Emuva ngoJanuwari, ngazinika inselele yokwenza ibhizinisi elizenzakalelayo le-AI ngokugcwele ekupheleni kuka-2025.   , ngacabanga. Ngokuthembakala? Kube buhlungu. Iyiphi indlela engcono yokusheshisa nge-AI  Ngo-Agasti 2025, 'senginabenzeli' abangu-2 abangathemba, bakhe, futhi bakhiphe uhla lwemibhalo lwewebhu emizuzwini engu-30 ngokucela imvume yomuntu emi-3.  Nokho, muva nje, ngibe nalo muzwa ocwilayo ofinyelele i-prescendo ekhona.  Yaqala Kanjani  NgoJanuwari, ngaqala ngenjabulo ukuhlola i-AI tooling out, ngenza  nokuzilungiselela ukubhekana nale nselelo.    , ukubuyekezwa kwethuluzi  Ngangazi indlela yami yokuzungeza ama-AI API amakhulu futhi ngase ngibhale inala yekhodi ezenzakalelayo. Bengilindele izinhlaka 'ze-AI ejenti' ukuthi zivuthwe ngokufanelekile - ngangikholelwa i-hype.  Lesi kwakuyisambulo sokuqala:   I-'AI' iyi-hype emangalisayo futhi eyize.  Ukuhlola Okungokoqobo: Isitimela se-AI Hype  Kunamazwi amaningi amemezayo 'AI, thola i-AI yakho lapha' njengamanje, kunzima ukuzizwa ucabanga.  Kuwo wonke amathuluzi e-AI engiwahlolayo, okuyiwona, amahle ngempela, ama-4 noma ama-5 noma ayi-10 awelela kumaki. Abaningi abasebenzi.   Lokhu kufanele kuzwakale njengokuthungatha igolide kwezamabhizinisi, kodwa kwesinye isikhathi, ngiyazibuza ukuthi ngabe kufana yini neMidlalo ye-squid.  Kuhamba kanjani: Amaphupho kanye ne-Duct Tape   Isikhathi sokwakha umkhombandlela esimaphakathi: 27-36 min; ukuthintwa kwabantu: 3; izinga lokwehluleka ekufakweni kuhlu kokuqala: ~ 40%  Ngiqale ngemodeli yebhizinisi elula, engicabange ukuthi akufanele ibe nzima kangako ukuzenzela nge-AI: Imibhalo yewebhu. Ngacabanga ukuthi uma uhla lwemibhalo lungasebenzi, ngingaphinda ngisebenzise okuningi engikufundayo futhi ngikusebenzise kunoma iyiphi enye inkampani ephambi kwewebhu.        Umhlabathi we-SEO ngaphansi kwami uzwakala udwetshiwe. Ngisanda kuqala umsebenzi wokumaketha okuzenzakalelayo.  Kodwa sesiku-August .  Ngiyazitshela ukuthi, '   ' futhi khumbula ukuthi ngikwenze okuningi ukuze ngikufunde, nokwabelana, njengokwenza uhla lwemibhalo lwe-AI oluqhakazayo. uthathe le nselele ewumsangano  Kodwa ukusekela leyo nselele ewuhlanya kunzima ngezinye izikhathi.  Amabili “ama-ejenti” ami okuqala ayasebenza. Bangakwazi ukuthola i-niche ngocwaningo lwegama elingukhiye, bangabhalisa isizinda, baguqule amaseva egama, bakhiqize ama-logo, iphalethi yombala, futhi empeleni bakhe isiza sombhalo esidlulayo.   Lokho kuzwakala kukuhle, futhi ngicabanga ukuthi empeleni kunjalo.  Kodwa akunjalo   . uzizwe njenge-AI  Eqinisweni, noma ngisebenzisa i-AI (LLMs) kulo lonke, eqinisweni, lawa ukuhamba komsebenzi okuzenzakalelayo nama-micro-ejenti. Basafundeka njengokuthi "inqwaba yamasu ahlanganisiwe ahlanganiswe ndawonye, asebenzisa njalo ama-LLM ukugcwalisa izikhala."   Ekuqaleni kwale nselelo, ngiphuphe ama-agent e-AI azimele; yokugxila kwifilosofi ejulile yebhizinisi. Ukuhlanganisa ndawonye amathuluzi esimanje kanye nemiyalelo ecijwa kancane yamandla amakhulu.  Akukho lapho okukhona.  I-AI Entrepreneurship: Ithuba vs Ukungaqiniseki  Njengamanje, izingcezu nezingcezu zephazili ziyasebenza:    Amamodeli amasha anamandla kakhulu. Lokhu kusho ukuthi singakwazi ukusebenzisa i-vibecode amathuluzi angaphakathi nezimpahla ezincane zokumaketha. Ama-LLM aba ngcono ngaso sonke isikhathi.    Singakwazi ukuzenzela izinqubo eziningi zebhizinisi futhi sithuthukise ama-ops ngama-LLM. Izinkundla ezizenzakalelayo ziba namandla ngokwengeziwe.    -  (isimo sezulu sika-Kokotajlo et al.) siyaqhubeka sizwakala siyiqiniso ngokuthi ama-agent okwamanje 'awathembeki', kodwa sesingakwazi kakade ukuwasebenzisa ukwenza imisebenzi elula. Ama-AI Agents aphezu kwesiphetho sokuba 'ngokoqobo'   I-AI 2027  Njengosomabhizinisi besoftware, singakha amathuluzi ngokushesha kunangaphambili. Lokhu kuqinisa kufinyeleleka kakhulu kuwo wonke umuntu onoxhumo lwe-net kanye nemali eyanele yokukhokhela ukufinyelela (okushibhile uma kuqhathaniswa).   …kodwa akukhona konke okuhle.   : Kunezinkinga ezingaxazululeki ngezamabhizinisi ngo-2025    , kodwa ngokufanayo, uma sitshala   , sizifaka engcupheni yokumosha isikhathi esibucayi nezisetshenziswa njengoba amathuluzi amasha enza umsebenzi wethu ungafuneki. Uma singatshali ku-AI, sizifaka engcupheni yokusalela emuva ngendlela engafanele    ngamaphimbo amaningi kangaka, kuba nzima ngokuphindwe ka-10 ukwazi ukuthi kufundwa kuphi/kuyini. Njengoba ama-LLM eba ngcono ekuncengeni, kuzoba nezingibe eziningi nezingibe okufanele zigwenywe phakathi kwe-hype kanye nokuthunyelwe kwenqubekelaphambili. Sidabula izindaba ze-AI nsuku zonke;    - kodwa awakabi khona - futhi ngenxa yokuthi akekho owaziyo ukuthi azoguquka nini aze afike ezingeni lokushintsha kakhulu ibhizinisi, ukungaqiniseki kubusa.  Ama-AI Agents aphezu kwamathuba okuba 'angempela'  Ingabe lesi sigqi esiphindaphindayo ekungabazeni kwami osomabhizinisi abancane noma ukukhathazeka okunengqondo? Kunzima ukungangabazi ukuthi sakhani phakathi kwakho konke lokhu. Ngizithola ngibuza:  Osomabhizinisi besoftware balengeza kuphi inani emhlabeni we-AI-dev?   Kufanele siyichithe kuphi i-delta v phakathi kwezinxushunxushu ezinjena?  Njengamanje, ayikho impendulo ecacile.  Kuvele Kuthi Sisesekuseni  Angikalilahli iphupho lami, kodwa ukugibela kuhle kakhulu.  Njengoba ungabona kumkhosi wezindaba we-AI onamagwebu, izinto zithuthuka ngokushesha. Kancane kancane, ngicabanga ukuthi sizofika lapho.  Kodwa ngiyaqinisa ukuthi akekho kithi ongacabanga ngempela ukuthi 'lapho' kuzobukeka kanjani. Ikusasa elinzima ukulibikezela.   Izowela kuphi imingcele? Ingabe ama-super agents angazilinganisa izinkundla zokudlala? Ingabe ukumaketha kuba umzabalazo wamandla wekhompuyutha njengempi ye-AIs yokwenza ama-algorithms angcono nangcono ukuze uvikele ukunakwa?  Ngabe usuku empilweni yosomabhizinisi lubukeka kanjani ngo-2026?  Noma ngiseduze kakhulu naleli gagasi, ngiyazabalaza ukubona ukuthi konke lokhu kuhamba kanjani.  Ngenze inqubekelaphambili ethile ngale nselele, kodwa angikwazi ukuzibamba kodwa nginomuzwa wokuthi ngidinga   . ukuqina okwengeziwe  Landela i-Build  Uma ufuna ukufunda izingodo zokwakha zamasonto onke ze-AI, izincazelo ezizenzakalelayo, nokuhlolwa kwebhizinisi le-AI esophayo, landela lapho ngizama ukwenza ibhizinisi le-AI elizwakala njenge-AI ngempela:   . bhalisa lapha

The is an opinion piece based on the author’s POV and does not necessarily reflect the views of HackerNoon.

100% AI Business Challenge Ride-along

Woody Hayday.com

Niche Report.io

Read My Stories

Lo msindo ukhiqizwa ngolimi lokuqala lwendaba!

Ukwakha Ibhizinisi le-'AI' ngo-2025: I-Duct Tape, Hhayi i-AGI

About Author

IMIBONO

HANG TAGS

LESI SIHLOKO SETHULWE NGAPHAKATHI

Related Stories

The Billionaires F*cked Up

Zulu Network: Moving The Bitcoin Economy Forward With a Two-Tiered Bitcoin Layer 2 Architecture

The Billionaires F*cked Up

Zulu Network: Moving The Bitcoin Economy Forward With a Two-Tiered Bitcoin Layer 2 Architecture

Light-Mode

Classic

Newspaper

Dark-Mode

Neon Noir

Minty

HN StartUps