🧪 I-Florida AG iqala uphenyo nge-OpenAI ngaphambi kwe-IPO engaba khona ↗
Ummeli-jikelele waseFlorida uvule uphenyo nge-OpenAI, futhi lokho akuyona into encane yokulawula. Uphenyo luphathelene nezinkinga zokuphepha kwezwe nokuthi idatha ye-OpenAI noma ubuchwepheshe be-AI bungagcina buvezwe yini ezitheni zakwamanye amazwe.
Umbuso uphinde waxhumanisa i-ChatGPT nochungechunge lwezingozi, okuhlanganisa nokukhuthaza ukuzilimaza kanye nokusetshenziswa okusolwayo ekuhleleni ukudubula kwabantu abaningi muva nje. Kubikwa ukuthi amasamanisi asesendleleni, kanti i-OpenAI ayizange iphendule ku-Reuters ngesikhathi kushicilelwa - ngakho-ke ingcindezi yezomthetho lapha ibonakala ikhuphuke kakhulu. (Reuters)
🖥️ I-Intel ne-Google zizosebenzisa ama-CPU e-AI ngokubambisana okwandisiwe ↗
I-Intel ne-Google bandisa ubudlelwano babo bengqalasizinda ye-AI, kodwa iphuzu elithakazelisayo kakhulu yilapho imakethe ibonakala ishintshashintsha khona. Akukhona nje kuphela ngama-accelerator amakhulu manje - bathambekele kakhulu kuma-CPU kanye nama-processor engqalasizinda ngokwezifiso ukuze bathole isiqondiso kanye neminye imisebenzi ebanzi ye-AI. Mhlawumbe ayikhangi kangako, kodwa isabalulekile kakhulu.
I-Google izoqhubeka nokusebenzisa ama-chip e-Intel Xeon, okuhlanganisa ne-Xeon 6, kuyilapho izinkampani ezimbili zijulisa umsebenzi wazo kuma-IPU enziwe ngokwezifiso. Isibonakaliso esikhulu ukuthi i-AI e-ejenti iqhuba isidingo sezinhlelo ezilinganiselayo, hhayi nje amalungelo okuqhosha e-GPU angavuthiwe, okuzwakala njengokuphenduka kwenkathi ye-AI ngokusobala. (Reuters)
🪖 Ukuxoshwa kwePentagon kwi-Anthropic kuvula iminyango yezimbangi ezincane ze-AI ↗
Ukuhlukana kwe-Anthropic nePentagon kungazelelwe kudala isikhala sezinkampani ezincane ze-AI zokuzivikela, futhi lokho kungase kugcine sekudwebe kabusha imephu yabaphakeli be-AI yezempi. I-Reuters ibika ukuthi izinkampani ezifana ne-Smack Technologies kanye ne-EdgeRunner AI manje sezidonsela isithakazelo kojenene, amaqembu okuthenga, kanye nabatshalizimali ababengakayiboni indlela yabo ngaphambili.
Ukungezwani kwePentagon ne-Anthropic kugcizelele ukuthi kungaba yingozi kangakanani ukuncika kakhulu kumthengisi oyedwa. Manje labo badlali abancane babona izinkontileka zihamba ngokushesha, kucelwa ama-demos amaningi, izingxoxo eziningi, kuphuthuma kakhulu - kufana nomnyango omkhulu ovaliwe futhi iminyango eyisithupha eseceleni ivuleka. (Reuters)
🧠 Okukhethekile: Imithombo ithi i-Anthropic ihlela ukwakha ama-chip ayo e-AI ↗
Kubikwa ukuthi i-Anthropic ihlola ukuthi ingaziklama yini ama-chip ayo e-AI, okuzwakala sengathi kukhulu ngoba kunjalo. Umzamo usasese sekuqaleni futhi inkampani ingase inqume ukungawuqhubeki nawo, kodwa iqiniso lokuthi icabangela lesi sinyathelo libonisa ukuthi ukucindezelwa kwama-chip kusalokhu kukhulu kangakanani kuma-lab asemngceleni.
Isidingo sikaClaude sikhuphuke kakhulu, kanti i-Anthropic isivele isebenzisa ingxube yama-TPU e-Google nama-chip e-Amazon. I-Reuters ithi ukwakha i-chip ye-AI ethuthukisiwe kungabiza cishe isigamu sezigidi zamaRandi, ngakho-ke lokhu akuyona into evamile kodwa "ingabe nathi sifuna ukuba yinkampani eyingxenye yama-chip?". (I-Reuters)
⚖️ I-xAI ka-Elon Musk imangalela iColorado ngomthetho omusha we-AI wesifundazwe ↗
I-xAI imangalele iColorado ukuze ivimbele ukuphoqelelwa komthetho we-AI wesifundazwe, yandisa impi yokuthi ngabe ukuqapha i-AI kufanele kwenzeke esifundazweni ngasinye noma ezingeni likahulumeni. Leli cala lihlose imithetho yezinhlelo "ezinobungozi obukhulu" ezisetshenziswa ezindaweni ezifana nemisebenzi, izindlu, imfundo, ukunakekelwa kwempilo kanye nezezimali.
I-xAI ithi umthetho wephula iSichibiyelo Sokuqala futhi uzophoqa izinguquko kuGrok mayelana nokuhlukahluka kanye nezindinganiso zokubandlulula. Ngakho-ke, yebo, lokhu kuyimpi yokuthobela imithetho - kodwa futhi kuyisilingo esibanzi sezepolitiki sokuthi ubani obhala incwadi yemithetho ye-AI kwasekuqaleni. (Reuters)
🛠️ I-Meta idlulisela onjiniyela abaphezulu eqenjini elisha lamathuluzi e-AI ↗
I-Meta ihambisa onjiniyela abaphezulu eyunithi entsha ye-Applied AI Engineering, futhi ukudluliselwa kubikwa ukuthi akusekho okukhethwa kukho. Leyo mininingwane ngokwayo isho okuningi. Le nkampani iphatha leli qembu njengenye yezinto eziza kuqala njengoba ihlela kabusha i-AI futhi ilungiselela ukudilizwa.
Le yunithi igxile kumathuluzi nokuhlolwa kwama-ejenti e-AI angabhala ikhodi futhi aphathe imisebenzi eyinkimbinkimbi ngokuzimela, lapho abasebenzi abangabantu beqapha umkhiqizo. Kuyinguqulo engokoqobo yomugqa othi "I-AI izoshintsha indlela esisebenza ngayo" - akuseyona nje isiqubulo, kufana nohlelo lokusebenza lwangaphakathi. (Reuters)
Imibuzo Evame Ukubuzwa
Kungani lezi zindaba ze-AI zihlanganiswa ndawonye njengamanje?
Bakhomba ushintsho olukhulu olufanayo: I-AI ayisekho nje ngekhwalithi yemodeli. Lesi sihloko sixhumanisa imithethonqubo, ukunikezwa kwama-chip, amacala, ubudlelwano bokuzivikela, kanye nezinguquko zangaphakathi zeqembu njengezindawo zokucindezela zokuncintisana. Uma zihlanganiswa, lezi zindaba zibonisa ukuthi ubuholi be-AI manje buncike kunqubomgomo, ingqalasizinda, ukubambisana, kanye nokusebenza, hhayi nje ekuthuthukisweni kocwaningo.
Kusho ukuthini uphenyo lwe-OpenAI embonini ebanzi ye-AI?
Uphenyo luvame ukukhombisa ukuthi izinkampani ze-AI zibhekene nokuhlolwa okukhulu ngokuphathwa, ukuncintisana, ukuphepha, noma imikhuba yebhizinisi. Ngisho noma inkampani eyodwa igxile, imiphumela ivame ukusabalala emakethe ngokulindela okuqinile okuvela kubalawuli, ozakwethu, kanye namakhasimende. Kuyo yonke imigudu eminingi, lokho kuholela ekuqapheleni okwengeziwe mayelana nokucaca, ukuthobela imithetho, kanye nesakhiwo sezivumelwano.
Ibaluleke kanjani imizamo ye-Intel-Google AI CPU uma ukunakwa okuningi kuya kuma-GPU?
Umsebenzi we-AI ogxile ku-CPU ubalulekile ngoba akuwona wonke umthwalo womsebenzi odinga iphrofayili efanayo yehadiwe. Indlela evamile ukusebenzisa ama-CPU ukuze kutholakale izinkomba, ukuhlelwa kabusha, ukuthunyelwa kwebhizinisi, noma imisebenzi ebucayi lapho ukuguquguquka kanye nengqalasizinda ekhona kusabalulekile. Ukufakwa kwe-Intel ne-Google kulesi sihloko kusikisela ukuthi ukuncintisana kuyanda ngale kwezihloko ze-GPU kuya isu lokubala elibanzi.
Kungani ukuhlukana kwePentagon ne-Anthropic kungabaluleka ngale kokubambisana okukodwa?
Ubudlelwano bezokuvikela buvame ukwakha ukwethembeka, umfutho wokuthengwa kwempahla, kanye nombono womphakathi wezinkampani ze-AI zasemngceleni. Ukuhlukana kungabaluleka ngoba kungabonisa izinto eziza kuqala ezahlukene mayelana nokufinyelela, ukuphepha, inqubomgomo, noma amazinga okuthunyelwa. Ngokubanzi, kugqamisa indlela ukwamukelwa kwe-AI kahulumeni okuncike ngayo hhayi kuphela emandleni obuchwepheshe kodwa futhi nasekuthembeni, ekuhambisaneni, kanye nasekwakhekeni kwenkontileka.
Yini okufanele abafundi bayifunde emibikweni mayelana nama-chip e-Anthropic noma ihadiwe eyenziwe ngokwezifiso?
Okubalulekile ukuthi izinkampani ezihamba phambili ze-AI zifuna ukulawula okwengeziwe ekusebenzeni, izindleko, kanye nemikhawulo yokuhlinzeka. Ngokuvamile, ama-chip enziwe ngokwezifiso noma isu lehadiwe eliqinile lingathuthukisa ukusebenza kahle futhi linciphise ukuncika eqenjini elincane labathengisi. Ngokwemigomo yokukhiqiza, lokho kungathinta izindleko zokukhonza imodeli, izinhlelo zokukhulisa, kanye nokuthi izinhlelo ezintsha zifinyelela ngokushesha kangakanani kumakhasimende.
Icala le-xAI libaluleke kangakanani kubasebenzisi bomkhiqizo noma abathengi bebhizinisi?
Icala alishintshi njalo ukufinyelela komkhiqizo ngokushesha, kodwa lingadala ukungaqiniseki mayelana nokusebenza, idumela, noma izinqumo zesikhathi esizayo. Abathengi bebhizinisi bavame ukubheka imiphumela yokungena njengokuphazamiseka kobuholi, ukuqaliswa okulibazisekile, noma ulimi oluqinile lwezinkontileka. Lesi sihloko sibonakala sihlela le ndaba njengengxenye yephethini ebanzi yokwakha ingcindezi yezomthetho ezungeze izinkampani ze-AI.
Ukubuyekezwa kwethimba le-AI le-Meta kusikiselani ngokuthi imakethe iyaphi?
Ukuguqulwa kweqembu kuvame ukukhombisa ukuphuthuma mayelana nokuqaliswa, ukugxila komkhiqizo, noma ukuhambisana kwangaphakathi. Ezimakethe ze-AI ezihamba ngokushesha, izinkampani zivame ukuhlela kabusha lapho zidinga ubunikazi obucacile ocwaningweni, ingqalasizinda, kanye nokwenza imali. Kubafundi, umphumela ongokoqobo ukuthi ukwakheka kwenhlangano ngokwayo sekuyinto yokuncintisana, ikakhulukazi njengoba izinkampani zincintisana ukuguqula ikhono le-AI libe imikhiqizo esebenzisekayo.
Amabhizinisi kufanele ayichaze kanjani le nhlanganisela yezindaba zemboni ye-AI ngaphandle kokusabela ngokweqile?
Ilensi engcono kakhulu ukuhlukanisa umsindo wesihloko noshintsho lwesakhiwo. Leli qoqo lezindaba lisikisela ukuthi imboni ye-AI iyavuthwa, inengcindezi eyengeziwe ekuhambisaneni nemithetho, ukuqina kwehadiwe, isu lozakwethu, kanye nokuzithiba kokuphatha. Kumabhizinisi, lokho ngokuvamile kusho ukuhlola abathengisi ngokuthembeka kanye nokuphatha kanye nokusebenza kwemodeli eluhlaza noma umfutho wokuqalisa.