I-OpenAI izomisa umsebenzi othile kumodeli ye-AI i-Astra ngenxa yezinkinga zokuphepha ↗
I-OpenAI imisa umsebenzi wangaphakathi ku-Astra ongahlangabezani nezidingo zokuphepha eziqinile ngemuva kokuthi ukuhlolwa kuphakamise ukuthi amakhono ayo e-inthanethi kungenzeka ukuthi adlule umkhawulo "obucayi". Kubikwa ukuthi i-ejenti ingathola futhi isebenzise ubuthakathaka ngaphandle kokungenelela komuntu, noma ihlasele kusuka emgomweni ophakeme. Yebo... lowo umugqa onzima kakhulu okufanele uwele.
Le nkampani iqinisa ukuhlolwa okuhlukile, ukufinyelela kwenethiwekhi namathuluzi, ukuvikelwa kwesisindo semodeli, ukubethela kanye nokuqapha. I-Astra ayizange ihileleke esigamekweni sangaphambilini se-Hugging Face se-OpenAI, kodwa iphethini ebanzi yama-ejenti azimele ngokwandayo yenza ukulawulwa kube yinto ekhathazayo kakhulu. (Umthombo: The Guardian)
I-Apple ithi abasebenzisi be-Mac eShayina bangaxhuma kwinsizakalo ye-Alibaba's Qwen AI ↗
Abasebenzisi be-Mac abafanelekile ezwenikazi laseShayina manje sebengakwazi ukuxhuma i-Qwen ka-Alibaba ku-Siri kanye namathuluzi okubhala e-Apple. Lokho kusho izimpendulo ezinemininingwane eminingi kanye nokuhlaziywa kwamadokhumenti nezithombe, kanye nokukhiqizwa kombhalo nezithombe kusuka ezincazelweni.
Kungumshado othakazelisayo, ngokusobala - i-Apple ithola indlela ehambisanayo endaweni eya ezicini ze-AI eziqinile emakethe yama-PC encintisana kakhulu, kuyilapho i-Alibaba ifaka i-Qwen ejulile kakhulu ohlelweni lomunye umuntu. I-Apple ithi izinto ezithunyelwe ngesandiso azikwazi ukusetshenziswa yi-Alibaba ukuqeqesha amamodeli ayo. (Umthombo: Reuters)
I-Google DeepMind ingena enkathini entsha njengoba umsunguli ohlangene uDemis Hassabis eshintsha indima ye-AI ↗
UDemis Hassabis uyeka ubuholi bansuku zonke be-DeepMind, eba ngusihlalo ngenkathi ethatha nesikhundla sososayensi oyinhloko kwa-Alphabet. Uzakwabo osenesikhathi eside uKoray Kavukcuoglu uthatha ukulawula kokusebenza njengephini likamongameli omkhulu kunokuba yi-CEO - ushintsho oluncane lwesihloko olusho okuningi.
Ukushintshashintsha kufika kanye nabanye abacwaningi abakhulu abashiyayo futhi kunemibuzo ngokuthi ngabe i-DeepMind ayisazimeli kakhulu futhi igxile kakhulu emshinini we-AI wezohwebo we-Google. I-Google iyaphikisana ngokuthi lolu shintsho luhlobene nokusebenza kokuncintisana kwe-Gemini. (Umthombo: The Guardian)
Inani elikhulayo lezingane zase-UK libika ukuthi zibona izithombe ezijulile zazo ↗
Imibiko evela ezinganeni zase-UK mayelana nezithombe zabo ezingamanga noma ezikhohlisiwe iye yanda. I-Report Remote ithole imibiko engu-420 ezinyangeni eziyisithupha nje kuphela, isivele idlule inani layo eliphelele langaphambilini lonyaka elingu-397. Amathuluzi "okususa" asebenzisa i-AI enza inkinga ibe lula kakhulu ukuyikhulisa.
I-Internet Watch Foundation ithi izithombe eziningi ezibikiwe ezishintshiwe zazinkulu ngokwanele ukuba zidlule umkhawulo wezinto zokuhlukumeza ngokocansi zezingane. Ezinye izisulu nazo zazihloselwe ukuhlukumeza ngokocansi, okwenza isithombe esenziwe nge-AI saba yisikhali esibonakalayo. (Umthombo: The Guardian)
Ucwaningo luthole ukuthi izici ze-AI zikhula phambili kunenani lebhizinisi ↗
Izinkampani zesofthiwe zamabhizinisi ziveza izici ze-AI, kodwa amakhasimende awasebenzisi ngempela. Ocwaningweni lwamabhizinisi esofthiwe angaphezu kuka-260 aqondile, ama-51% athi angaphansi kwengxenye eyodwa kwezine yamakhasimende awo asebenzisa amakhono e-AI awakhile. Kunzima kakhulu ngemva kwakho konke lokho kuthunyelwa...
Inkinga ayibukeki njengekhono le-AI kodwa ifana kakhulu nokuhlanganiswa, ukuphepha, ubumfihlo, izinguquko zomsebenzi kanye nokuqinisekisa inani eliqondile lebhizinisi. Isenzakalo esisodwa esingavamile kakhulu "i-tokenmaxxing" - ukusetshenziswa okuningi kwe-AI okubukeka kumangalisa kudeshibhodi ngaphandle kobufakazi obucacile bokuthi kukhiqiza imiphumela engcono. (Umthombo: Forbes)
I-AI ishintsha umsebenzi ngokushesha kunokuba idatha ingaqhubeka ↗
Impikiswano nge-AI-jobs iyaxaka kakhulu, ayihlanzeki. Ucwaningo oluhlanganisa izinkampani zase-US ezingaphezu kuka-21,000 luthole ukuthi abantu abasebenzisa i-AI kakhulu bakhule ngenani labantu abasebenza ngobuchwepheshe besimanje ngo-10% kanye nokuqashwa kwabantu abasafufusa ngo-12%, kuyilapho abantu abasebenzisa ubuchwepheshe besimanje bengabonisi ukukhula kwenani labantu abasebenza ngobuchwepheshe besimanje.
Kodwa olunye ucwaningo luthole ingcindezi yokuqashwa phakathi kwabasebenzi abasebasha emisebenzini evezwe yi-AI, kanti izazi zomnotho zixwayisa ngokuthi izinkampani ezikhulayo zisengaqasha ngenkathi i-AI ishintsha imisebenzi kuwo wonke umnotho. Ngakho-ke, i-AI ibulala imisebenzi futhi iyidala... noma kunjalo amasethi edatha aphikisanayo angakwenza ukholwe. (Umthombo: Fortune)
Imibuzo Evame Ukubuzwa
Yiziphi intuthuko enkulu yezindaba ze-AI njengamanje?
Izindaba zakamuva ze-AI zikhomba ezindikimbeni eziningana ezihambisanayo: amakhono aqinile okuzimela, izilawuli zokuphepha eziqinile, ukuhlanganiswa okujulile kwepulatifomu, izinguquko zobuholi, izingozi zokuphepha ku-inthanethi, kanye nenani lebhizinisi elingaqinisekile. I-OpenAI iqinisa izivikelo ezizungeze i-Astra, i-Apple ixhumanisa i-Qwen ne-Macs eShayina, kanti i-DeepMind ishintsha ubuholi. Ngesikhathi esifanayo, ukukhathazeka kuyanda mayelana nama-deepfake, ukwamukelwa kwamabhizinisi, kanye nomthelela we-AI emisebenzini.
Kungani i-OpenAI imise umsebenzi othile ku-Astra?
Kubikwa ukuthi i-OpenAI imise umsebenzi wangaphakathi we-Astra ongahlangabezani nezidingo zokuphepha eziqinile ngemuva kokuthi ukuhlolwa kuphakamise ukuthi amakhono e-cyber emodeli kungenzeka ukuthi afinyelele ezingeni elibucayi. Ukukhathazeka ukuthi i-ejenti ethuthukisiwe ingathola futhi isebenzise ubuthakathaka ngokungenelela okulinganiselwe noma okungekho muntu. Ukuphendula, i-OpenAI iqinisa ukuhlolwa okuhlukanisiwe, izilawuli zokufinyelela, ukubethela, ukuqapha, kanye nokuvikelwa okuzungeze isisindo semodeli.
I-Alibaba Qwen izosebenza kanjani ne-Apple Macs eShayina?
Abasebenzisi be-Mac abafanelekile ezwenikazi laseShayina bangaxhuma isevisi ye-Qwen ye-Alibaba ku-Siri kanye namathuluzi okubhala e-Apple. Ukuhlanganiswa kungasekela izimpendulo ezinemininingwane eminingi, ukuhlaziywa kwamadokhumenti kanye nezithombe, kanye nokukhiqizwa kombhalo noma izithombe ezivela ezincazelweni. Ku-Apple, leli lungiselelo linikeza indlela ehambisana nendawo yokusebenza kwe-AI eqinile, kuyilapho i-Alibaba ithola ukufinyelela okujulile ohlelweni lwe-Apple. I-Apple ithi izinto ezithunyelwe azikwazi ukusetshenziselwa ukuqeqesha amamodeli e-Alibaba.
Kusho ukuthini ukushintsha kobuholi be-DeepMind ku-Google AI?
UDemis Hassabis uyasuka ebuholini bansuku zonke be-DeepMind ngenkathi eba ngusihlalo futhi ethatha indima yososayensi oyinhloko kwa-Alphabet. UKoray Kavukcuoglu uzothatha ukulawula kokusebenza njengephini likamongameli omkhulu. Ushintsho luphakamise imibuzo mayelana nokuthi i-DeepMind iya ngokuya ihlanganiswa kakhulu nemisebenzi ye-AI yezentengiselwano ebanzi ye-Google, yize i-Google iphikisana neziphakamiso zokuthi ukuguqulwa kwe-DeepMind kuhlobene nokusebenza kokuncintisana kwe-Gemini.
Kungani ama-deepfake e-AI eba yinkinga enkulu yokuphepha kwezingane?
Amathuluzi okuphatha izithombe asebenzisa i-AI angenza kube lula ukudala izithombe ezingamanga ezingokoqobo ngezinga elikhulu. Izingane zase-UK ziya ngokuya zibika izithombe zazo ezikhohlisiwe noma ezenziwe ngazo, futhi amanye amacala abe makhulu ngokwanele ukuhlangabezana nomkhawulo wezinto zokuhlukunyezwa ngokocansi kwezingane. Lezi zithombe zingasetshenziselwa futhi ukuthunjwa ngokocansi, okusho ukuthi isithombe esenziwe singase sibangele ukulimala okukhulu kwedumela, ngokomzwelo, nangokwezimali.
Ingabe amabhizinisi nabasebenzi bayazuza ku-AI okwamanje?
Izindaba zakamuva ze-AI ziphakamisa ukuthi impendulo ixubile. Izinkampani eziningi zesofthiwe zengeza izici ze-AI ngokushesha kunamakhasimende azisebenzisayo, kanti ukuhlanganiswa, ubumfihlo, ukuphepha, kanye nenani lebhizinisi elilinganiswayo kusalokhu kuyizithiyo ezinkulu. Idatha yokuqashwa iyinkimbinkimbi ngendlela efanayo: ezinye izinkampani ezisebenzisa i-AI enzima ziyanda ukuqasha, kanti olunye ucwaningo lukhomba ekucindezelweni kwabasebenzi abasebasha ezindimeni ezivezwe yi-AI. Umthelela jikelele uncike kakhulu embonini, emsebenzini, kanye nendlela i-AI esetshenziswa ngayo.