🤝 I-Microsoft yembula ukuthuthukiswa kwe-AI, yethula i-Copilot Cowork kumakhasimende afinyelela kusenesikhathi ↗
I-Microsoft iqhubekisele phambili i-Copilot endaweni yamamodeli amaningi, okuzwakala sengathi yisihloko esiyinhloko manje - hhayi imodeli eyodwa ewinayo, kodwa eziningana zithungwe ndawonye. Ukugeleza kwayo okusha kwe-"Critique" kwenza i-GPT ibhale impendulo ngenkathi uClaude eyihlola ukuthi inembile yini futhi ikhwalithi, kanti i-Microsoft ithi ifuna ukuthi leyo loop yokubuyekeza ibe yindlela yesibili kamuva.
Iphinde yethula i-"Council" yokuqhathaniswa kwamamodeli eceleni futhi yandisa ukufinyelela ku-Copilot Cowork ngohlelo lwe-Frontier lokufinyelela kusenesikhathi. Inkulumo ilula ngokwanele: imibono embalwa engabonakali, umsebenzi osheshayo, umphumela ongcono - noma kunjalo, kubonakala kanjalo, ngoba leso sithembiso sesibe ulimi olujwayelekile lwemboni. (Reuters)
🇰🇷 I-Rebellions, inkampani eqala i-AI chip yaseNingizimu Korea, iqoqe imali eyizigidi ezingama-400 zamaRandi emzuliswaneni wakamuva wokuxhasa ngezimali ↗
AmaRebellions athole u-$400 million ngenani elilinganiselwa ku-$2.34 billion, okuyisilinganiso esikhulu esibonisa ukuthi abatshalizimali basazimisele ukusekela imidlalo yengqalasizinda ye-AI, ikakhulukazi ngaphandle kwabasolwa abavamile base-US. Le nkampani ithe imali izosiza ukuthi ikhule e-US, ikhulise ipulatifomu yayo yeRebel100, futhi iqhubekele phambili ku-IPO ekugcineni.
Okugqama kakhulu ukwakheka kwezepolitiki nezimboni okuzungezile. Isikhwama sokukhulisa iKorea senze utshalomali oluqondile ngaphansi kokucindezela kwezwe "i-K-Nvidia", kuyilapho iRebellions ibheja ukuthi inkundla yempi yangempela ye-AI isishintshe kusukela kuma-chatbot akhangayo yaya emishinini eshibhile nesebenza kahle ngaphansi kwawo. Lokho kuzwakala kungakhangi kangako, kubaluleke kakhulu. (Reuters)
🇫🇷 I-Mistral yaseFrance iqongelela isikweletu esingu-$830 million sokwakhiwa kwezikhungo zedatha ze-AI ↗
I-Mistral ithole imali eyizigidi ezingama-830 zamaRandi ukuze yakhe isikhungo sedatha eduze kwaseParis futhi ithenge ama-chips e-Nvidia angu-13,800. Lokho kuyinsimbi enkulu, amandla amakhulu, kanye nesitatimende esicacile sokuthi iYurophu ayifuni ukuqasha ikusasa layo le-AI unomphela kuma-giants aseMelika.
Indawo kulindeleke ukuthi iqale ukusebenza kwikota yesibili futhi iyingxenye yohlelo olubanzi lokufinyelela ama-megawatts angu-200 omthamo wokubala kulo lonke elaseYurophu ekupheleni kuka-2027. I-Mistral iphinde ikhonze amakhasimende okuhlanganisa namabutho aseFrance, ngakho-ke lokhu akuyona nje indaba yokukhula kwenkampani entsha - ingqalasizinda, ubukhosi, kanye namasu konke kuhlanganiswe ndawonye njengezintambo ngaphansi kwephansi eliphakanyisiwe. (Reuters)
🍎 I-Apple Intelligence yethulwe ngephutha eShayina. ↗
I-Apple yadalula ngephutha izici ze-Apple Intelligence kwabanye abasebenzisi be-iPhone eShayina, yabe isizihoxisa ngemuva kokusabalala kwemibiko ku-inthanethi. Ngokusho kwesifinyezo semibiko ye-Verge, ukukhishwa kwenzeke ngephutha, okuyinto engeyinhle neze, uma kubhekwa indlela izici ze-AI ezilawulwa ngayo emakethe yaseShayina.
Ingxenye exakile ukuthi i-Apple isadinga umlingani wendawo ukuze anikeze amandla amathuluzi e-AI lapho, kanti i-Alibaba ingenye yezinkampani okukhulunywe ngazo ngaphambilini kulowo mongo. Ngakho lokhu kwakungeyona nje i-switch switch flip - kwaveza kafushane ukuthi isu le-Apple le-China AI lisaqediwe kangakanani. (The Verge)
⚖️ Ucwaningo luthole ukuthi iningi lamajaji kahulumeni wase-US asebenzisa i-AI ↗
Ucwaningo olusha luthole ukuthi amajaji kahulumeni wase-US angama-60% asebenzisa okungenani ithuluzi elilodwa le-AI emsebenzini wobulungiswa, yize ama-22% kuphela athi alisebenzisa nsuku zonke noma masonto onke. Iningi lithembele ezinhlelweni ezithile zomthetho kune-chatbots ejwayelekile, futhi icala lokusetshenziswa eliphambili ucwaningo lwezomthetho, olulandelwa ukubuyekezwa kwedokhumenti.
Lokho kuyishintsho elikhulu kancane. Izinkantolo zivame ukuba yisithwathwa, bese kungazelelwe azikho, futhi lapha ijaji elilodwa kwabathathu lithe liyakuvumela noma likhuthaza ukusetshenziswa kwe-AI emakamelweni kuyilapho ama-20% ekwenqabela ngokusemthethweni. Ngesikhathi esifanayo, abangaphezu kuka-45% bathi abaphathi bezinkantolo abazange banikeze ukuqeqeshwa kwe-AI, okuzwakala njengephethini ejwayelekile yobuchwepheshe besimanje - ukwamukelwa kuqala, izithiyo kamuva. (Reuters)
📉 Njengoba abantu baseMelika abaningi besebenzisa amathuluzi e-AI, bambalwa abathi bangayethemba imiphumela ↗
Ucwaningo olusha lubonise ukuthi abantu baseMelika basebenzisa i-AI kakhulu kuyilapho bengayethembi kakhulu, okuzwakala kuphikisana kuze kube yilapho ukhumbula ukuthi yilokho abantu abaphatha ngakho ingxenye ye-inthanethi. Ucwaningo luthole ukuthi ama-70% acabanga ukuthi ukuthuthukiswa kwe-AI kuzonciphisa amathuba emisebenzi, kuyilapho ama-7% kuphela ecabanga ukuthi kuzodala imisebenzi eminingi.
Isimo sengqondo sibi ngaphezu kwemisebenzi. Izingxenye ezimbili kwezintathu zabaphenduli zithe izinkampani azibonakali ngokwanele mayelana nokusetshenziswa kwe-AI, futhi ingxenye efanayo ithi uhulumeni akakulawuli ngokwanele. Ngakho-ke ukwamukelwa kuyanda, kodwa intshiseko ayinjalo - ukuhlala ndawonye okubi kakhulu kunokuthandana. (TechCrunch)
Imibuzo Evame Ukubuzwa
Kungani iMicrosoft ihlanganisa amamodeli amaningi e-AI ngaphakathi kwe-Copilot?
Kubonakala sengathi iMicrosoft iqondisa iCopilot ekusethweni kwamamodeli amaningi ngoba imodeli eyodwa ingakhiqiza kuyilapho enye ibuyekeza. Kulesi simo, i-GPT ibhala impendulo bese uClaude eyihlola ukuze athole ikhwalithi nokunemba. Inhloso ewusizo ukunciphisa ukubona izinto ezingekho futhi iqinise umphumela ngaphandle kokucela uhlelo olulodwa ukuthi lwenze konke ngesikhathi esisodwa. Lokho kwenza iCopilot ingabi njenge-chatbot ezimele kodwa ifane nomsebenzi olawulwayo.
Yini ngempela eyenziwa yisici esisha se-Critique se-Microsoft?
I-Critique ichazwa njengohlelo lokubuyekeza lapho imodeli eyodwa ibhala khona impendulo bese enye iyihlola. Lokho kubalulekile ngoba ingeza ukudlula kwesibili ngaphambi kokuba umphumela ufike kumsebenzisi. Amaqembu avame ukufuna lolu hlobo lokusetha ukuze abambe amaphutha, ukucabanga okubuthakathaka, noma amagama angacacile ekuqaleni kwenqubo. I-Microsoft iphinde yabonisa ukuthi ifuna lolu hlelo lokuphendula lube nokuxhumana kakhudlwana ngokuhamba kwesikhathi.
Kungani abatshalizimali besasekela izinkampani ze-AI chip ezifana neRebellions?
Lesi sihloko sibonisa ukuthi abatshalizimali basabona izinzuzo ezinkulu engqalasizinda ye-AI. I-Rebellions izibeka endaweni yokubala eshibhile, esebenza kahle kakhulu kunezimangalo ze-chatbot ezibhekene nabathengi. Leyo nkulumo ingaba yinto ekhangayo ngoba isidingo se-AI sisaxhomeke ku-hardware engaphansi kwamamodeli. Umjikelezo wayo wakamuva wokuxhasa ngezimali uphinde uphakamise ukuthi ohulumeni bezifunda kanye nezikhwama bafuna abanqobi bendawo kule ngxenye ye-stack.
Kungani ukwanda kwesikhungo sedatha saseMistral kubalulekile esulweni le-AI laseYurophu?
Ukwanda kweMistral kubalulekile ngoba akugcini nje ngokukhula kwenkampani, kodwa futhi nokulawula amandla okusebenzisa ikhompyutha. Ukuthenga izinkulungwane zama-chip eNvidia nokwakha indawo eduze kwaseParis kubonisa ukuthi iYurophu ifuna ingqalasizinda yayo ye-AI eyengeziwe ezweni lasekhaya. Kuwo wonke amapayipi amaningi e-AI, ukufinyelela ezikhungweni zedatha, amandla, nama-chip kusiza ekunqumeni ukuthi ubani ongakha futhi asebenzise amamodeli ngezinga elikhulu.
Kungani i-Apple Intelligence yavela eShayina into enkulu kangaka?
Kwaba yinkinga enkulu ngoba izici ze-AI eShayina zisebenza endaweni elawulwa kahle. Kubikwa ukuthi i-Apple yadalula amathuluzi ngephutha ngaphambi kokuwabuyisela emuva, okwaqokomisa ukuthi ukukhishwa kwayo kwe-AI yendawo kusabonakala kungaqediwe. Lesi sihloko siphinde siphawule ukuthi i-Apple idinga umlingani wendawo lapho, ngakho-ke lesi sigameko sasingeyona nje inkinga yesofthiwe, kodwa inkinga yesu eyabonakala.
Kungani abantu abaningi besebenzisa amathuluzi e-AI yize ukwethembana ku-AI kuqhubeka kwehla?
Lesi sihloko sikhomba iphethini lapho ukwamukelwa kukhula khona ngoba i-AI iyasebenza, noma ngabe ukuzethemba emiphumeleni kuhlala kubuthakathaka. Amajaji ase-US asevele esebenzisa i-AI ocwaningweni nasekubuyekezweni kwamadokhumenti, kuyilapho ucwaningo lubonisa ukuthi abantu abaningi baseMelika basakhathazeka ngemisebenzi, ukucaca, kanye nomthethonqubo. Lokho kuhlanganiswa kusikisela ukuthi i-AI ingena emsebenzini ngaphambi kokuba ukwethembana, ukuqeqeshwa, kanye nokuphatha kufinyelele ngokugcwele.