I-Anthropic Isezingxoxweni Zokusebenzisa Ama-chip e-AI e-Microsoft ↗
Kubikwa ukuthi i-Anthropic ihlola isivumelwano sokuqasha amaseva asebenzisa ama-chip e-Maia AI e-Microsoft. Inkulumo ilula: ukubala okwengeziwe kukaClaude, ngaphandle kokuthembela kakhulu kwingqalasizinda ye-Nvidia kuphela.
Kusesekuseni kakhulu, futhi kungase kungafiki ndawo. Kodwa isignali inomsindo noma kunjalo - amalabhu e-AI asemngceleni aseba ochwepheshe bezokwakha ingqalasizinda abaphelele. Amamodeli ayingxenye ekhazimulayo, ama-chips ayizinto zokufaka amanzi okungekho muntu ongazenaki.
Isizindalwazi se-D&B samabhizinisi ayizigidi ezingu-642 sakhiwe ngabantu, hhayi ama-AI agents. Ngakho baphinde basakha kabusha. ↗
I-Dun & Bradstreet yakha kabusha izinhlelo zayo zedatha yezentengiselwano ukuze ama-AI agents akwazi ukuzisebenzisa kahle. Abahlaziyi babantu bangabekezelela ukungacaci kanye nokusesha okuhamba kancane. Ama-agents, hhayi kangako: adinga umongo ohlanzekile kanye nokuqiniseka okusheshayo.
Lokho kuzwakala kuncane, kodwa kuyinkomba enkulu yebhizinisi. Kwakhelwe idatha eningi yebhizinisi abantu abachofoza kumadeshibhodi - hhayi abasebenzi besofthiwe abenza izinqumo ngesivinini somshini.
Ama-ejenti akho e-AI adinga i-terminal, hhayi nje i-vector database ↗
Abacwaningi bacindezela umqondo ohlukile wokuthola ama-ejenti: vumela ama-ejenti e-AI aseshe amafayela angahluziwe kanye nama-corpora ngqo ngamathuluzi omugqa wemiyalo, esikhundleni sokuthembela kuphela kudathabheyisi ye-vector kanye neziqeshana eziqoqiwe.
Kungushintsho oluncane lobuchwepheshe olunencazelo enkulu. Mhlawumbe ama-ejenti awadingi nje inkumbulo engcono - mhlawumbe adinga izandla ezingcono. I-terminal ayiyona into ekhangayo, kodwa futhi ayiyona i-spanner, futhi lokho kulungisa izinto.
🔐 Izitifiketi ezivumelekile, ama-akhawunti abiwe: indlela abahlaseli abaphule ngayo isignali yokugcina yokwethembana ka-npm ↗
Kubikwa ukuthi igagasi lamaphakheji e-npm anonya lidlule ekuhlolweni kwemvelaphi ngoba abahlaseli basebenzise iziqinisekiso zomgcini ezithintekile. Izitifiketi zazibonakala zisebenza, okuyiyona nkinga - ilokhi yayicwebezela, kodwa okhiye babebiwe.
Ukukhathazeka okukhulu kuvela ngqo emisebenzini yokubhala ikhodi ye-AI. Imishini yonjiniyela, ama-ejenti okubhala ikhodi, abaphathi bamaphakheji, amapayipi e-CI - konke kuba yindawo eyodwa yokuhlasela exakile. Kulula, yebo. Kukhululekile, hhayi ngisho nakancane.
Sethula i-OpenAI yaseSingapore ↗
I-OpenAI iyanda eSingapore ngomkhankaso omusha we-AI osetshenzisiwe, okuhlanganisa ne-Applied AI Lab yayo yokuqala ngaphandle kwe-US. Ukugxila kusetshenziswa ngendlela engokoqobo, ithalente lendawo, kanye nokufinyelela emkhakheni womphakathi.
Lokhu akulona nje elinye ithuba lokuvulwa kwamahhovisi. Kubeka iSingapore ngokujulile ebalazweni lengqalasizinda ye-AI, futhi kunikeza i-OpenAI enye indawo e-Asia-Pacific. Inkulu kancane, njengefriji ekhala ekamelweni elilandelayo.
Imibuzo Evame Ukubuzwa
Kusho ukuthini ukusebenzisa ama-chip e-AI e-Microsoft i-Anthropic?
Kubikwa ukuthi i-Anthropic ihlola ukusetshenziswa kwama-chip e-Maia AI e-Microsoft ukuqasha amandla okusebenzisa i-Claude. Inhloso eyinhloko ukunciphisa ukuthembela kwingqalasizinda ye-Nvidia kuphela ngenkathi kuqinisekiswa amandla amaningi okucubungula. Ngisho noma isivumelwano singaqhubeki, kukhombisa ukuthi ama-lab e-AI asemngceleni aseqala ukucabanga njengezinkampani zengqalasizinda, hhayi nje abakhi bemodeli.
Kungani ama-ejenti e-AI eshintsha indlela izinhlelo zedatha yebhizinisi ezakhiwa ngayo?
Ama-ejenti e-AI adinga ukufinyelela kolwazi okuhlanzekile, okusheshayo, nokuhlelekile kunalokho okunikezwa yizingosi eziningi zedatha ezibhekene nabantu. Ukwakhiwa kabusha kwe-Dun & Bradstreet kugcizelela ukuthi izinhlelo zedatha zebhizinisi ezakhelwe amadeshibhodi zingase zingawafanele ama-ejenti esofthiwe enza izinqumo ngesivinini somshini. Emigwaqweni eminingi yamabhizinisi, ukungacaci abantu abangakubekezelela kuba yinkinga yokwenza izinto ngokuzenzakalela.
Ingabe ama-ejenti e-AI adinga okungaphezu kwesizindalwazi se-vector?
Yebo, emisebenzini eminingi yokusebenza, ama-ejenti e-AI angadinga amathuluzi ngale kwemininingwane ye-vector. Lesi sihloko siphawula intshisekelo ekhulayo yokuvumela ama-ejenti ukuthi aseshe amafayela angahluziwe kanye nama-corpora ngqo ngamathuluzi omugqa wemiyalo. Lokhu kusikisela ukuthi ama-ejenti angazuza "ezandleni" ezisebenzayo, hhayi nje inkumbulo engcono, ikakhulukazi lapho edinga ukuhlola, ukuthola, noma ukuphatha ulwazi kuzo zonke izinhlelo eziyinkimbinkimbi.
Ama-akhawunti omgcini wezimpahla ezibiwe angakuthinta kanjani ukuphepha kwephakheji ye-npm?
Iziqinisekiso zomgcini ezibiwe zingenza amaphakheji anonya abonakale ethembekile lapho abahlaseli besebenzisa ama-akhawunti nezitifiketi ezivumelekile. Esibonelweni se-npm, ukuhlolwa kwemvelaphi kubikwa ukuthi kudlulile ngoba isignali yobunikazi ibonakala isemthethweni. Ingozi inkulu kakhulu ekusebenzeni kwekhodi ye-AI, lapho imishini yonjiniyela, abaphathi bephakheji, ama-ejenti okubhala ikhodi, kanye nezinhlelo ze-CI kungaba yindawo eyodwa yokuhlasela exhunyiwe.
Kungani ukunwetshwa kwe-OpenAI eSingapore kubalulekile ekusetshenzisweni kwe-AI?
Ukwandiswa kwe-OpenAI eSingapore kufaka phakathi i-Applied AI Lab yayo yokuqala ngaphandle kwe-US, ngokugxila ekusetshenzisweni okusebenzayo, ithalente lendawo, kanye nokufinyelela emkhakheni womphakathi. Lokhu kunikeza iSingapore indima enkulu kumephu yengqalasizinda ye-AI yase-Asia-Pacific. Kuphinde kubonise ukuthi izinkampani ezinkulu ze-AI zanda kanjani ngale kokuthuthukiswa kwemodeli ziye ekusetshenzisweni kwesifunda kanye namacala okusetshenziswa okusetshenzisiwe.