🪓 Okukhethekile: Ukuhlela kwe-Meta kukhuphula ukudilizwa njengoba izindleko ze-AI zikhuphuka ↗
Kubikwa ukuthi i-Meta inesisindo esikhulu sokunciphisa - cishe ingxenye yesihlanu yenkampani - ukusiza ukumboza ukusetshenziswa kwayo kwe-AI kanye ne-data center okukhuphukayo. Umqondo uzwakala ulula: ama-chip amaningi, ukubala okuningi, abantu abambalwa.
Ingxenye ephawulekayo ukuthi lesi skripthi manje sibonakala sijwayelekile kangakanani. Ukutshalwa kwezimali kwe-AI kuphathwa njengomoya-mpilo, kuyilapho konke okunye kulinganiswa kube yizicucu, noma kubonakala kanjalo.
🎞️ I-ByteDance imisa ukwethulwa kwemodeli ye-AI yevidiyo ngemuva kwezingxabano ze-copyright, kubika i-The Information ↗
Kubikwa ukuthi i-ByteDance imise ukukhishwa kwemodeli yokukhiqiza amavidiyo ngemuva kokubhekana nezingxabano ze-copyright. Lokho kuyisikhumbuzo esikhulu sokuthi ku-AI ekhiqizayo, imodeli iyinxenye yendaba kuphela - idatha yokuqeqeshwa ingaphenduka emuva.
Kuluhlobo lokubuyela emuva olubonakala lulandela inqubo uma luvela ngaphandle, kodwa lungamisa umfutho ngokushesha. I-AI yevidiyo iyaqhubeka nokugijima, kodwa indawo yezomthetho ngaphansi kwayo isazwakala njengopende omanzi.
🔌 Indlela yokusebenzisa ukuhlanganiswa kohlelo lokusebenza olusha lwe-ChatGPT, okuhlanganisa i-DoorDash, i-Spotify, i-Uber, nezinye ↗
Ukuhlanganiswa kohlelo lokusebenza olusha lwe-OpenAI kusondeza i-ChatGPT ekubeni umsizi owenza izinto, hhayi nje okhulumayo. Xhuma izinsizakalo ezifana ne-Spotify, i-Uber, ne-DoorDash, bese ucela uhlu lwadlalwayo, ukuhamba, noma ukudla - okulula ephepheni, okubaluleke kakhulu kunalokho okubonakala ekuqaleni.
Yilapho umjaho womsizi uba lukhuni khona. Umkhiqizo ophumelelayo ungase ungabi yimodeli ehlakaniphe kakhulu ekuchazeni, kodwa yilowo osebenza buthule emisebenzini yakho yansuku zonke, okuzwakala kuyinto evamile, futhi yingakho kubalulekile.
🛠️ 'Ayakhiwanga kahle okokuqala' - I-xAI kaMusk iqala kabusha ... futhi, futhi ↗
Kubikwa ukuthi i-xAI iqala kabusha umzamo wayo wokwakha ithuluzi lokubhala ikhodi le-AI, iletha ubuholi obusha obuvela ku-Cursor. Lokho akuyona into encane - kufana nokusetha kabusha ngemuva kokuba inguqulo yokuqala yehlulekile ukufika kahle.
Nokho, kusenesibonakaliso esinamandla esifihliwe ekusabalaleni. Ithalente elivela emikhiqizweni ephezulu yokubhala amakhodi e-AI eliya ema-lab anokufinyelela kwamamodeli kanye nokubala okukhulu likutshela ukuthi isikhungo samandla adonsela phansi sishintsha kuphi, noma ngabe isithombe sonke sisabonakala singazinzile.
🪖 Ibutho lase-US limemezela inkontileka no-Anturil ebiza amaRandi ayizigidi ezingama-20 ↗
Ibutho lase-US lisayine isivumelwano sesikhathi eside no-Anturil esingafinyelela inani elikhulu kakhulu, okuqinisa indlela ubuchwepheshe bokuzivikela obusebenzisa i-AI obungena ngayo kakhulu emphakathini. Hhayi into engavamile, hhayi eyokuhlola - izinga lokuthengwa kwezikhali.
U-Anturil uqhubeka nokuguqula ingxoxo yokuzivikela-i-AI isuke ekucabangeni kweqembu lokuziphatha ibe yisabelomali esiqondile kanye nezinhlelo zasensimini. Mhlawumbe kuyadabukisa, kodwa umkhondo wemali uqala ukubukeka njengomzila ovulekile.
🏭 UMusk uthi iphrojekthi ye-mega AI chip enhle kakhulu yeTesla izoqalwa ezinsukwini eziyisikhombisa ↗
U-Elon Musk uthe iphrojekthi enkulu yokukhiqiza ama-chip e-AI kaTesla isiseduze nokuqaliswa, okunye okubonisa ukuthi umjaho we-AI awusasewona nje amamodeli - umayelana nokuba ne-hardware stack ngaphansi kwawo. Ama-chip, izinto ezinhle, amandla, ukupholisa... izingxenye ezinhle manje seziyizimboni.
Iphethini ebanzi kunzima ukuyikhohlwa. Izinkampani zizama ukuvala ingqalasizinda ngaphambi kokuba kukhuphuke isidingo esilandelayo, ngoba ukuqasha ikusasa lakho komunye umuntu ngokuzumayo kubukeka sengathi kuyisivumelwano esibi.
Imibuzo Evame Ukubuzwa
Kungani izindleko ze-AI ezikhuphukayo ziholela ekuxoshweni kwabasebenzi ezinkampanini ezifana ne-Meta?
Lesi sihloko sibonisa ukuthi i-Meta ingase icabangele ukunciphisa okukhulu ukusiza ukuxhasa ngezimali ukwandisa ukusetshenziswa kwe-AI kanye ne-data center. Lokho kubonisa ukuhwebelana okubanzi okubonakalayo manje embonini yonke: izinkampani ziphatha ama-compute, ama-chip, kanye nengqalasizinda njengokubalulekile, kuyilapho zinciphisa ezinye izindleko ngamandla. Kulolo hlaka, inani labasebenzi liba ngelinye lamathuluzi asheshayo ukudonsa.
Zibonisani izingxabano ze-copyright ze-ByteDance ngokwakha amamodeli e-AI evidiyo?
Zibonisa ukuthi ikhwalithi yemodeli iyingxenye nje yenselele. Idatha yokuqeqesha ingaba yisithiyo esikhulu, ikakhulukazi kuvidiyo, lapho imibuzo ye-copyright kunzima ukuyishaya indiva futhi kulula ukuyikhulisa. Ngisho noma ukuqaliswa kubonakala kuseduze ngaphandle, ukungaqiniseki kwezomthetho kunganciphisa noma kumise inqubekela phambili ngokushesha okukhulu.
Ukuhlanganiswa kohlelo lokusebenza lwe-ChatGPT kukushintsha kanjani lokho umsizi we-AI angakwenza ngempela?
Ushintsho lusukela ekuphenduleni imibuzo kuya ekusizeni ukuqeda imisebenzi kuzo zonke izinsizakalo zansuku zonke. Esibonelweni lapha, ukuhlanganiswa nezinhlelo zokusebenza ezifana ne-Spotify, i-Uber, kanye ne-DoorDash kwenza umsizi abe usizo kakhulu ezinqumweni nasezenzweni ezivamile. Lokho kubalulekile ngoba ukulula, hhayi nje ubuhlakani bemodeli eluhlaza, kungabumba ukuthi yimuphi umsizi oba yingxenye yokuphila kwansuku zonke.
Kungani kubikwa ukuthi i-xAI iqala kabusha umzamo wayo wethuluzi lokubhala ikhodi ye-AI futhi?
Ukuqaliswa kabusha okubikiwe kusikisela ukuthi inguqulo yangaphambilini ayizange ifike ngokuqinile ngokwanele ukuze iqhubeke nokwakha njengoba injalo. Ukuletha ubuholi obuvela kumkhiqizo ofana ne-Cursor kukhomba ekusethweni kabusha okujulile kunokuba kube ukuthuthuka okuncane. Kuphinde kubonise ukuthi ithalente lokubhala ikhodi le-AI elinolwazi selibaluleke kangakanani, ikakhulukazi uma lihambisana nokufinyelela okuqondile kwemodeli kanye nezinsizakusebenza zekhompyutha ezisindayo.
Ithini inkontileka ka-Anturil mayelana nokuthi ukutshalwa kwezimali kwe-AI kuya kuphi?
Kubonisa ukuthi i-AI idlula izingxoxo zokubonisana kanye nocwaningo iye ekuthengeni isikhathi eside kanye nokuthunyelwa ensimini. Isilinganiso salokho kuphakamisa ukuthi abathengi bezokuvikela babona kakhulu izinhlelo ezisebenzisa i-AI njengamathuluzi okusebenza, hhayi ukubheja okuqagelayo. Lesi sihloko sichaza lokhu njengophawu lokuthi izabelomali manje sezilandela i-AI ezindaweni eziqondile, zezikhungo.
Kungani ama-chip fabs, izikhungo zedatha, kanye nehadiwe kuba yinto ebalulekile ekutshalweni kwezimali kwe-AI?
Ngoba umncintiswano awusagcini nje ngokuba ngubani onemodeli engcono kakhulu. Lesi sihloko sikhomba ekugxileni okukhulayo esitokisini esingokoqobo ngaphansi kwe-AI, okuhlanganisa ama-chips, ukwenziwa, ugesi, kanye nokupholisa. Izinkampani zibonakala zizama ukuvikela ingqalasizinda kusenesikhathi, njengoba ukuthembela ngokuphelele kumandla angaphandle kungaba kubi kakhulu, kube nomkhawulo, noma kube yingozi ngokwecebo.