🤖 I-Apptronik, inkampani entsha ye-Humanoid, iqoqa u-$520 million ngokusekelwa yi-Google kanye ne-Mercedes-Benz ↗
I-Apptronik isanda kuthola umklomelo omkhulu we-$520M, i-Google ne-Mercedes-Benz zihlangene - kanye nenqwaba yabanye abatshalizimali abakhulu. Intengo ecatshangelwayo iyanda... inkulu. Njengokuthi, "ngokungathi sína" inkulu.
Iphimbo lilula: khulisa i-Apollo, irobhothi labo elifana nomuntu, bese ulifaka ezimbonini nasezindaweni zokugcina impahla lapho umsebenzi oyisicefe nophindaphindayo ungapheli khona. Kukhona futhi i-engeli ejulile lapha - ukuhlanganiswa okuqinile ne-Google DeepMind ukuze i-Apollo ikwazi ukusebenza ngobuhlakani kumamodeli asekelwe ku-Gemini, okuwumzila osheshayo "wokwenza irobhothi lingabi yisiwula".
🧠 Ikomiti ephezulu yeDemocrat kwiNdlu yase-US eChina ivulelekile ekuthengisweni kweNvidia H200 ↗
URo Khanna ukhombisa ukuthi akamelene ngokuzenzakalelayo nokuthengisa ama-chip amadala e-H200 eNvidia eShayina - okuwushintsho oluphawulekayo ezwini uma kuqhathaniswa nesimo esinzima ikomidi elibe naso. Ingqikithi: uma i-US ihlala phambili ngezizukulwane ezimbalwa, ama-chip amadala angase abekezeleleke ukusetshenziswa okungeyona eyezempi… noma kubonakala kanjalo.
Kodwa udweba umugqa ocacile ngezinto ezintsha - izakhiwo ezintsha akufanele zibe semenyu. Okwamanje, ubuholi bekomidi ngakolunye uhlangothi busaphakamisa ifulegi "lokuhlangana kwezempi zomphakathi", ngokuyisisekelo bephikisana ngokuthi "okungeyona eyezempi" kuyindaba enenhlonipho yokulala.
🏗️ Imithombo ithi i-Modal Labs, inkampani eqalayo ye-AI, izoqala ingxoxo yokuqoqa imali engu-$2.5B ↗
Kubikwa ukuthi iModal Labs isezingxoxweni zokuqoqa imali cishe ngenani lika-$2.5B, okukutshela ukuthi "ukuqonda" kusashisa kangakanani. Ukuqeqeshwa kuthola ubuhle, kodwa ukuqonda yingxenye oyikhokhela unomphela, njengokubhalisela ongeke ukwazi ukukukhansela.
Uma lokhu kufika, kuyisibonakaliso sokuthi abatshalizimali baphatha ukulethwa kwekhompyutha kanye nemodeli njengokujaha okusha kwegolide. Kufana nokungathekisana okuncane, kodwa kuyafaneleka - wonke umuntu ufuna ukuthengisa amapiki namafosholo, noma ngabe abavukuzi bangama-algorithms.
📺 Ukwamukelwa kwezikhangiso ze-AI "kubi kakhulu" njengoba i-brand beefs .. ↗
Kuvela ukuthi abantu abakhathali ngempela ngenxa yokukhangisa okune-AI - okunye ukusabela kuchazwa ngokuthi "kubi kakhulu," futhi ngiyaqonda. Kukhona ukukhohlisa okuthile okumangalisayo okwenza ubuchopho bakho buthi "cha," ngisho noma izithombe zihlaba umxhwele.
Futhi, idrama yomkhiqizo we-AI iya ngokuya ihlekisa obala: I-Anthropic yaqhuba umkhankaso kaClaude othatha izinguquko ekukhangiseni kwepulatifomu ye-AI, futhi umbhalo ongezansi uzwakala… ungacacile. Kufana nokulwa komcamelo wenkampani, ngaphandle kokuthi imicamelo yenziwe ngemali yokuqalisa ibhizinisi.
⚖️ Umthetho Ocacile Uzosungula Izidingo Zesaziso Semisebenzi Ene-Copyright kuDatha Yokuqeqeshwa kwe-AI ↗
Umthethosivivinywa ohlongozwayo wase-US obizwa ngokuthi i-CLEAR Act ucindezela izidingo zesaziso mayelana nemisebenzi enelungelo lobunikazi esetshenziswa kumasethi edatha okuqeqeshwa kwe-AI - ngokuyisisekelo "tshela abantu ukuthi usebenzise ini," okungenani ngendlela ehlelekile nehlelekile. Akuyona "ukukuvimbela" kodwa "kufakazisa ukuthi awuzenzisi."
Uma iqhubekela phambili, izokwengeza olunye ungqimba lokuthobela imithetho ezinkampanini ezikhipha i-AI ekhiqizayo kwezentengiselwano. Okungaba kuhle! Noma ingaba yisisombululo samaphepha, kuye ngokuthi imithetho ibhalwa kanjani futhi iphoqelelwe kanjani… ukubuyela emuva okuncane, kodwa leyo ngxenye ibalulekile.
💾 Isitoko seNvidia Siyakhuphuka. I-Broadcom Ivala Igebe Lama-AI Chips, Kusho Umhlaziyi. ↗
Abahlaziyi baphikisana ngokuthi i-Broadcom isondela ku-Nvidia kuma-AI chips, ikakhulukazi nge-silicon eyenziwe ngokwezifiso - ikakhulukazi umsebenzi ohlobene nama-TPU e-Google. Lapho izinkampani eziningi zifuna ukuqagela okushibhile, kulapho ihadiwe "enhle ngokwanele futhi engabizi" iqala ukubonakala iyingozi enkosini ephezulu.
Lokhu akusho ukuthi iNvidia isiphekiwe ngokuzumayo - ingasondelene kakhulu - kodwa kusikisela ukuthi imakethe ihlukaniswe yaba imizila emibili: izilo zokuqeqesha ezisezingeni eliphezulu, kanye namahhashi asebenzayo okubikezela angcono kakhulu. Umdlalo ofanayo, izigaba ezahlukene zesisindo.
Imibuzo Evame Ukubuzwa
Kwenzekeni kulesi sifinyezo sezindaba zobuchwepheshe be-AI?
Lokhu kubuyekezwa kulandelela ukunyakaza okukhulu kuwo wonke amarobhothi, ama-chip, i-compute, izikhangiso, kanye nenqubomgomo ye-copyright. I-Apptronik iqoqe imali eningi yokukhulisa irobhothi layo elifana ne-Apollo lomsebenzi wefektri kanye nendawo yokugcina impahla. Kanye nalokho kukhona izimpawu ezishintshayo ekuthunyelweni kwama-chip e-Nvidia e-China, ukwanda kwenani eliphezulu okubikiwe kwe-inference startup Modal Labs, ukugxekwa kwabathengi ngezikhangiso ezisindayo ze-AI, uMthetho we-CLEAR ohlongozwayo ogxile esazisweni sedatha yokuqeqesha, kanye nezinkomba zokuthi i-Broadcom inciphisa izingxenye zegebe lama-chip e-AI.
Kungani i-Apptronik iqoqa imali engaka ngerobhothi layo elifana ne-Apollo?
Inhloso eshiwoyo ukukhulisa i-Apollo nokuyibeka emafektri nasezindaweni zokugcina impahla lapho imisebenzi ephindaphindwayo ivame khona. Lo mbuthano uhlanganisa abaxhasi abakhulu, okubonisa ukuzethemba kwabatshalizimali ezenzakalelayo ezibonakalayo njengesigaba esikude. Indaba iphinde igcizelele i-engeli ejulile: ukuhlanganiswa okuqinile ne-Google DeepMind ukuze i-Apollo ikwazi ukugijima ngokuhlakanipha kakhulu kumamodeli asekelwe ku-Gemini, iqhubekisele phambili ikhono kunokwandisa inani lamayunithi akhiqizwayo kuphela.
Amamodeli asekelwe ku-Google DeepMind kanye ne-Gemini angayenza kanjani i-Apollo ibe “hlakaniphile”?
Isisekelo siwukuthi amamodeli e-AI aqinile angathuthukisa indlela irobhothi elinobuntu elihumusha ngayo imiyalelo, lihlela izenzo, futhi lizivumelanise nezimo ezididayo phansi. Esikhundleni sokuziphatha okubuthakathaka, okubhalwe kusengaphambili, irobhothi lingaba nezimo eziguquguqukayo ngolimi nokuqonda. Ezindleleni eziningi, lokho kuholela ekusetshenzisweni okungcono kwemisebenzi kanye nokuphindaphinda okusheshayo emisebenzini emisha, ikakhulukazi lapho kuhanjwa phakathi kwezindawo ezahlukene zefektri noma zesitolo.
Ingabe abenzi bomthetho base-US bavulelekile ngempela ekuthengisweni kwama-chip e-Nvidia H200 eShayina?
Esinye isibonakaliso esiphawulekayo ukuthi uMmeli uRo Khanna uphakamise ukuthi akaphikisani ngokuzenzakalelayo nokuthengisa ama-chip amadala e-Nvidia H200 eShayina, kuye ngokuthi yiziphi izindlela zokuvikela kanye nokusetshenziswa kwawo. Uhlaka luwukuthi ukuhlala "ezizukulwaneni ezimbalwa ezizayo" kungenza ama-chip amadala angazweli kakhulu ekusetshenzisweni okungeyona eyezempi. Udweba umugqa ocacile ngezakhiwo ezintsha, kuyilapho amanye amazwi ekomidi esagcizelela izingozi zokuhlanganiswa kwezempi phakathi kwabantu bomphakathi.
Kungani izinkampani eziqalayo ze-AI ezifana neModal Labs zithola amanani aphezulu kangaka?
Isiphetho "yizindleko zaphakade" ze-AI: uma amamodeli esesetshenzisiwe, ukuwakhonza ngezinga eliphezulu kuba yizindleko eziqhubekayo. Lokho kwenza ukulethwa kokubala, ukukhonza amamodeli, kanye nokwenza ngcono kuzwakale njengesendlalelo sebhizinisi esiqinile - esiseduze nengqalasizinda kunokwakhiwa kanye. I-Modal Labs ebikwe ezingxoxweni eziphathelene nokuhlolwa kwentengo engu-$2.5B kubonisa isifiso sabatshalizimali sezinkampani zokukhetha nokufaka amafosholo ezisiza abanye ukuthi basebenzise i-AI ngokwethembeka nangokushibhile.
Sisho ukuthini isiphakamiso se-CLEAR Act ngemisebenzi enelungelo lobunikazi kudatha yokuqeqeshwa kwe-AI?
Isiphakamiso esichazwe lapha sigxile ezidingweni zesaziso - "tshela abantu ukuthi usebenzise ini" kunoku "vimbela ukuqeqeshwa." Uma sithuthuka, singangeza izinyathelo zokuthobela imithetho yezinkampani ezithengisa i-AI ekhiqizayo, okungenzeka kudinge ukudalulwa okuhlelekile mayelana nokufakwayo okune-copyright. Ukuthi lokho kuba sobala noma "isuphu yamaphepha" kuncike ekutheni imithetho itholakala kanjani, ukuthi imithwalo yabelwe kanjani, nokuthi ukuphoqelelwa kubukeka kanjani ekusebenzeni.