🏛️ UTrump uqondisa izinhlangano zase-US ukuthi zilahle i-AI ka-Anthropic njengoba iPentagon ibiza inkampani entsha ngokuthi ingozi yokuhlinzeka ngezinsizakalo ↗
Umongameli wase-US uyale izinhlangano zikahulumeni ukuthi ziyeke ukusebenzisa ubuchwepheshe be-Anthropic, ngefasitela eliphoqelelwe lokuqeda umsebenzi eliguqula ukuhlukana kwabathengisi kube yisikhathi senqubomgomo esiphelele. I-Pentagon ibeke i-Anthropic "njengengozi yochungechunge lokuphakelwa" - ilebula elihlaba umxhwele elizobekwa elebhu enkulu ye-AI yase-US. (Reuters)
I-Anthropic ibonise ukuthi izolwa nalesi sinyathelo, futhi lesi siqephu siveza ngokucacile umbango omkhulu: imithetho yokuphepha kwenkampani vs izidingo zikahulumeni zokuguquguquka okukhulu. Uma ungumthengi webhizinisi obukela eceleni, inkinga eyinhloko ilula - uma ikhasimende kuyisifundazwe, imigomo yesifundazwe ivame ukunquma isimo sezulu. (Reuters)
⚖️ IPentagon imemezela i-Anthropic njengosongo ekuphepheni kwezwe ↗
Le nguqulo yengeza ukuthungwa okwengeziwe endleleni yesitayela sohlu olumnyama nokuthi kusho ukuthini kosonkontileka - hhayi izinhlangano kuphela - ezithinta uhlelo lwezemvelo lukahulumeni. Luhlobo lomthetho oluphuma ngokushesha, njengoyinki emanzini, ngaphandle kokuthi uyinki uyiphepha lokuthenga. (I-Washington Post)
Kukhona futhi ukungqubuzana okuqondile mayelana nokuthi izinqubomgomo zokusetshenziswa komhlinzeki oyimodeli zingavimbela yini ukusetshenziswa kwezempi, ikakhulukazi ezimweni zokusetshenziswa ezibucayi. Ukusabela kwemboni kuzwakale kushubile; isibonelo yilokho okukhathaza abantu, hhayi inkampani eyodwa kuphela ekhonjwayo. (I-Washington Post)
☁️ I-OpenAI ne-Amazon bamemezela ukubambisana okuneqhinga ↗
I-OpenAI imemezele ukubambisana okunesu ne-Amazon okuletha ipulatifomu ye-OpenAI Frontier ku-AWS, yanda lapho amakhasimende angasebenzisa futhi aphathe khona izinhlelo ze-OpenAI-grade. Uma ubulokhu ulandela ukuthi “ubani osingatha ini” ku-AI, lokhu kufana nokushintsha kwamandla okugcina izwi lawo liphansi ngenkathi kuhanjiswa ifenisha eningi. (OpenAI)
Kuphinde kufike njengempendulo ekucindezelweni kwesidingo - izinketho eziningi zengqalasizinda, izindlela eziningi zokusabalalisa, imigwaqo emincane enomzila owodwa. Ukuthi lokhu kwenza ukuthunyelwa kube lula noma kumane kunikeza wonke umuntu izinkinobho ezengeziwe zokuhlela kabi kusazobonakala. (OpenAI)
🧠 I-OpenAI yethula i-AI ecacile ku-AWS, ikhombisa ukushintsha kwamandla endiza yokulawula ↗
Umbono we-Computerworld: "i-AI eqondile" ku-AWS imayelana nokungaphezu kokusingatha - imayelana nokuthi indiza yokulawula ihlala kuphi, nokuthi ubani ohlela ubunikazi, inkumbulo, kanye nokusebenza kwayo phakathi nezikhathi. Izinhlelo eziqondile zingazizwa "zifana ne-agent" kakhulu, zibe ngcono futhi zizizwe zinjalo nje. (Computerworld)
Ingqikithi iwuhlobo lokuncintisana: amafu afuna ukuba ngumnikazi wesendlalelo sepulatifomu, amalebhu e-AI afuna ukuba ngumnikazi womphezulu womkhiqizo, kanti amakhasimende afuna ungaphuki ngo-2 ekuseni. Wonke umuntu ufuna isiteringi - noma ngabe enza sengathi akasifuni. (Computerworld)
🔐 Ushintsho olubalulekile lwe-Google API 'oluthule' luveze idatha ye-Gemini AI - CSO ↗
Isexwayiso sokuphepha sizungeza okhiye be-Google Cloud API kanye ne-Gemini - kanye nokubika ukuthi izinguquko endleleni okhiye abasebenza ngayo (noma abaphathwa ngayo) zingenza "ukucabanga okuphephile ukushumeka" kube ingozi enkulu, ekhukhulekayo. Ingxenye ephazamisayo ukuthi ungakwenza kalula kangakanani konke "njengoba ubuhlala wenza," kodwa ugcine ungaphandle kwemingcele yokuvikela. (CSO Online)
Abacwaningi bakhombe okhiye abaveziwe kabanzi kuwo wonke ama-org, okungeyona inkinga eyodwa kodwa kuyisikhumbuzo sokuthi ukuhlanganiswa kwe-AI kwandisa ububanzi bokuphathwa kwezimfihlo ezindala okuyisicefe. Yizinto ezingafanele eziluma kakhulu. (SC Media)
📱🎶 I-Gemini Drops: Izibuyekezo ezintsha kuhlelo lokusebenza lwe-Gemini, uFebhuwari ... ↗
Isibuyekezo sohlelo lokusebenza lwe-Gemini lwe-Google siqhakambisa ukuthuthukiswa kwe-Gemini 3.1 kanye nemodi yokucabanga ethi “Deep Think” ebekelwe izinkinga zesayensi ezisindayo kanye nesitayela sobunjiniyela, kanye nokufakwa kwegethi kwesigaba sababhalisile. Imodi ehlakaniphile, uthango oluphezulu, inhlanganisela yakudala. (blog.google)
Futhi: I-Lyria 3 ibizwa ngokuthi imodeli yomculo engakhiqiza amathrekhi amafushane kusuka kumbhalo noma izithombe ku-beta. Kuyathandeka ukuthi uhlelo olufanayo lokunikeza ukucabanga okuqinile luphinde lunikeze imisindo esheshayo, eyenziwe ngokwezifiso - amagiya amabili, ibhokisi elilodwa legiya. (blog.google)
Imibuzo Evame Ukubuzwa
Yini eshintsha izikhungo zikahulumeni wase-US ezivimba ukusetshenziswa kobuchwepheshe be-Anthropic?
Iguqula ukukhetha komthengisi kube umthetho wokuthenga, onesikhathi esinqunyiwe sokuphuma esikhundleni sezinqumo zesikhashana, zeqembu neqembu. Uhlaka lwe-Pentagon "lwengozi yokunikezela" luphakamisa izingqinamba futhi lubonise ukuthi ukufaneleka kungabunjwa yinqubomgomo ngaphezu kokufaneleka komkhiqizo. Kubathengi, kugcizelela ukuthi izidingo zomkhakha womphakathi zingadlula kanjani imodeli yokusebenza ekhethwa ngumhlinzeki.
Ilebula lePentagon elithi “ingozi yokunikezela” lingabathinta kanjani osonkontileka kanye nabathengisi abangaphansi komfula?
Umbiko uphakamisa ukuthi umthelela awugcini nje kuma-ejensi; ungasakazekela kosonkontileka abahlangana nesimiso semvelo sikahulumeni. Ngisho noma ungalokothi uthenge imodeli ngqo, i-stack yakho ingazuza imikhawulo ngokusebenzisa izinkontileka eziyinhloko, izigaba zokugeleza, kanye nokuhlolwa kokuhambisana nomthetho. Yingakho "ubani osebenzisa lokho" eba yinkinga yamaphepha ngokushesha, hhayi nje impikiswano yokwakha.
Yini okufanele abathengi bebhizinisi bayenze uma umhlinzeki we-AI oyinhloko ebanjwa ekuvinjelweni kokuthengwa kwe-AI kahulumeni?
Qala ngokumaka lapho umhlinzeki ebonakala khona: ukusetshenziswa kwe-API eqondile, izici ezifakiwe, kanye nokuncika kwabathengisi. Yakha uhlelo lokushintshana oluhlanganisa ama-endpoints emodeli, amathempulethi asheshayo, izisekelo zokuhlola, kanye nokuvunyelwa kokubusa, ukuze ukuphuma kwesigaba kungabi yinto engasebenzi. Emipayipini eminingi, ukusethwa kwabahlinzeki ababili kanye neziqephu eziphathwayo kunciphisa i-radius yokuqhuma lapho inqubomgomo ishintsha ngobusuku obubodwa.
Ingabe izinqubomgomo zokusetshenziswa komhlinzeki we-AI zingangqubuzana nezimfuneko zikahulumeni noma zempi?
Yebo - lesi simo siqokomisa ukungqubuzana okuqondile mayelana nokuthi imithetho yokusebenzisa yomhlinzeki oyisibonelo ingazivimbela yini izinhlelo zokusebenza ezibucayi. Ohulumeni bavame ukuphoqelela ukuguquguquka okuphezulu, kuyilapho ama-lab engase aphoqelele imingcele eqinile ezimweni ezithile zokusebenzisa. Uma ukhonza amakhasimende omkhakha womphakathi, hlela imigomo yenkontileka ebeka phambili izidingo zomsebenzi futhi engadinga izilawuli noma iziqinisekiso ezahlukene.
Kusho ukuthini ukubambisana kwe-OpenAI-Amazon lapho ungaqhuba khona izinhlelo ze-OpenAI?
Inwebeka lapho amakhasimende angasebenzisa khona futhi aphathe izinhlelo ze-OpenAI-grade ngokuletha ipulatifomu ye-OpenAI Frontier ku-AWS. Empeleni, lokho kungasho izinketho eziningi zengqalasizinda kanye nezindlela ezimbalwa zokuthunyelwa komzila owodwa. Kungaphinde kushintshe imithwalo yemfanelo: izinkinobho eziningi zokushintsha ubunikazi, ukufinyelela, kanye nokusebenza, okungasekela ukuqina kodwa futhi kwandisa ingozi yokucushwa.
Kuyini "i-AI ecacile" ku-AWS, futhi kungani indiza yokulawula ibalulekile?
I-AI ethi “Stateful” isho izinhlelo ezingathwala umongo kuzo zonke izikhathi, ezingazizwa zifana kakhulu ne-ejenti emisebenzini yangempela yokusebenza. Umbuzo wendiza yokulawula ugxile ekutheni ubani ohlela ubunikazi, inkumbulo, kanye nomsebenzi wesikhathi: ifu lakho, ilebhu ye-AI, noma ungqimba lwakho lweplatifomu. Lokho kubalulekile ekubuseni, ekulungiseni amaphutha, kanye nokuphendula kwezehlakalo lapho kukhona okungahambi kahle ngo-2 ekuseni.
Ushintsho "oluthule" lwe-Google API lungaholela kanjani engcupheni yokudalulwa kwedatha ye-Gemini?
Uma ukuziphatha okubalulekile noma okulindelwe ekuphatheni okhiye kushintsha, imikhuba eyayibonakala iphephile - njengokushumeka okhiye - ingaba yingozi ngaphandle kokuqaphela kwamaqembu. Ukubika kubeka lokhu njengenkinga yokuphathwa kwezimfihlo ekhuliswa ukuhlanganiswa kwe-AI, hhayi iphutha elilodwa elihlukanisiwe. Indlela evamile ukuphatha zonke okhiye njengeziyingozi enkulu, ukujikeleza njalo, nokuzigcina ziseceleni kweseva ngemuva kwezilawuli zokufinyelela eziqinile.
Yini entsha kusibuyekezo sohlelo lokusebenza lwe-Gemini sikaFebhuwari 2026, futhi ubani ositholayo?
I-Google igqamisa ukuthuthukiswa kwe-Gemini 3.1 kanye nemodi yokucabanga ethi “Deep Think” ebekelwe izinkinga zesayensi nesitayela sobunjiniyela ezisindayo. Lesi sibuyekezo siphinde sigcizelele ukulingana kwezinga lababhalisile, okusho ukuthi ikhono kanye nokufinyelela kungahluka ngohlelo. Ngokwehlukana, i-Lyria 3 ikhulunywa ngayo njengemodeli yomculo engakhiqiza amathrekhi amafushane avela kumbhalo noma ezithombeni ku-beta, yandisa amathuluzi okudala ohlelo lokusebenza.
Imibuzo Evame Ukubuzwa
Yini eshintsha izikhungo zikahulumeni wase-US ezivimba ukusetshenziswa kobuchwepheshe be-Anthropic?
Iguqula ukukhetha komthengisi kube umthetho wokuthenga, onesikhathi esinqunyiwe sokuphuma esikhundleni sezinqumo zesikhashana, zeqembu neqembu. Uhlaka lwe-Pentagon "lwengozi yokunikezela" luphakamisa izingqinamba futhi lubonise ukuthi ukufaneleka kungabunjwa yinqubomgomo ngaphezu kokufaneleka komkhiqizo. Kubathengi, kugcizelela ukuthi izidingo zomkhakha womphakathi zingadlula kanjani imodeli yokusebenza ekhethwa ngumhlinzeki.
Ilebula lePentagon elithi “ingozi yokunikezela” lingabathinta kanjani osonkontileka kanye nabathengisi abangaphansi komfula?
Umbiko uphakamisa ukuthi umthelela awugcini nje kuma-ejensi; ungasakazekela kosonkontileka abahlangana nesimiso semvelo sikahulumeni. Ngisho noma ungalokothi uthenge imodeli ngqo, i-stack yakho ingazuza imikhawulo ngokusebenzisa izinkontileka eziyinhloko, izigaba zokugeleza, kanye nokuhlolwa kokuhambisana nomthetho. Yingakho "ubani osebenzisa lokho" eba yinkinga yamaphepha ngokushesha, hhayi nje impikiswano yokwakha.
Yini okufanele abathengi bebhizinisi bayenze uma umhlinzeki we-AI oyinhloko ebanjwa ekuvinjelweni kokuthengwa kwe-AI kahulumeni?
Qala ngokumaka lapho umhlinzeki ebonakala khona: ukusetshenziswa kwe-API eqondile, izici ezifakiwe, kanye nokuncika kwabathengisi. Yakha uhlelo lokushintshana oluhlanganisa ama-endpoints emodeli, amathempulethi asheshayo, izisekelo zokuhlola, kanye nokuvunyelwa kokubusa, ukuze ukuphuma kwesigaba kungabi yinto engasebenzi. Emipayipini eminingi, ukusethwa kwabahlinzeki ababili kanye neziqephu eziphathwayo kunciphisa i-radius yokuqhuma lapho inqubomgomo ishintsha ngobusuku obubodwa.
Ingabe izinqubomgomo zokusetshenziswa komhlinzeki we-AI zingangqubuzana nezimfuneko zikahulumeni noma zempi?
Yebo - lesi simo siqokomisa ukungqubuzana okuqondile mayelana nokuthi imithetho yokusebenzisa yomhlinzeki oyisibonelo ingazivimbela yini izinhlelo zokusebenza ezibucayi. Ohulumeni bavame ukuphoqelela ukuguquguquka okuphezulu, kuyilapho ama-lab engase aphoqelele imingcele eqinile ezimweni ezithile zokusebenzisa. Uma ukhonza amakhasimende omkhakha womphakathi, hlela imigomo yenkontileka ebeka phambili izidingo zomsebenzi futhi engadinga izilawuli noma iziqinisekiso ezahlukene.
Kusho ukuthini ukubambisana kwe-OpenAI-Amazon lapho ungaqhuba khona izinhlelo ze-OpenAI?
Inwebeka lapho amakhasimende angasebenzisa khona futhi aphathe izinhlelo ze-OpenAI-grade ngokuletha ipulatifomu ye-OpenAI Frontier ku-AWS. Empeleni, lokho kungasho izinketho eziningi zengqalasizinda kanye nezindlela ezimbalwa zokuthunyelwa komzila owodwa. Kungaphinde kushintshe imithwalo yemfanelo: izinkinobho eziningi zokushintsha ubunikazi, ukufinyelela, kanye nokusebenza, okungasekela ukuqina kodwa futhi kwandisa ingozi yokucushwa.
Kuyini "i-AI ecacile" ku-AWS, futhi kungani indiza yokulawula ibalulekile?
I-AI ethi “Stateful” isho izinhlelo ezingathwala umongo kuzo zonke izikhathi, ezingazizwa zifana kakhulu ne-ejenti emisebenzini yangempela yokusebenza. Umbuzo wendiza yokulawula ugxile ekutheni ubani ohlela ubunikazi, inkumbulo, kanye nomsebenzi wesikhathi: ifu lakho, ilebhu ye-AI, noma ungqimba lwakho lweplatifomu. Lokho kubalulekile ekubuseni, ekulungiseni amaphutha, kanye nokuphendula kwezehlakalo lapho kukhona okungahambi kahle ngo-2 ekuseni.
Ushintsho "oluthule" lwe-Google API lungaholela kanjani engcupheni yokudalulwa kwedatha ye-Gemini?
Uma ukuziphatha okubalulekile noma okulindelwe ekuphatheni okhiye kushintsha, imikhuba eyayibonakala iphephile - njengokushumeka okhiye - ingaba yingozi ngaphandle kokuqaphela kwamaqembu. Ukubika kubeka lokhu njengenkinga yokuphathwa kwezimfihlo ekhuliswa ukuhlanganiswa kwe-AI, hhayi iphutha elilodwa elihlukanisiwe. Indlela evamile ukuphatha zonke okhiye njengeziyingozi enkulu, ukujikeleza njalo, nokuzigcina ziseceleni kweseva ngemuva kwezilawuli zokufinyelela eziqinile.
Yini entsha kusibuyekezo sohlelo lokusebenza lwe-Gemini sikaFebhuwari 2026, futhi ubani ositholayo?
I-Google igqamisa ukuthuthukiswa kwe-Gemini 3.1 kanye nemodi yokucabanga ethi “Deep Think” ebekelwe izinkinga zesayensi nesitayela sobunjiniyela ezisindayo. Lesi sibuyekezo siphinde sigcizelele ukulingana kwezinga lababhalisile, okusho ukuthi ikhono kanye nokufinyelela kungahluka ngohlelo. Ngokwehlukana, i-Lyria 3 ikhulunywa ngayo njengemodeli yomculo engakhiqiza amathrekhi amafushane avela kumbhalo noma ezithombeni ku-beta, yandisa amathuluzi okudala ohlelo lokusebenza.