⚖️ I-Anthropic imangalela ukuvimba ukufakwa ohlwini lwePentagon oluvinjelwe ngenxa yemikhawulo yokusebenzisa i-AI ↗
I-Anthropic ithathe impi yayo nePentagon yayisa enkantolo ngemuva kokubizwa ngokuthi ingozi yokuthengiswa kwempahla. Le nkampani ithi lesi sinyathelo sasiwukuziphindiselela okungekho emthethweni ngokwenqaba ukuxegisa izivikelo ezizungezile izikhali ezizimele kanye nokuqapha kwasekhaya - okubeka le mpikiswano phakathi kwezindawo ezishisa kakhulu kunqubomgomo ye-AI njengamanje. (Reuters)
Uhulumeni ufuna ukuguquguquka "kwanoma yikuphi ukusetshenziswa okusemthethweni," kuyilapho i-Anthropic ithi amalabhorethri azimele kufanele asavunyelwe ukudweba imigqa eqinile yokuphepha. Lokhu kube ngenye yezivivinyo ezicacile kuze kube manje zokuthi inkampani ye-AI ingayiphika yini imigomo yezempi ngaphandle kokuphoqelelwa. (Reuters)
🧑💻 Izisebenzi kulo lonke i-OpenAI kanye ne-Google zisekela icala lika-Anthropic ngokumelene nePentagon ↗
Lelo cala lasheshe lathola ukwesekwa okuvela ngaphakathi kwamalebhu aphikisanayo. Cishe abasebenzi abangu-40 abavela ku-OpenAI kanye ne-Google basekela i-Anthropic esifinyezweni sobungani, bephikisana ngokuthi ukuziphindiselela ezinkampanini ngemigqa ebomvu yokuphepha kwe-AI kunciphisa intshisekelo yomphakathi. (I-Verge)
Ngakho-ke yebo, abancintisana nabo bazithole besohlangothini olufanayo - okungenani kuleli phuzu. Lesi sihloko sigxile ekuqapheni okukhulu kanye nezikhali ezizimele ezingathembekile, okwenza yonke indaba izwakale ingafani ne-Silicon Valley sniping ejwayelekile kodwa ifane nomugqa wangempela wemboni esihlabathini... noma mhlawumbe udaka. (The Verge)
🛡️ I-OpenAI ithola i-Promptfoo ukuze ivikele ama-ejenti ayo e-AI ↗
I-OpenAI ithe ithenga i-Promptfoo, inkampani entsha egxile ekuvikeleni amamodeli ezilimi ezinkulu ekuhlaselweni okuphikisayo. Uhlelo luwukufaka ubuchwepheshe bayo ku-OpenAI Frontier, ipulatifomu yebhizinisi yenkampani yama-ejenti e-AI. (TechCrunch)
Kuyisinyathelo esibalulekile. Wonke umuntu ufuna i-AI e-ejensi yenze okwengeziwe, ngokushesha, yonke indawo - kodwa uhlangothi lwezokuphepha selusele ngemuva, kungakhathaliseki ukuthi luncane noma lukhulu. Lesi sivumelwano sibonisa ukuthi i-OpenAI icabanga ukuthi umjaho omkhulu olandelayo awusiwo owama-ejensi ahlakaniphile kuphela, kodwa futhi nawaphephile. (TechCrunch)
🧪 I-Anthropic yethula ithuluzi lokubuyekeza ikhodi ukuhlola ukugcwala kwekhodi ekhiqizwe yi-AI ↗
I-Anthropic yethule i-Code Review ngaphakathi kwe-Claude Code, ehlose amaqembu acwila ezicelweni zokudonsa ezidalwe ngamathuluzi okubhala ikhodi e-AI. Isebenzisa ama-ejenti amaningi ngesikhathi esisodwa ukuskena ikhodi, ukumaka izinkinga ze-logic, ubukhali bezinga, nokushiya amazwana ngqo ku-GitHub. (TechCrunch)
Inkulumo ilula ngokwanele - i-AI ikhiqiza ikhodi ngokushesha kunokuba abantu bangayibuyekeza ngokuhlakanipha, ngakho-ke manje i-AI kufanele ibuyekeze i-AI. Ilungiselelo elidla umsila wayo kancane, kodwa elisebenzayo. I-Anthropic ithi ihloselwe abasebenzisi bebhizinisi asebevele bebona umkhiqizo omkhulu wekhodi ovela ku-Claude Code. (TechCrunch)
💰 I-Nscale esekelwa yiNvidia ibiza u-$14.6 billion emzuliswaneni omusha wokuxhasa ngezimali ↗
Inkampani yengqalasizinda ye-AI yase-UK i-Nscale iqoqe amaRandi ayizigidi eziyizinkulungwane ezimbili emzuliswaneni we-Series C, yafinyelela esilinganisweni samaRandi ayizigidi eziyizinkulungwane eziyi-14.6. Abasekeli babehlanganisa i-Aker, i-8090 Industries, i-Nvidia, i-Citadel, i-Dell, kanye ne-Jane Street - okulingana nevoti elikhulu lokuzethemba. (Reuters)
Lokhu kubalulekile ngoba akuyona enye indlela yokuqalisa noma yokulungisa i-chatbot. Kuyicala lokukhetha nokushaya ihlombe - ukubala, ingqalasizinda, umthamo, yonke imishini esindayo ngemuva kwesihenqo. Mhlawumbe akuyona into ekhangayo, kodwa yilapho imali eningi manje inyathelwa khona. (Reuters)
🧠 Ama-AMI Labs kaYann LeCun aqoqe amaRandi ayizigidi eziyizinkulungwane ezingu-1.03 ukwakha amamodeli omhlaba ↗
Ibhizinisi elisha likaYann LeCun, i-AMI Labs, liqoqe amaRandi ayizigidi eziyizinkulungwane ezingu-1.03 ngenani lamaRandi ayizigidi eziyizinkulungwane ezingu-3.5 ngaphambi kokusetshenziswa kwemali. Le nkampani iphishekela "amamodeli omhlaba" - izinhlelo ze-AI ezenzelwe ukufunda eqinisweni ngokwalo kunokuba zifunde kakhulu olimini. (TechCrunch)
Lokho kuyinguquko eqondile yefilosofi ekuvumelaneni kwamanje okunzima kwe-LLM, futhi kuyinto ecacile kakhulu. ULeCun ubelokhu ephikisana iminyaka eminingi ngokuthi amamodeli olimi anamuhla ngeke asiyise ekuhlakanipheni kwezinga lomuntu, ngakho-ke lokhu kukhushulwa kuguqula leyo mpikiswano ibe ukuhlolwa okubiza kakhulu. (TechCrunch)
🇨🇳 Izikhungo zobuchwepheshe zaseShayina zikhangisa i-ejenti ye-OpenClaw AI naphezu kwezixwayiso zokuphepha ↗
Ohulumeni bendawo abaningana baseShayina basekela i-OpenClaw, i-ejenti ye-AI esakazeka ngokushesha naphezu kokukhathazeka kokuphepha okuhlobene nokufinyelela kwayo kudatha yomuntu siqu. Ngakho-ke isimo sengqondo esisemthethweni sibonakala siwukuthi: yebo, kunezingozi - futhi yebo, ake sikwenze kube ngcono noma kunjalo. (Reuters)
Lokho kuhlukana kuyindaba. Izikhungo zendawo zifuna ukuthuthuka komnotho kanye nomfutho we-ecosystem, kuyilapho abalawuli bexwayisa ngokuchayeka kwedatha. Sekuyiphethini ejwayelekile ye-AI njengamanje - qala ngokusheshisa, ulungise izivikelo kamuva, noma kubonakala kanjalo. (Reuters)
Imibuzo Evame Ukubuzwa
Kungani i-Anthropic imangalela iPentagon ngemikhawulo yokusebenzisa i-AI?
I-Anthropic ithi iPentagon ibize lokhu njengengozi yochungechunge lokuphakelwa kwempahla ngemuva kokuba inkampani yenqabe imigomo ebanzi engadlulela ezikhalini ezizimele kanye nokuqapha kwasekhaya. Lokho kwenza icala libe ngaphezu kwesimo somthengisi. Ihlola ukuthi ngabe ilebhu ye-AI ingagcina yini imingcele eqinile yokuphepha endaweni yayo futhi iqhubeke nokuncintisana ngomsebenzi kahulumeni ngaphandle kokujeziswa.
Kungani abasebenzi be-OpenAI kanye ne-Google besekela i-Anthropic kule mpikiswano yokuphepha kwe-AI?
I-amicus brief ikhombisa ukuthi abantu abaningi ngaphakathi kwamalabhorethri aphikisanayo babona lokhu njengenkinga yokuphepha kwe-AI ebeka isibonelo, hhayi nje impi phakathi kwenkampani eyodwa ne-ejensi eyodwa. Ukukhathazeka kwabo ukuthi ukujezisa umthengisi ngokugcina imigqa ebomvu kungacindezela imakethe ebanzi ukuthi yenze buthaka izivikelo. Empeleni, lokho kungabumba indlela okuxoxiswana ngayo ngezinkontileka ze-AI zokuvikela zesikhathi esizayo kanye nezemboni yomphakathi.
Yini engashintsha icala le-Anthropic-Pentagon ngenqubomgomo ye-AI kanye nezinkontileka zokuzivikela?
Uma i-Anthropic iwina, izinkampani ze-AI zingase zibe nesisekelo esiqinile sokuchaza ukusetshenziswa okungamukeleki ngisho nalapho zithengisa ezindaweni ezibucayi zikahulumeni. Uma ilahlekelwa, izinhlangano zingathola ithuba lokufuna imigomo ebanzi "yokusetshenziswa okusemthethweni" kubaphakeli. Noma ngabe yikuphi, le mpikiswano cishe izothinta ulimi lokuthengwa kwempahla, ukubuyekezwa kwezingozi, kanye nendlela izivikelo zokuphepha ezibhalwe ngayo ezivumelwaneni zokuzivikela.
Kungani i-OpenAI ithenge i-Promptfoo yama-ejenti e-AI?
I-Promptfoo yaziwa ngokuhlola amamodeli amakhulu olimi ngokumelene nezifiso zokuphikisana kanye nobunye ubuthakathaka bokuphepha. Ukugoqa lolo hlobo lwamathuluzi epulatifomu ye-ejenti yebhizinisi le-OpenAI kusikisela ukuthi inkampani ibona ukuthunyelwa okuphephile njengenzuzo yokuncintisana kunokuba umsebenzi oseceleni. Njengoba ama-ejenti e-AI enza umsebenzi obalulekile, ukuhlolwa kokuqina nokuhlukunyezwa kuba nzima kakhulu ukunganakwa.
Amaqembu angabhekana kanjani nokwanda kwekhodi ekhiqizwe yi-AI ngokuphephile kakhudlwana?
Isici esisha se-Anthropic se-Code Review ngaphakathi kwe-Claude Code sihloselwe amaqembu agcwele izicelo zokudonsa ezikhiqizwe ngamathuluzi okubhala ikhodi e-AI. Sisebenzisa ama-ejenti amaningi ngesikhathi esifanayo ukuthola izinkinga ze-logic, ubukhali bezinga, nokushiya amazwana ku-GitHub. Ngokuvamile, amathuluzi anjengalawa asiza ekuhloleni ivolumu, kodwa ababuyekezi abangabantu basabalulekile ekwakhiweni, kumongo, nasekuvunyelweni kokugcina.
Kungani ingqalasizinda ye-AI ithola utshalomali oluningi kangaka njengamanje?
Umjikelezo wakamuva wokuxhaswa ngezimali we-Nscale uqokomisa ukuthi imboni ye-AI isathela imali enkulu ekubaleni, engqalasizinda, kanye namandla. Lokho kusebenzisa imali kungase kungabonakali kangako njengokuqaliswa kwemodeli ekhangayo, kodwa kuyisisekelo sako konke okunye. Lapho isidingo sokuqeqeshwa nokuthunyelwa siqhubeka sikhula, izinkampani ezithengisa amapiki namafosholo zivame ukuba ezinye zezinqobi ezinkulu.
Ayini amamodeli omhlaba, futhi kungani uYann LeCun ebheja kuwo?
Amamodeli omhlaba ayizinhlelo ze-AI ezenzelwe ukufunda endleleni umhlaba oziphatha ngayo, kunokuthembela kakhulu kudatha yolimi. Lokho kubalulekile ngoba uYann LeCun ubelokhu ephikisana ngokuthi amamodeli olimi ngokwawo cishe awakwazi ukufinyelela ubuhlakani bezinga lomuntu. I-AMI Labs iguqula lowo mbono ube ukubheja okukhulu kwezentengiselwano endleleni ehlukile yocwaningo lwe-AI oluthuthukisiwe.
Kungani izikhungo zobuchwepheshe zaseShayina zisekela i-OpenClaw naphezu kwezixwayiso zokuphepha?
Indaba ye-OpenClaw ikhombisa ukuhlukana okujwayelekile ngaphakathi kwezimakethe zobuchwepheshe ezihamba ngokushesha: ohulumeni basekhaya bafuna ukukhula, uxhaso, kanye nomfutho wesimiso sezinto eziphilayo, kuyilapho abalawuli bekhathazeka ngokudalulwa kwedatha kanye nokuphepha. Ukusekela i-ejenti naphezu kwezixwayiso kusikisela ukuthi izikhuthazo zomnotho ziyanda kwezinye izindawo. Kwabaqapheli, kuyisikhumbuzo esinye sokuthi ukwamukelwa kuvame ukuhamba ngokushesha kunokuqapha.