Kwenzekeni
I-Anthropic ihlola ishejuli yokukhishwa esheshisiwe yemodeli entsha ye-AI ukuze iqhudelane ne-OpenAI esanda kwethulwa i-GPT-6 Astra. Ngokwemibiko ecashunwe yi-SBS, lolu shintsho luza njengoba i-Anthropic ilungiselela umnikelo wokuqala womphakathi (IPO) manje ohloselwe uNovemba. Lesi sinyathelo silandela idatha evela ezisekelweni ezifana ne-Ramp kanye ne-OpenRouter ebonisa ukuthi imodeli yakamuva ye-OpenAI isizuze kakhulu ekusetshenzisweni kwemali kwebhizinisi kanye nethrafikhi yonjiniyela, okungase kube inselele isikhundla semakethe se-Anthropic.
I-Anthropic kubikwa ukuthi icubungula ukukhishwa okusheshisiwe kwemodeli entsha ye-AI ukuze kubhekwane nosongo lokuncintisana olubangelwe yi-OpenAI's GPT-6 Astra. Lesi sinqumo silandela imibiko yokuthi i-Astra ibone ukutholwa ngokushesha ezindaweni zamabhizinisi, ngedatha evela kunkundla yokuphatha ukuchitha imali i-Ramp ebonisa i-Astra ithatha cishe u-13% wezindleko zebhizinisi ze-AI, uma kuqhathaniswa no-8% we-Anthropic's Claude Feble.
Isimo sokuncintisana sesishintshile nasezinkundleni ezigxile kunjiniyela. I-OpenRouter ibike ukuthi ukusetshenziswa kwemali kumamodeli e-OpenAI kudlule i-Anthropic okokuqala ngqa eminyakeni emibili, okubonisa ukuguguleka okungenzeka kokuma kwe-Anthropic phakathi konjiniyela. Lezi zinguquko zenzeka njengoba i-Anthropic ilungiselela i-IPO, ephushelwe kuNovemba.
Lo mbiko uphawula ukuthi ubuholi be-Anthropic, okuhlanganisa ne-CEO u-Dario Amodei, phambilini bebekhuthaza ukuthi kube nesivinini esihamba kancane, esigxile kwezokuphepha. Ukusheshisa okungenzeka kokukhishwa kwemodeli entsha kubukwa izibukeli zemboni njengovivinyo olubucayi lokuthi inkampani ingakwazi yini ukugcina isithunzi sayo sokuqala sokuphepha ngenkathi iphendula izingcindezi zemakethe.
Imininingwane yomthombo: news.sbs.co.kr ↗
Kungani kubalulekile
Ukusheshisa okungaba khona kwemephu yomgwaqo yomkhiqizo we-Anthropic kugqamisa ingcindezi enkulu yokuncintisana emkhakheni we-AI webhizinisi, lapho ubuholi bemakethe bukhula bushintshashintsha. Lokhu kuthuthukiswa kubaluleke kakhulu ngoba kuhlola ukuzibophezela kwenkampani okushiwo ekuthuthukisweni kwejubane 'lokuphepha kuqala', obekuyinsika ewumgogodla wobunikazi bayo. Ukulinganisa isidingo sokusungula izinto ezintsha ngokushesha ukuze kwanelise abatshalizimali bangaphambi kwe-IPO ngokumelene nezingozi zokuthunyelwa ngaphambi kwesikhathi kusewukushuba kwesimo esibucayi enkampanini. Ngaphezu kwalokho, ukukhuphuka kwamamodeli wemithombo evulekile nesisindo esivulekile kwengeza ungqimba lobunkimbinkimbi, njengoba lezi zindlela ezihlukile zisongela ukwenza ingqalasizinda ye-AI futhi yehlise amandla entengo yabathuthukisi abakhulu bamamodeli okuphathelene afana ne-Anthropic kanye ne-OpenAI.
Umbango phakathi kwe-Anthropic ne-OpenAI ukhombisa isimo semakethe ye-AI yamanje, lapho izinzuzo zobuchwepheshe zivame ukuhlala isikhashana. Abatshalizimali baqaphe ngeso lokhozi ukuthi i-Anthropic ingakwazi yini ukuvikela isabelo sayo semakethe njengoba umjikelezo wonjiniyela abakhulu ngokusebenzisa izizukulwane zamamodeli amasha.
Ngale kokuncintisana okuqondile ne-OpenAI, ukuvela kwamamodeli omthombo ovulekile kanye nesisindo esivulekile kuletha inselele yesakhiwo. Lawa mamodeli angase ehlise izindleko zamathokheni futhi avumele amabhizinisi ukuthi akhe ingqalasizinda yobunikazi, okungenzeka acekele phansi amamodeli ebhizinisi ezinkampani ezithembele ekufinyeleleni kwemodeli ebiza kakhulu, yokuphathelene.
Ingcindezi yokubonisa inzuzo ngaphambi kwe-IPO iphoqa isenzo esinzima sokulinganisa se-Anthropic. Inkampani kufanele ilinganise uhlobo lokuqeqeshwa oludinga imali enkulu futhi ikhiphe amamodeli amasha ngokumelene nesidingo sokukhombisa abatshalizimali indlela ecacile yembuyiselo yezezimali endaweni yokwenyuka kwezinga lenzalo.
I-Interactive Mechanism: Indlela Esebenza Ngayo Ngempela
Hlola ubuchwepheshe obuyisisekelo ngemuva kwalokhu kuthuthukiswa ngokuhlanganyela.
Which component of an AI application is the machine-learning model itself?
Ongakubuka ngokulandelayo
Izibukeli kufanele ziqaphe ukuthi i-Anthropic iqinisekisa ngokusemthethweni idethi yokukhishwa kwemodeli entsha nokuthi inkampani ihlanganisa kanjani lokhu kusheshisa nokugcizelela kwayo umphakathi kwangaphambilini ekunciphiseni ukuthuthukiswa kwe-AI ngenxa yezizathu zokuphepha. Ukwengeza, umthelela waleli su ekulinganisweni kwe-IPO kwe-Anthropic kaNovemba kuzoba inkomba ebalulekile yokuthembela kwabatshalizimali emandleni enkampani okugcina umkhawulo wokuncintisana uma iqhathaniswa ne-Anthropic ye-GPT-6 Astra. Ukuthambekela okubanzi kwezinguquko zokusebenzisa imali kwebhizinisi phakathi kwalaba bahlinzeki ababili kuzosebenza futhi njengesilinganiso sokusimama kokuguquguquka kwemakethe ye-AI yamanje.
Buka izimemezelo ezisemthethweni mayelana nomugqa wesikhathi wokukhishwa kwemodeli elandelayo ye-Anthropic nanoma imaphi amadokhumenti okuphepha angase abhekane nesimo sangaphambilini senkampani mayelana nesivinini sokuthuthukiswa.
Qapha inqubo ye-IPO yangoNovemba ukuze uthole izibuyekezo zokuthi ukuncintisana kwezimakethe nokusebenza kwamamodeli amanje kuthonya kanjani ukulinganisa kwenkampani kanye nomuzwa wabatshalizimali.
Landelela idatha yezindleko zebhizinisi namamethrikhi eplathifomu kanjiniyela ukuze ubone ukuthi ithrendi yamanje evuna i-OpenAI's GPT-6 Astra isaqhubeka noma uma ukukhishwa okusha okungaba khona kwe-Anthropic kuphinda kubambe kabusha isabelo semakethe ngempumelelo.