Kwi-HackerNoon, siyawaxabisa amabali malunga nokwakha itekhnoloji ye-groundbreaking. Le mibuzo ayibhekiseli kumatshini we-AI kuphela-imalunga nokwenziwa kwezinto ezintsha, imiceli mngeni, kunye nobuchule bokuzisa izixhobo eziya kumgangatho ophezulu ebomini, ezilungele uluntu lwethu lweenkokeli zetekhnoloji, abakhi, kunye nabafundi abacinga ngekamva.
Igama lam ndinguAniruth . Ndisebenza kwiqela lokugcina kwi -Databricks , apho sisebenza ekuvumela ukuba abathengi bagcine inani elikhulu ledatha kwifomathi evulekileyo, enokunyuka kunye nePlatform yethu yeNgqondo yeDatha. Ngokukodwa, ndisebenza kwiinzame zethu zokusebenzisana kunye neDelta Lake kunye ne-Apache Iceberg.
I-Aniruth : I-Databricks idibanisa idatha kunye ne-AI ukunika abathengi ubukrelekrele obusebenzayo-into esiyibiza ngokuba bubukrelekrele bedatha. Oku kubandakanya ukungeniswa kwedatha enkulu, i-ETL, ukugcinwa kwezinto ezinkulu, imibuzo yobukrelekrele beshishini, kunye nomthwalo we-AI. Ubuchwephesha obusetyenzisiweyo ekufundeni koomatshini kwiminyaka elishumi edlulileyo bekukho ukususela kwiminyaka yoo-1980; ukunyuka kwedatha enkulu kwenza ukuba kube nokwenzeka ukuqhuba i-algorithms kwisikali.
Ubuchwephesha obufana nokuqaliswa kokudutyulwa okumbalwa okanye i-RAG ixhomekeke kwidatha ekumgangatho ophezulu. Iimodeli ezinedatha engcono ziphumelela ngokuchasene nezo zinezakhiwo ezingcono. I-Databricks ibeke utyalo-mali olubalulekileyo kwiinzame ezikhokelayo kwindawo yedatha, iphayona i-architecture ye-lakehouse kunye neefomathi zedatha evulekileyo kunye nolawulo oluvulekileyo, apho abathengi bayakwazi ukufumana ulwazi olungcono kunye nokusebenza kakuhle ukusuka kumachibi edatha.
Sisebenzisa iimodeli ze-AI ngeendlela ezininzi kwimveliso - njengeLlama 3 yoMncedisi we-AI. Sikholelwa kwidatha evulekileyo kunye ne-AI ecosystem kwaye sikhuthaza abathengi bethu ukuba basebenzise nayiphi na imodeli abayithandayo. Sinceda ukuqinisekisa ukuba abathengi banolawulo oluphela kwisiphelo somjikelo wobomi be-AI kungakhathaliseki iimodeli abazisebenzisayo, ngoko banokugxila ekwenzeni iimodeli zabo zibe yinjongo-eyakhelwe kwiimeko zabo zokusetyenziswa.
Sichithe umzamo omkhulu kunye notyalo-mali ekubekeni phambili ukuchaneka kunye neempendulo ezingakhethi cala zokusetyenziswa kwe-AI kwiimveliso zethu, kwaye siqhubeke nokuqhuba uvavanyo rhoqo.
Idatha kunye nesithuba se-AI sivela ngokukhawuleza, ngoko ke kubaluleke kakhulu ukugcina usesikhathini. Usuku lwam lunokubandakanya ukuthetha nabathengi, uhlalutyo lwemarike, ukudibanisa uxwebhu lweemfuno zemveliso, okanye ukulungiselela izixhobo zokuthengisa. Elona candelo ndilithandayo kukwenza imizobo ebonisa indlela izinto eziza kusebenza ngayo, njengoko kumnandi kakhulu ukuguqula umbono ube ngumbono.
Zininzi iimpumelelo ezinkulu ezizayo kungekudala. Enye into endinomdla kuyo kukugqithiswa komxholo. Kule minyaka ilishumi idlulileyo, iintengiso zilungiswe kakuhle kumlindi othile. Ezinye izinto zomxholo zilungisiwe, ezifana nokuba yeyiphi i-thumbnail Netflix ebonisa umsebenzisi, kodwa umxholo wokwenene (ividiyo ngokwayo) ayizange ibekho. Kuya kuba mnandi ukubona indlela abalawuli / abavelisi abalinganisa ngayo ibali abafunayo kunye nezinto ezinomdla zomsebenzisi.
Ndisebenza kugcino lwedatha enkulu, enokubhideka kakhulu ukuqonda. Sinolungelelwaniso olwahlukeneyo lwe-AI kwidatha, kodwa kuhlala kukho imibuzo malunga nokuba ezi zilungiso ziqhutywa nini, zisebenza njani, zingagqunywa ntoni, njalo njalo. Ngolu hlobo lwemibuzo, kubalulekile ukuqinisekisa ukuba sinomyalezo ocacileyo, ongaguqukiyo malunga. yintoni esiyakhayo nokuba kutheni. Ndifumene ukuba ukuchaza unobangela wokusikelwa umda kuhambelana kakuhle nabathengi.
Iimodeli ze-Multimodal ziya kuba ngcono kakhulu kwiminyaka ezayo, eya kutshintsha indlela yethu yokuqala yokusebenzisana ne-AI. Ukufumana imvakalelo yomntu kulula kakhulu ukusuka kulwazi olubonakalayo okanye oluvakalayo xa kuthelekiswa nesicatshulwa. Ndicinga ukuba kukho ithuba lokudala unxibelelwano lwendalo ngakumbi kuluhlu olubanzi lweemeko.
Ngokuqhelekileyo sifuna ukubona impendulo efanelekileyo kunye nokusetyenziswa. Ndithetha nabathengi rhoqo ukuze ndiqonde ukuba kutheni kwaye kutheni becinga malunga neemveliso zethu, eyona nto iphambili ukucacisa ukuba kutheni sibona iindlela ezithile kwiimethrikhi.
Iimveliso zedatha ezinkulu zidume ngokuba ngumngeni ekusebenziseni. Imizekelo elula ilula ukuseta, kodwa umthwalo wemveliso wemveliso ubandakanya ukubhidanisa uqwalaselo kunye nekhowudi. Ibiyeyona nto iphambili kum ukwakha ukusebenza okufunwa ngabathengi, ngelixa ndisenza imveliso ibelula kakhulu ukuyisebenzisa.
Ikamva kulapho naliphi na ishishini lifumana ulwazi kwiidatha zabo ngokulula. Kwihlabathi langoku, ulwazi lweshishini oluqhutywa yidatha luthintelwa kwiinkampani ezinkulu - kodwa banokukhetha amava alula.
Ngomntu ngamnye, ndonwabile kakhulu ngokudityaniswa kwe-AI kwihardware. Ukuza kuthi ga ngoku, siyibonile kakhulu i-AI kwizicelo zesoftware efana neewebhusayithi. Zininzi izicelo ezinkulu zezixhobo zokwakha ezisebenzisa i-AI, kwaye sele siqala ukubona ezinye zeziphumo kwiimoto kunye neefowuni.
I-databricks isendleleni eya ekubeni ilula kwaye ibe namandla ngakumbi ngaxeshanye. Zininzi iinzame esisebenza kuzo kwibhodi iphela, ukusuka ekwenzeni idatha enkulu kunye nokubala ngokulula ukusebenza kunye nokuphucula ukusebenza kwimibuzo kunye nokuhamba komsebenzi. Ngokwam, ndicinga ukuba sineempawu ezinomdla ezizayo kungekudala kuyo yonke imveliso eyenza ukuhamba komsebenzi kube lula nge-AI. Imizekelo ibandakanya izimvo ezenziwe nge-AI kwidatha, iingcebiso zekhowudi ye-AI kubahleli beencwadana, kunye ne-AI yokujongana nokuxoxa neenkcukacha (umzekelo, i-Databricks AI / BI Genie).
Kukho iinkxalabo malunga nokuba i-AI ingalinciphisa na inani lemisebenzi. Iimveliso zethu ziyilelwe ukwandisa ulwazi oluxabisekileyo, oluhlala ludibana nabasebenzisi. Ngokomzekelo, nge-AI / BI Genie, abasebenzisi banokwenza i-interfaces kwidatha yabo. La ngamava omlingo, apho abasebenzisi banokubuza imibuzo kwaye bafumane iimpendulo ezithe ngqo kubo. Ngapha koko, abasebenzisi banokujonga iSQL esetyenziswayo ukuqinisekisa ukuba yile nto bayifunayo. Oku kusebenzisana nabahlalutyi, ukunciphisa ixesha abalithathayo ukusuka kwingcamango ukuya kwingqiqo.
Enye into eyandothusayo yayikukuntsonkotha kwezinye iinkampani ezinkulu. Oku kwazisa iimfuno kwimveliso ngendingazange ndiyithathele ingqalelo ngokwam. Umzekelo oqhelekileyo kucinga ngamaqhinga okufuduka xa uzisa imveliso entsha. Ngokwesiqhelo, iinkampani ezinkulu ziya kuba zidityanisile itekhnoloji esele ikho (eqhelekile isoftware evulelekileyo) okanye yakhe isoftware yesiko ukusombulula ingxaki ejongana nemveliso yethu entsha. Kudla ngokuthatha ixeshana ukuqonda ukuba kutheni kwaye zidityaniswe njani na ukuze kuqinisekiswe ukuba sinesisombululo esiquka konke okunokwenzeka.
Ngaba ungathanda ukuphendula eminye yale mibuzo?