedatha se-AI e-Abilene, eTexas ngemuva kokuba i-OpenAI ihoxile kuleyo ngxenye yephrojekthi, yize i-OpenAI isakha kwenye indawo ne-Oracle kanye ne-Crusoe.
Ukuhlukana akusho ukuthi kuzohlukana, impela. Kubonisa ukuthi izinhlangano ezinkulu ze-AI zithola i-modular eyengeziwe - abalingani abafanayo, ukubheja okuhlukile, ukubala okuningi, kanye nesidingo esikhulu samandla ngaphansi kwakho konke lokhu. (Izindaba ze-AP)
Imibuzo Evame Ukubuzwa
Kungani ingqungquthela ye-AI ihoxise ukuvinjelwa kwayo kwamaphepha avela ezinhlanganweni ezijezisiwe?
Lokhu kuguqulwa kwenzeke ngemuva kokungavunyelwa kwezemfundo zaseShayina, okwacindezela ingqungquthela ukuba ihoxe kwinqubomgomo ebhekwa kabanzi njengenecala lezepolitiki. Lesi siqephu sibonisa ukuthi ucwaningo lwe-AI lungangena ngokushesha kangakanani kwezepolitiki yezwe. Siphinde siphakamise ukuthi izingqungquthela ezinkulu zisafuna ukugcina ubuwena bomhlaba wonke, obuvulekile, ngisho noma izinkinga zokuphepha kwezwe ziqhubeka nokubumba imithetho yokubamba iqhaza.
Kusho ukuthini ukuthola inzalo ku-Huawei ku-ByteDance kanye ne-Alibaba nge-China AI stack?
Kubonisa okungaphezu kokuthengiswa okuvamile kwama-chip. Intshisekelo evela kubathengi ababili abakhulu be-AI baseShayina iphakamisa ukuthi ihadiwe yeHuawei iphathwa ngokungathi sína njengengxenye yomzamo waseShayina wokunciphisa ukuncika kwiNvidia. Emijikelezweni eminingi yengqalasizinda ye-AI, ukuqinisekiswa okuvela kumakhasimende amakhulu kungathonya ukwamukelwa okubanzi, ikakhulukazi lapho ezinye izindlela zasendaweni zisazama ukufakazela ukuthi zingakwazi ukubhekana nesidingo sokukhiqiza.
Kungani i-EU ivimbela izinhlelo zokusebenza zokususa izikhangiso ngenkathi ibambezela izingxenye zoMthetho we-AI?
Lezi zinyathelo ezimbili zibonisa izinto eziza kuqala ezahlukene ngaphakathi komthethonqubo we-AI. Abenzi bomthetho babonakala bekulungele ukuthatha isinyathelo ngokushesha ezingozini ezizwakala zishesha, zibonakala, futhi zilimaza umphakathi, njengokususa izinhlelo zokusebenza. Ngesikhathi esifanayo, izidingo zoMthetho we-AI ezibanzi zibonakala zihamba kancane ngoba ukuthobela izinhlelo ezinobungozi obukhulu kanye nezibopho zokumaka ama-watermark kwakunzima ukufeza isikhathi sokuqala.
Ukungenisa inkumbulo ye-chatbot kwenye i-Gemini kushintsha kanjani umncintiswano?
Kunciphisa izindleko zokushintshanisa kubasebenzisi asebevele bakhe imilando emide yengxoxo kwenye indawo. Lokho kushintsha umncintiswano womkhiqizo we-AI usuke empendulweni eyodwa yemodeli bese ubheke ekuqhubekeni, ukulula, kanye nokugcinwa. Ngokwezinto ezingokoqobo, amapulatifomu e-AI awancintisani nje kuphela ngobuhlakani, kodwa nangokuthi abasebenzisi bangakwazi yini ukuthwala izintandokazi zabo, imikhuba, kanye nomongo kanye nabo.
Kungani i-OpenAI imise izinhlelo zayo ze-chatbot ezivusa inkanuko isikhathi esingaziwa?
Lesi sihloko sikhomba ukukhathazeka kwangaphakathi mayelana nokulimala, ukulinganisela, kanye nemiphumela yesikhathi eside yokusebenzisana kwe-AI ngokocansi. Lokho kwenza isinqumo singabonakali njengokumiswa komkhiqizo wesikhashana kodwa sifane nokuhoxa okunesu esigabeni esiyingozi. Siphinde sifanelane nephethini ebanzi yokubeka phambili imikhiqizo ye-AI eyinhloko kunokubheja okuphambeneyo okunzima ukukulawula.
Ingabe iMicrosoft ithatha iphrojekthi yesikhungo sedatha saseTexas kusho ukuthi ukubambisana kwayo ne-OpenAI kuyawohloka?
Akunjalo ngempela. Ushintsho lubukeka njengokulungiswa kwe-modular ngaphakathi kwenethiwekhi ebanzi yokubambisana kunokuphuka okuhlanzekile. I-OpenAI isaqhubeka nokufuna ingqalasizinda kwenye indawo ne-Oracle ne-Crusoe, kuyilapho i-Microsoft iqhubekela phambili nalokhu kwandiswa okuthile. Ezivumelwaneni ezinkulu zengqalasizinda ye-AI, abalingani bavame ukuhlala behlelekile ngenkathi benza ukubheja okuhlukile kumasayithi ngamanye kanye namaphrojekthi okubala.