💰 I-Sequoia igxuma buthule ingena emzuliswaneni omkhulu we-Anthropic ↗
I-Sequoia - esivele ixhumene nama-lab amaningi amakhulu e-AI - kubikwa ukuthi ijoyina ukukhushulwa okukhulu kwe-Anthropic. Lolu uhlobo lwesinyathelo oluguqula yonke inkulumo ethi "izingxabano, akukho zingxabano" ibe ngcono, kungakhathaliseki ukuthi ubani uyakuvuma noma cha.
Kuthiwa lo mjikelezo uhlanganisa nezinye izivivinyo zesibhakela esindayo, uqhubekisela phambili i-Anthropic kulowo mzila ophezulu, onenani elikhulu. Ukuzwakala kwebhamuza kuhlala isikhathi eside. Lokhu kungaba yinto evamile entsha, ecasulayo.
📢 I-ChatGPT iqala ukudlala ngezikhangiso - empeleni kulokhu ↗
Kuthiwa i-OpenAI ihlola izikhangiso zabanye abasebenzisi base-US ezigabeni ezishibhile, kanti ezigabeni ezikhokhelwayo eziphezulu zihlala zingenazo izikhangiso. Isithembiso siwukuthi izikhangiso ngeke zithinte izimpendulo - okuqinisekisayo ngokombono, noma ngabe i-trust reflex ishintsha kancane.
Ingqikithi enkulu ilula: ukuqagela kuyabiza, futhi okubhaliselwe kukodwa kungase kungambozi yonke into unomphela... noma kubonakala kanjalo. Noma kunjalo, okokuqala uma ubona "kuxhaswe" noma kuphi eduze kwe-chatbot, kukhona okushintshayo ekhanda lakho. Isimo siyashintsha.
📚 Abashicileli bazama ukuhlanganisa icala lokuqeqeshwa kwe-AI le-Google ↗
Iqembu labashicileli lizama ukujoyina icala elisola i-Google ngokusebenzisa imisebenzi enamalungelo okushicilela ukuqeqesha izinhlelo zayo ze-AI. Le mpi yezomthetho iyaqhubeka nokukhula, njengoqhekeko eqhweni ohlala uluzwa kodwa ungakwazi ukuluthola.
Uma inkantolo ibavumela ukuba bangene, icala lingacacisa ukuthi “imvume” kanye “nenkokhelo” kufanele kusho ukuthini ngemininingwane yokuqeqeshwa. Wonke umuntu ufuna isibonelo - okungcono kakhulu esizobasiza, ngokusobala.
🕳️ Kubikwa ukuthi icebo "lokujova ngokushesha" liphazamisa iGemini ngemininingwane yokuhlangana ↗
Abacwaningi bachaze ukuhlaselwa kwesitayela "sokufaka okungaqondile ngokushesha" lapho imiyalelo enonya ifihlwa khona ngaphakathi kokuqukethwe okubukeka kujwayelekile, bese umsizi eyilandela kamuva lapho umsebenzisi ebuza okuthile okungenacala. Akukho malware, akukho ubuqili - umbhalo ohlotshiswe ngezikhali, ubuhle obumangalisayo futhi obubi kakhulu.
Kuyisikhumbuzo sokuthi “i-LLM ifunda umbhalo ongathembekile” akuyona into emnandi - iyindawo esongelayo ngokuphelele. Njengokuvumela abantu ongabazi ukuthi bafake amanothi emaphaketheni akho usuku lonke, bese bezenza sengathi bamangala lapho omunye wabo eyisithiyo.
🎮 I-CEO yeRazer ithi abadlali “sebevele bayayithanda i-AI” - bayayizonda nje le lebula ↗
Ingxoxo kaRazer nge-CES igxile ku-AI njengethuluzi eliwusizo lokusebenza kokuthuthukiswa kwemidlalo - i-QA, i-iteration, into enjalo - kanye neminye imiqondo efana nomsizi ezwakala iwusizo kancane, ingxenye yesayensi ye-fi.
Futhi empeleni bayavuma inkinga yokubeka uphawu: abadlali abafuni "i-AI slop," kodwa bafuna amathuluzi ahlakaniphile kanye nokuhlangenwe nakho okubushelelezi. Kubize ngokuthi "ukusiza" bese abantu bavuma ngekhanda. Kubize ngokuthi "AI" bese abantu bafinyelela ama-pitchforks... ngezinye izikhathi.
⚖️ Inkantolo ibeka imithetho yabameli abasebenzisa i-AI ekhiqizayo ↗
Inkantolo ishicilele isiqondiso esisho ukuthi: vele, sebenzisa i-genAI - kodwa usengumnikazi womsebenzi. Awukwazi ukunikeza ukwahlulela kwakho kobungcweti kumkhiqizi wombhalo bese ushaqeka lapho esungula okuthile ngokuzethemba.
Ngokuthakazelisayo, ukudalulwa akudingeki ngaphandle kokuthi ijaji licele - kodwa umlayezo wokuzibophezela yiwona oyinhloko yalokhu. I-AI ingaklama futhi ihlele… futhi ibukeke njenge-intern eqiniseka kakhulu enekhono lokubhala izindaba eziqanjiwe.
Imibuzo Evame Ukubuzwa
Kusho ukuthini ukuthi iSequoia ijoyina i-mega-round ka-Anthropic ekutshalweni kwezimali kwe-AI kanye nezingxabano?
Kusikisela ukuthi abatshalizimali abakhulu bangaqhubeka nokusekela amalebhu amaningi aphezulu e-AI ngesikhathi esisodwa, okubikezelwa ukuthi kuzovuselela impikiswano ethi "izingxabano, akukho zingxabano". Lapho isikhwama esifanayo sihileleke kumalebhu amaningana, abantu baqala ukuhlola izikhuthazo, ukufinyelela, kanye nemikhawulo yokuncintisana. I-mega-round ebikiwe iphinde igcizelele ukuqhubekela phambili ekuhlolweni okukhulu kanye nokuhlolwa kwamanani okuphakeme, noma ngabe "ama-bubble vibe" asalokhu ekhona emoyeni.
Ingabe i-ChatGPT ithola izikhangiso kumazinga wamahhala noma ashibhile, futhi ingabe zizothinta izimpendulo?
Umbiko uthi i-OpenAI ihlola izikhangiso zabanye abasebenzisi base-US ezigabeni ezishibhile, kuyilapho amazinga aphezulu ekhokhelwayo ehlala engenazikhangiso. Iphinde ithi izikhangiso ngeke zithinte izimpendulo, okuzwakala kuqinisekisa ephepheni kodwa kungashintsha indlela abantu ababona ngayo ukwethembana. Ingqikithi iwumnotho: ukuphetha kuyabiza, futhi okubhaliselwe kungase kungambozi yonke into unomphela.
Kungani abashicileli bezama ukujoyina icala lokuqeqeshwa kwe-AI le-Google?
Iqembu labashicileli lifuna ukujoyina icala elisola i-Google ukuthi isebenzise imisebenzi enamalungelo okushicilela ukuqeqesha izinhlelo ze-AI. Uma inkantolo ibavumela, icala lingacacisa ukuthi "imvume" kanye "nokukhokha" kufanele kubukeke kanjani ngedatha yokuqeqesha. Amaqembu amaningi avame ukusho ingcindezi eyengeziwe yesibonelo esicacile - ikakhulukazi mayelana nokuthi ubani okhokhelwayo, futhi ngaphansi kwaziphi izimo.
Kuyini ukuhlasela "kokujova okuqondile okusheshayo", futhi kungani kuyinto enkulu ezindabeni zobuchwepheshe be-AI?
Kuwukuhlasela lapho imiyalelo enonya ifihlwa ngaphakathi kokuqukethwe okubukeka kujwayelekile, futhi umsizi uyayilandela kamuva lapho umsebenzisi enza isicelo esingenacala. Inkinga eyinhloko ukuthi imodeli ifunda umbhalo ongathembekile, iguqula amadokhumenti nemiyalezo yansuku zonke ibe yindawo engaba yingozi. Kuyathandeka ngoba ingasebenza ngaphandle kwe-malware yendabuko - ulimi oluhlomile olufakwe kokuqukethwe.
Kungani abadlali begeyimu bengathandi ilebula ethi “AI” kodwa basafuna amathuluzi e-AI?
I-CEO kaRazer ithi abadlali sebevele bayayithanda izinzuzo ezisebenzayo - i-QA esheshayo, ukuphindaphinda okubushelelezi, kanye nabasizi bomsebenzi - kodwa basabela kabi ekufakweni komkhiqizo. Ukukhathazeka kuvame ukuba “i-AI slop,” noma okuqukethwe okuzwakala kungenzi lutho futhi kungelona iqiniso. Ukuyibiza ngokuthi “usizo” noma isici sosizo kungenza izwakale njengethuluzi elithuthukisa ulwazi kunokufaka esikhundleni sokudala.
Isho ukuthini imithetho yenkantolo kubameli abasebenzisa i-AI ekhiqizayo, futhi ingabe kufanele bayidalule?
Isiqondiso esichaziwe siqondile: abameli bangasebenzisa i-AI ekhiqizayo, kodwa bahlala benomthwalo wemfanelo ngomsebenzi futhi abakwazi ukunikeza ukwahlulela kochwepheshe kumkhiqizi wombhalo. Ingozi iwukubona izinto ezingekho - amaqiniso noma izingcaphuno ezisungulwe ngokuqiniseka - ngakho ukuqinisekiswa kanye nokuziphendulela kuhlala kuyinto ebalulekile. Kubikwa ukuthi ukudalula akudingeki ngaphandle kokuthi ijaji libuze, kodwa umyalezo uthi: ungumnikazi womphumela.