Iqembu le-Lloyds Banking lizoqasha ochwepheshe bezobuchwepheshe abangu-300 ukuze basebenze ku-AI ↗
ULloyds wengeza ochwepheshe bezobuchwepheshe abangu-300 ukuze bakhe i-AI ejensi, kanye nabasebenzi abaqashwayo abajoyina ithimba elibanzi le-AI labantu abayi-1,000. Ibhange lifuna izinhlelo ezisebenza ngokukhwabanisa, ukusesha amadokhumenti e-HR, kanye namathuluzi ebhange aku-inthanethi engeziwe.
Ukungezwani, ngokusobala, ukuthi lokhu kuyizindaba zokuqasha ezizungeze ukukhathazeka ngokuzenzakalelayo. ULloyds uthi i-AI isivele ilethe amashumi ezigidi zenzuzo, kuyilapho ishiya ithuba lokuncishiswa kwemisebenzi njengoba izindima ziqhubeka nokushintsha.
I-Atlantic idale isizindalwazi somculo esingasetshenziswa ukuqeqesha i-AI ↗
Isizindalwazi esingasesheka manje sivumela abantu ukuthi babheke amasethi edatha omculo asetshenziselwa ukuqeqeshwa kwe-AI. Amanye makhulu - elinye linamathrekhi ayizigidi ezingu-12, elinye linama-9 million, kanye namaqoqo amancane asadlula izingoma eziyi-100,000.
Ingxenye engakhululekile: amathrekhi angafakwa ohlwini olukhululekile ku-inthanethi ngenkathi engakaqedwa ukuze aqeqeshwe ngokwezentengiselwano. Ingenye yalezo zinkinga "zomphakathi kodwa ezingenayo imvume" eziqhubeka nokuguqula i-AI-copyright crank.
Ku-Weights kukhona usesho lwakho olusha lwe-vanity olugxile ku-AI ↗
Ku-Weights kushintsha ukufuna i-ego kube ibhodi lamaphuzu lesikhathi se-AI. Esikhundleni sokubuza i-Google ukuthi ungubani, ibuza amamodeli afana no-GPT, uClaude, uGemini, uGrok, noLlama ukuthi “bakhumbulani” ngegama.
Isayithi linikeza abantu amaphuzu anamandla futhi libonisa lapho amamodeli enza khona izinto ezingabonakali. Kumnandi ngendlela engavamile, kuqalekisiwe kancane, futhi mhlawumbe kufana ne-maze encane yesibuko sayo yonke into ethanda kakhulu ethi “ingabe ngikhona kudatha yokuqeqeshwa?”.
UMeredith Whittaker weSignal ufuna ukhumbule ukuthi ama-chatbot e-AI 'awabona abangane bakho' ↗
UMeredith Whittaker uxwayise ngokuphatha ama-chatbot njengabangane noma abalingani bokucabanga abathembekile. Iphuzu lakhe lalicacile: lezi zinhlelo aziqapheli, azibona abangane, futhi azizona izimfihlo zangasese.
Uphinde wagxeka abasizi be-ejenti abadinga ukufinyelela ezingxoxweni, amakhadi esikweletu, iziphequluli, amakhalenda, kanye nemiyalezo. Ezweni le-Signal, lolo hlobo lokunethezeka luqala ukubukeka njengomnyango wangemuva.
UJohn Jumper, owine umklomelo weNobel, ushiya iDeepMind ukuze athathe imbangi yakhe i-Anthropic ↗
UJohn Jumper, owaziwa ngomsebenzi wakhe we-AlphaFold, ushiya i-Google DeepMind ngenxa ye-Anthropic. Lokho akusikho nje ukunyakaza kwabasebenzi - kuyisibonakaliso esicacile sokuthi ithalente eliphezulu le-AI lisaqhubeka nokusebenza phakathi kwama-lab asemngceleni.
Isikhathi sifika kanye nezinye izindlela ezidumile zokususa i-DeepMind. Ngakho-ke, yebo, ilebhu isenedumela elikhulu, kodwa amandla adonsela phansi azungeze i-Anthropic abukeka engathi sína.
I-AI yenza ukuphepha kwe-crypto kushibhile, kusheshe futhi kube nzima ukukushalazela ↗
Amathuluzi okuphepha e-AI afana ne-Mythos angasunduza ukuhlolwa kwe-smart-contract ekubuyekezweni okushibhile nokuqhubekayo. Lokho kubalulekile ngoba ukuhlolwa kwe-crypto kuvame ukubiza, kuhamba kancane, futhi kubhekwa njengokuzikhethela ngamaqembu amancane.
Abacwaningi basaxwayisa ngokuthi i-AI ngeke ithathe indawo yokwahlulela kwabantu. Izinhlekelele eziningi ze-crypto zivela ezitifiketini ezisengozini, ubunjiniyela bezenhlalo, noma ukusebenza okubuthakathaka - hhayi nje ikhodi yephutha.
IJapan izothuthukisa ukubambisana komhlaba wonke ezingozini ze-AI ↗
IJapan ikhiphe uhlaka lokubuyekezwa kohlelo lwayo oluyisisekelo lwe-Artificial Intelligence, igcizelela kakhulu ukubambisana nohulumeni bangaphandle kanye nabathuthukisi be-AI. Okugxilwe kukho ukusetshenziswa kabi, ukuhlaselwa kwe-cyber, kanye nezingozi zemodeli ezihamba ngokushesha.
Kuwukusetha kabusha kwenqubomgomo okwenziwe ezinyangeni ezimbalwa nje ngemuva kokwakhiwa kohlelo okokuqala. Lokho kusho ingxenye ethule ngokuzwakalayo: Ukuphathwa kwe-AI kuguga njengobisi egagasini lokushisa.
Uhlelo olusha lomhlaba wonke: Abaphathi be-AI njengezinhloko zezizwe ↗
Engqungqutheleni ye-G7, ama-CEO e-AI ahlaliswe nabaholi bomhlaba endaweni eyayizwakala, ngokusobala, njengegumbi lomkhandlu lesayensi. Abaholi be-OpenAI, i-Google DeepMind, i-Anthropic, kanye ne-Salesforce baphathwe njengabalingisi abaphambili ekuphepheni nasekubuseni komhlaba wonke.
Inkinga eyinhloko amandla. Ohulumeni bafuna abalingani be-AI abathembekile, kodwa futhi abafuni ukuthi ama-lab azimele anqume ikusasa ngokwawo - noma yilowo nalowo usho kanjalo, kuze kube yilapho kufika inkinga elandelayo.
Imibuzo Evame Ukubuzwa
Kungani i-Lloyds Banking Group iqasha ochwepheshe be-AI abengeziwe?
I-Lloyds iqasha ochwepheshe bezobuchwepheshe abangu-300 ukuze bandise umsebenzi wayo ku-AI ejensi, yengeza eqenjini elibanzi le-AI labantu abangaba yi-1,000. Ibhange lifuna ukusebenzisa lezi zinhlelo ezindaweni ezifana nokutholwa kokukhwabanisa, ukusesha amadokhumenti e-HR, kanye nokubhanga okuku-inthanethi okwenziwe ngezifiso. Lesi sihloko siphinde siphawule ukungezwani phakathi kwezinzuzo eziqhutshwa yi-AI kanye nokungaqiniseki ngokuthi ukuzenzekela kungashintsha kanjani imisebenzi.
Kungani isizindalwazi sokuqeqeshwa komculo se-AI sibalulekile?
Isizindalwazi somculo esingasesheka sibalulekile ngoba sivumela abantu ukuthi bahlole amathrekhi afakwe kumasethi edatha asetshenziselwa ukuqeqeshwa kwe-AI. Amanye amasethi edatha makhulu, aqukethe izigidi zamathrekhi. Inkinga eyinhloko ukuthi umculo ungafakwa ohlwini esidlangalaleni ku-inthanethi ngaphandle kokuthi uvunyelwe ukuqeqeshwa kwe-AI yezentengiselwano, okushiya imibuzo ye-copyright kanye nemvume ingaxazululwa.
Kuyini "Ezisindweni" futhi kuhlobene kanjani nokusesha kwe-AI?
I-In the Weights ichazwa njengethuluzi lokusesha i-vanity lesikhathi se-AI. Esikhundleni sokukhombisa imiphumela ejwayelekile yewebhu, ibuza amamodeli afana ne-GPT, Claude, Gemini, Grok, kanye ne-Llama ukuthi babonakala “bekhumbula” ini ngegama. Ibe isihlinzeka ngamaphuzu wamandla futhi ibonise lapho amamodeli enza khona izinto ezingabonakali, okwenza kube ukubonakaliswa okujabulisayo kodwa okungakhululekile kwesifiso sokuqeqeshwa kwedatha.
Kungani abameli bobumfihlo bexwayisa ngama-chatbot e-AI?
UMeredith Whittaker weSignal uxwayisa ngokuthi ama-chatbot akufanele aphathwe njengabangane, izidalwa eziqaphelayo, noma abantu abayimfihlo. Ukukhathazeka kwakhe kukhulu kakhulu ngabasizi be-ejenti abadinga ukufinyelela imiyalezo, amakhalenda, iziphequluli, amakhadi esikweletu, kanye nezingxoxo. Ezinhlelweni ezigxile kubumfihlo, lolo hlobo lokulula lungaqala ukubukeka njengobuthakathaka obukhulu bokuphepha.
Kusikiselani ukusuka kukaJohn Jumper ku-DeepMind ukuya ku-Anthropic?
Ukusuka kukaJohn Jumper esuka ku-Google DeepMind kuya ku-Anthropic kusikisela ukuthi ithalente le-AI eliphezulu lisaqhubeka ngamandla phakathi kwama-lab asemngceleni. Ngenxa yokuthi i-Jumper yaziwa nge-AlphaFold, ukuhamba kwayo kungaphezu nje kwezindaba zokuqasha ezivamile. Lesi sihloko sikubeka njengesinye isibonakaliso sokuthi i-Anthropic isibe yi-talente ekhangayo, yize i-DeepMind isalokhu idumile kakhulu.