Ingabe i-AI izothatha indawo yabahleli? Okokugcina, vala isihleli sekhodi.

“ Okokugcina, vala umhleli wekhodi. ” Lesi sisho solimi esihlathini besilokhu sijikeleza ezinkundleni zonjiniyela, sikhombisa amahlaya akhathazekile ngokukhuphuka kwabasizi bekhodi be-AI. Njengoba amamodeli e-AI ekhula ekwazi ukubhala amakhodi, abahleli bohlelo abaningi bayabuza ukuthi ingabe abathuthukisi babantu baphokophele esiphethweni esifanayo nesama-elevator opharetha noma ama-switchboard opharetha - imisebenzi eyenziwa yisidala ngokuzenzakalelayo. Ngo-2024, izihloko ezigqamile zamemezela ukuthi ubuhlakani bokwenziwa maduzane bebungabhala yonke ikhodi yethu, kushiye abathuthukisi abangabantu bengenakwenza lutho. Kodwa ngemuva kwe-hype ne-sensationalism, iqiniso lihluke kakhulu.

Yebo, i-AI manje isingakwazi ukukhiqiza ikhodi ngokushesha kunanoma yimuphi umuntu, kodwa inhle kangakanani leyo khodi, futhi ingakwazi yini ukuphatha i-AI yonke umjikelezo wokuphila wokuthuthukiswa kwesofthiwe iyodwa? Ochwepheshe abaningi bathi “akusheshi kangako.” Abaholi bobunjiniyela be-software njengo-CEO we-Microsoft uSatya Nadella bagcizelela ukuthi “i-AI ngeke ithathe indawo yabahleli bohlelo, kodwa izoba ithuluzi elibalulekile endaweni yabo yezikhali. Imayelana nokunikeza abantu amandla okwenza okwengeziwe, hhayi okuncane.” ( Ingabe I-AI Izothatha Indawo Yabahleli Bezinhlelo? Iqiniso Elingemuva Kwe-Hype | yi-PyCoach | Ikhona Lokwenziwa | Mar, 2025 | Medium ) Ngokufanayo, isikhulu se-AI ye-Google uJeff Dean siphawula ukuthi nakuba i-AI ikwazi ukusingatha imisebenzi yokubhala ikhodi evamile, "isantula ubuhlakani kanye namakhono okuxazulula izinkinga" - zona kanye izimfanelo ezilethwa abathuthukisi betafula. Ngisho no-Sam Altman, oyi-CEO ye-OpenAI, uyavuma ukuthi i-AI yanamuhla “yinhle kakhulu emisebenzini” kodwa “yimbi kakhulu emisebenzini egcwele” ngaphandle kokuqondisa komuntu. Ngamafuphi, i-AI inhle ekusizeni ngezingcezu zomsebenzi, kodwa ayikwazi ukuthatha ngokuphelele umsebenzi womdidiyeli kusukela ekuqaleni kuya ekupheleni.

Leli phepha elimhlophe libheka ngokwethembeka, nokulinganiselayo embuzweni othi “Ingabe i-AI izothatha indawo yabahleli bohlelo?” Sihlola ukuthi i-AI iyithinta kanjani indima yokuthuthukiswa kwesoftware namuhla nokuthi yiziphi izinguquko ezisazokwenzeka. Ngezibonelo zomhlaba wangempela namathuluzi akamuva (kusuka ku-GitHub Copilot kuya ku-ChatGPT), sihlola ukuthi onjiniyela bangalungisa kanjani, bazivumelanise, futhi bahlale bebalulekile njengoba i-AI ishintsha. Kunokuba kube nempendulo elula ethi yebo-noma-cha, sizobona ukuthi ikusasa ukusebenzisana phakathi kwe-AI nabathuthukisi abangabantu. Umgomo uwukugqamisa imininingwane esebenzayo kulokho onjiniyela abangakwenza ukuze baphumelele eminyakeni yobudala ye-AI - kusukela ekutholeni amathuluzi amasha kuya ekufundeni amakhono amasha kanye nokuveza ukuthi imisebenzi yokubhala amakhodi ingase ithuthuke kanjani eminyakeni ezayo.

I-AI Ekuthuthukisweni Kwesoftware Namuhla

I-AI isheshe yazihlanganisa nomsebenzi wesimanje wokuthuthukisa isoftware. Kunokuba kube inganekwane yesayensi, amathuluzi asekelwe ku-AI asebhala futhi abuyekeza ikhodi , azenzela imisebenzi eyisicefe, futhi athuthukisa ukukhiqiza konjiniyela. Onjiniyela namuhla basebenzisa i-AI ukuze bakhiqize amazwibela ekhodi, imisebenzi yokuqedela ngokuzenzakalela, ukubona iziphazamisi, ngisho namacala okuhlola umsebenzi wezandla ( Ingabe Likhona Ikusasa Lonjiniyela Besofthiwe? Umthelela We-AI [2024] ) ( Ingabe Likhona Ikusasa Lonjiniyela Besofthiwe? Umthelela we-AI [2024] ). Ngamanye amazwi, i-AI ithatha umsebenzi we-grunt kanye ne-boilerplate, ivumela abahleli bohlelo ukuthi bagxile ezicini eziyinkimbinkimbi zokwakhiwa kwesoftware. Ake sibheke amanye amakhono namathuluzi e-AI avelele aguqula izinhlelo njengamanje:

• Ukukhiqiza Ikhodi & Ukuqedela ngokuzenzakalela: Abasizi bokufaka amakhodi be-AI besimanje bangakhiqiza ikhodi ngokusekelwe ekwazisweni kolimi lwemvelo noma umongo wekhodi ingxenye. Isibonelo, i-GitHub Copilot (eyakhelwe kumodeli ye-OpenAI's Codex) ihlanganisa nabahleli ukuze iphakamise umugqa olandelayo noma ibhulokhi yekhodi njengoba uthayipha. Isebenzisa isethi enkulu yokuqeqeshwa yekhodi yomthombo ovulekile ukuze inikeze iziphakamiso eziqaphela umongo, ngokuvamile ekwazi ukuqedela imisebenzi yonke ukusuka kumazwana noma igama lomsebenzi. Ngokufanayo, i-ChatGPT (GPT-4) ingakwazi ukukhiqiza ikhodi yomsebenzi othile uma uchaza okudingayo ngesiNgisi esilula. Lawa mathuluzi angabhala ikhodi ye-boilerplate ngamasekhondi, kusukela emisebenzini elula yomsizi ukuya emisebenzini ye-CRUD evamile.

• Ukutholwa Nokuhlolwa Kweziphazamisi: I-AI futhi isiza ukubamba amaphutha futhi ithuthukise ikhwalithi yekhodi. Amathuluzi okuhlaziya anamandla we-AI nama-linter angamaka iziphazamisi ezingaba khona noma ubungozi bokuphepha ngokufunda kumaphethini eziphazamisi zesikhathi esidlule. Amanye amathuluzi e-AI akhiqiza ngokuzenzakalelayo ukuhlolwa kweyunithi noma aphakamise izimo zokuhlola ngokuhlaziya izindlela zekhodi. Lokhu kusho ukuthi unjiniyela angathola impendulo esheshayo ngamakesi asemaphethelweni okungenzeka ukuthi awaphuthelwe. Ngokuthola iziphazamisi kusenesikhathi futhi iphakamise ukulungiswa, i-AI isebenza njengomsizi we-QA ongakhathali osebenza eduze nonjiniyela.

• Ukwenziwa Kwekhodi Nokwenza Kabusha: Okunye ukusetshenziswa kwe-AI kuphakamisa ukuthuthukiswa kwekhodi ekhona. Ngokunikezwa amazwibela, i-AI ingancoma ama-algorithms asebenza kahle kakhulu noma ukusetshenziswa okuhlanzekile ngokubona amaphethini kukhodi. Isibonelo, ingase iphakamise ukusetshenziswa okunengqondo kakhulu kwelabhulali noma ikhodi engasasebenzi efulegi engase yenziwe kabusha. Lokhu kusiza ekwehliseni izikweletu zobuchwepheshe kanye nokwenza ngcono ukusebenza. Amathuluzi okwenza kabusha asuselwa ku-AI angaguqula ikhodi ukuze ihambisane nemikhuba emihle noma ibuyekeze ikhodi ibe izinguqulo ze-API, konga isikhathi sonjiniyela ekuhlanzeni mathupha.

• I-DevOps ne-Automation: Ngaphandle kwekhodi yokubhala, i-AI inikela ekwakhiweni nasekusetshenzisweni kwezinqubo. Amathuluzi e-CI/CD ahlakaniphile asebenzisa ukufunda komshini ukuze abikezele ukuthi iziphi izivivinyo okungenzeka zehluleke noma zibeke phambili imisebenzi ethile yokwakha, okwenza ipayipi lokuhlanganisa eliqhubekayo lisheshe futhi lisebenze kahle kakhudlwana. I-AI ingahlaziya amalogi okukhiqiza namamethrikhi okusebenza ukuze ikhombe izinkinga noma iphakamise ukulungiselelwa kwengqalasizinda. Empeleni, i-AI ayisizi nje kuphela ekubhaleni amakhodi, kodwa kuwo wonke umjikelezo wokuphila wokuthuthukiswa kwesofthiwe - kusukela ekuhleleni kuya ekulungisweni.

• I-Natural Interfaces Nokubhalwa Kwezilimi: Siphinde sibone i-AI inika amandla ukusebenzisana okungokwemvelo okwengeziwe ngamathuluzi okuthuthukisa. Onjiniyela bangacela ngokoqobo i -AI ukuthi yenze imisebenzi (“yenza umsebenzi owenza u-X” noma “chaza le khodi”) futhi bathole imiphumela. Ama-chatbots e-AI (njenge-ChatGPT noma abasizi be-dev abakhethekile) bangaphendula imibuzo yokuhlela, basize ngombhalo, futhi babhale imibhalo yephrojekthi noma benze imilayezo esekelwe ekushintsheni kwekhodi. Lokhu kuvala igebe phakathi kwenhloso yomuntu kanye nekhodi, okwenza intuthuko ifinyeleleke kalula kulabo abakwazi ukuchaza abakufunayo.

• 

Onjiniyela abasebenzisa amathuluzi e-AI: Inhlolovo yango-2023 ikhombisa ukuthi inani eliphakeme kakhulu elingu-92% lonjiniyela basebenzise amathuluzi okufaka amakhodi e-AI ngandlela thile - kungaba emsebenzini, kumaphrojekthi abo, noma kokubili. Bangu-8% kuphela ababike ukuthi abasebenzisi noma yiluphi usizo lwe-AI ekubhaleni ikhodi. Leli shadi libonisa ukuthi izingxenye ezimbili kwezintathu zonjiniyela zisebenzisa amathuluzi e-AI ngaphakathi nangaphandle komsebenzi, kuyilapho ikota iwasebenzisa ngokukhethekile emsebenzini kanye nedlanzana elincane kuphela langaphandle komsebenzi. Okuthathwayo kucacile: Ukubhala ngekhodi okusizwa yi-AI sekuvele kwanda ngokushesha phakathi konjiniyela ( Inhlolovo iveza umthelela we-AI kulwazi lonjiniyela - Ibhulogi ye-GitHub ).

Lokhu kwanda kwamathuluzi e-AI ekuthuthukisweni kuholele ekwenyukeni kokusebenza kahle kanye nokunciphisa isidina ekubhaleni amakhodi. Imikhiqizo idalwe ngokushesha njengoba i-AI isiza ukukhiqiza ikhodi ye-boilerplate futhi isingathe imisebenzi ephindaphindwayo ( Ingabe Likhona Ikusasa Lonjiniyela Besofthiwe? Umthelela We-AI [2024] ) ( Ingabe I-AI Izoshintsha Onjiniyela ngo-2025: Ukubheka Ikusasa ). Amathuluzi afana ne-Copilot angakwazi ngisho nokuphakamisa ama-algorithms aphelele noma izisombululo “ezingase zingabonakali ngokushesha kubathuthukisi abangabantu,” ngenxa yokufunda kumadathasethi amaningi wekhodi. Izibonelo zomhlaba wangempela ziningi: unjiniyela angacela i-ChatGPT ukuthi isebenzise umsebenzi wokuhlunga noma ithole iphutha kukhodi yayo, futhi i-AI izokhiqiza isisombululo esisalungiswa ngemizuzwana. Izinkampani ezifana ne -Amazon ne -Microsoft zithumele abahleli bezinhlelo ze-AI (i-CodeWhisperer ye-Amazon kanye ne-Microsoft's Copilot) emaqenjini abo onjiniyela, zibika ukuqedwa ngokushesha kwemisebenzi kanye namahora ambalwa avamile asetshenziswe ku-boilerplate. Eqinisweni, i-70% yabathuthukisi abacwaningwa ku-2023 Stack Overflow inhlolovo bathi sebevele basebenzisa noma bahlela ukusebenzisa amathuluzi e-AI ohlelweni lwabo lokuthuthukisa ( 70% yabathuthukisi basebenzisa amathuluzi ekhodi ye-AI, 3% bathemba kakhulu ukunemba kwabo - ShiftMag ). Abasizi abadume kakhulu yi-ChatGPT (esetshenziswa ~83% yabaphenduli) kanye ne-GitHub Copilot (~56%), okubonisa ukuthi izingxoxo ezijwayelekile ze-AI kanye nabasizi abahlanganisiwe be-IDE bobabili abadlali ababalulekile. Onjiniyela ngokuyinhloko baphendukela kulawa mathuluzi ukuze bakhulise ukukhiqiza (okucashunwe ngu-~33% wabaphenduli) futhi basheshise ukufunda (25%), kuyilapho cishe u-25% bewasebenzisa ukuze basebenze kahle kakhulu ngokwenza umsebenzi ophindaphindayo.

Kubalulekile ukuqaphela ukuthi indima ye-AI ekuhleleni ayisiyona entsha ngokuphelele - izici zayo sezineminyaka ikhona (cabanga ngokuqedela ngokuzenzakalela kwekhodi kuma-IDE noma izinhlaka zokuhlola ezizenzakalelayo). Kodwa eminyakeni emibili edlule kube yiphuzu elibalulekile. Ukuvela kwamamodeli wolimi amakhulu anamandla (njengochungechunge lwe-OpenAI's GPT kanye ne-DeepMind's AlphaCode) kukhulise ngendlela emangazayo okungenzeka. Isibonelo, -AlphaCode yaba sematheni ngokudlala ngezinga lokuncintisana lokuhlela izinhlelo , yafinyelela cishe izinga eliphezulu elingu-54% ezinseleleni zokubhala amakhodi - ngokuyisisekelo ifanelana nekhono lomuntu oqhudelana naye omaphakathi ( I-AlphaCode ye-DeepMind ifana nekhono elimaphakathi lomklami ). Bekungokokuqala ngqa uhlelo lwe-AI lwenza ngokuncintisana emiqhudelwaneni yokuhlela. Kodwa-ke, kuyasho ukuthi ngisho ne-AlphaCode, nawo wonke amandla ayo, yayisekude nokushaya amakhodi abantu angcono kakhulu. Kuleyo miqhudelwano, i-AlphaCode ingaxazulula cishe u-30% wezinkinga phakathi kwemizamo evunyelwe, kuyilapho abahleli bezinhlelo abangabantu abaphezulu bexazulula> izinkinga ezingu-90% ngomzamo owodwa. Leli gebe ligqamisa ukuthi nakuba i-AI ikwazi ukusingatha imisebenzi ye-algorithmic echazwe kahle kuze kufike eqophelweni elithile, izinkinga ezinzima kakhulu ezidinga ukucabanga okujulile nobuhlakani zihlala ziyinqaba yomuntu .

Kafushane, i-AI izitshale yaqina kukhithi yamathuluzi yansuku zonke yonjiniyela. Kusukela ekusizeni ekubhaleni ikhodi kuya ekuthuthukiseni ukusetshenziswa, ithinta yonke ingxenye yenqubo yokuthuthukisa. Ubudlelwano namuhla buyavumelana kakhulu: I-AI isebenza njengomshayeli wendiza (oqanjwe ngokufanelekile) esiza onjiniyela bafake ikhodi ngokushesha nangokukhungatheka okuncane, kunomshayeli ozimele ongakwazi ukundiza yedwa. Esigabeni esilandelayo, sizocubungula ukuthi lokhu kuhlanganiswa kwamathuluzi e-AI kuyishintsha kanjani indima yonjiniyela kanye nemvelo yomsebenzi wabo, kube ngcono noma kubi.

I-AI Ishintsha Kanjani Izindima Zonjiniyela Nokukhiqiza

Njengoba i-AI iphatha umsebenzi omningi ojwayelekile, indima yonjiniyela wesoftware isiqala ukuvela. Kunokuba bachithe amahora bebhala ikhodi ye-boilerplate noma ukulungisa amaphutha avamile, abathuthukisi bangakwazi ukulayisha leyo misebenzi kubasizi babo be-AI. Lokhu kuguqula ukugxila konjiniyela kufinyelele ekuxazululeni izinkinga kwezinga eliphezulu, izakhiwo, kanye nezici zokudala zobunjiniyela besofthiwe. Empeleni, i-AI yandisa onjiniyela, ibavumela ukuthi bakhiqize kakhudlwana futhi babe namandla amasha. Kodwa ingabe lokhu kuhumusha imisebenzi embalwa yokuhlela, noma uhlobo oluhlukile lomsebenzi? Ake sihlole umthelela ekukhiqizeni nasezindimeni:

Ukuthuthukisa Ukukhiqiza: Ngama-akhawunti amaningi nezifundo zakuqala, amathuluzi wokubhala amakhodi we-AI akhulisa ukukhiqiza konjiniyela kakhulu. Ucwaningo lwe-GitHub luthole ukuthi onjiniyela abasebenzisa i-Copilot bakwazi ukuqeda imisebenzi ngokushesha kakhulu kunalabo abangenalo usizo lwe-AI. Kwesinye isilingo, onjiniyela baxazulule umsebenzi wokubhala amakhodi ngokushesha okungaba ngu-55% ngokwesilinganiso ngosizo lwe-Copilot – kuthathe cishe ihora elingu-1 amaminithi angu-11 esikhundleni samahora angu-2 amaminithi angu-41 ngaphandle kwawo ( Ucwaningo: ukulinganisa umthelela we-GitHub Copilot ekukhiqizeni nasekujabuleni konjiniyela - I-GitHub Blog ). Lokho kuyinzuzo emangazayo ngesivinini. Akusona isivinini kuphela; abathuthukisi babika ukuthi usizo lwe-AI lusiza ukunciphisa ukukhungatheka “nokuphazamiseka kokugeleza”. Ezinhlolovo, i-88% yabathuthukisi abasebenzisa i-Copilot bathi ibenze bakhiqiza kakhulu futhi yabavumela ukuba bagxile emsebenzini owanelisayo ( Imaphi amaphesenti onjiniyela athi i-github copilot yenza ... ). Lawa mathuluzi asiza abahleli bezinhlelo ukuthi bahlale “endaweni” ngokuphatha izingcezu eziyisicefe, zona ezigcina amandla engqondo ezinkingeni ezinzima. Ngenxa yalokho, onjiniyela abaningi banomuzwa wokuthi ukubhala ngekhodi sekujabulisa kakhulu - umsebenzi omncane wokukhononda nokudala okwengeziwe.

Ukushintsha Umsebenzi Wansuku zonke: Ukuhamba komsebenzi kwansuku zonke komklami kuyashintsha ngokuhambisana nalezi zinzuzo zokukhiqiza. Iningi "lomsebenzi omatasa" - ukubhala i-boilerplate, ukuphindaphinda amaphethini avamile, ukucinga i-syntax - kungalayishwa ku-AI. Isibonelo, esikhundleni sokubhala ngokuzenzela ikilasi ledatha ngama-getter nama-setters, umthuthukisi angavele atshele i-AI ukuthi iyikhiqize. Esikhundleni sokuhlanganisa imibhalo ukuze uthole ikholi ye-API efanele, unjiniyela angabuza i-AI ngolimi lwemvelo. Lokhu kusho ukuthi onjiniyela bachitha isikhathi esincane ngokuqhathaniswa ekubhaleni ngekhanda kanye nesikhathi esengeziwe emisebenzini edinga ukwahlulela komuntu . Njengoba i-AI ithatha umsebenzi wokubhala amakhodi alula angu-80%, umsebenzi womthuthukisi ushintshela ekuqondiseni okukhiphayo kwe-AI (ukubuyekeza iziphakamiso zekhodi, ukuzihlola) nokubhekana no-20% wezinkinga ezikhohlisayo i-AI engakwazi ukuzithola. Empeleni, umthuthukisi angase aqale usuku lwakhe ngokuhlola izicelo zokudonsa ezenziwe nge-AI noma abuyekeze iqoqo lokulungiswa okuphakanyiswe yi-AI, kunokuba abhale zonke lezo zinguquko kusukela ekuqaleni.

Ukusebenzisana kanye ne-Team Dynamics: Kuyathakazelisa ukuthi i-AI iphinde ibe nomthelela ekusebenzeni kweqembu. Ngemisebenzi evamile ezenzakalelayo, amaqembu angakwazi ukufeza okwengeziwe ngonjiniyela abambalwa abancane abanikezwe umsebenzi wokugruneka. Ezinye izinkampani zibika ukuthi onjiniyela bazo abakhulu bangakwazi ukuzimela - bangakwazi ukulinganisa izici ngokushesha ngosizo lwe-AI, ngaphandle kokudinga umfundi omncane ukwenza okusalungiswa kokuqala. Kodwa-ke, lokhu kuphakamisa inselele entsha: ukweluleka nokwabelana ngolwazi. Kunokuba abancane bafunde ngokwenza imisebenzi elula, bangadinga ukufunda ukuphatha ngempumelelo imiphumela ye-AI. Ukubambisana kweqembu kungase kushintshele emisebenzini efana nokucwenga ngokuhlanganyela imiyalo ye-AI noma ukubuyekeza ikhodi ekhiqizwe yi-AI ukuze uthole izingibe. Ngasohlangothini oluhle, lapho wonke umuntu oseqenjini enomsizi we-AI, ingalinganisa inkundla futhi ivumele isikhathi esengeziwe sezingxoxo zokuklama, ukucabanga okuhlakaniphile, kanye nokubhekana nezidingo eziyinkimbinkimbi zabasebenzisi okungekho i-AI okwamanje eyaziqonda ngaphandle kwebhokisi. Eqinisweni, abathuthukisi abangaphezu kwabane kwabahlanu bakholelwa ukuthi amathuluzi okubhala amakhodi e-AI azothuthukisa ukusebenzisana kweqembu noma okungenani alikhulule ukuze lisebenzisane kakhudlwana ekwakhiweni nasekuxazululeni izinkinga, ngokusho kwemiphumela yocwaningo lwe-GitHub lwango-2023 ( Ucwaningo luveza umthelela we-AI kulwazi lonjiniyela - Ibhulogi Ye-GitHub ).

Umthelela Ezindimeni Zomsebenzi: Umbuzo omkhulu wukuthi i-AI izonciphisa yini isidingo sabahleli bohlelo (njengoba umhleli wezinhlelo ngamunye manje esekhiqiza kakhulu), noma izomane iguqule amakhono afunekayo. Izibonelo zomlando nokunye okuzenzakalelayo (njengokukhuphuka kwamathuluzi e-devops, noma izilimi zokuhlela zezinga eliphezulu) kuphakamisa ukuthi imisebenzi yonjiniyela ayisuswanga kangako njengoba iphakanyiswa . Ngempela, abahlaziyi bezimboni babikezela ukuthi izindima zobunjiniyela besoftware zizoqhubeka nokukhula , kodwa uhlobo lwalezo zindima luzoshintsha. Umbiko wakamuva we-Gartner ubikezela ukuthi ngo-2027, i-50% yezinhlangano zobunjiniyela bezinhlelo zokusebenza zizosebenzisa amapulatifomu e-AI-augmented "software engineering intelligence" ukuze kukhuliswe umkhiqizo , kusukela ku-5% kuphela ngo-2024 ( Ingabe Likhona Ikusasa Lonjiniyela Besofthiwe? Umthelela we-AI [2024] ). Lokhu kukhombisa ukuthi izinkampani zizohlanganisa kabanzi i-AI, kodwa kusho ukuthi onjiniyela bazobe besebenza nalezo zinkundla ezihlakaniphile. Ngokufanayo, iphrojekthi yokubonisana nenkampani iMcKinsey ethi nakuba i-AI ingase izenzele imisebenzi eminingi, cishe u-80% wemisebenzi yokuhlela isazodinga umuntu onolwazi futhi ihlale “igxile kumuntu” . Ngamanye amazwi, sisazodinga abantu ezikhundleni eziningi zonjiniyela, kodwa izincazelo zomsebenzi zingashintsha.

Olunye ushintsho olungase lube khona ukuvela kwezindima ezifana ne- "AI Software Engineer" noma "Prompt Engineer" - onjiniyela abasebenza ngokukhethekile ekwakheni noma ekuhleleni izingxenye ze-AI. Sesivele sibona isidingo sonjiniyela abanobuchwepheshe be-AI/ML bukhuphuka kakhulu. Ngokusho kokuhlaziywa okwenziwe Ngempela, imisebenzi emithathu ehlobene ne-AI edingeka kakhulu isayensi yedatha, unjiniyela wesoftware, nonjiniyela wokufunda ngomshini , futhi isidingo salezi zindima siphindwe kabili kule minyaka emithathu edlule ( Ingabe Likhona Ikusasa Lonjiniyela Besoftware? Umthelela we-AI [2024] ). Onjiniyela bezinhlelo zendabuko balindeleke ngokwandayo ukuthi baqonde izisekelo zokufunda komshini noma bahlanganise amasevisi e-AI ezinhlelweni zokusebenza. Kunokuba yenze onjiniyela bangabi namsebenzi, "i-AI ingawuphakamisa umsebenzi, ivumele abathuthukisi ukuthi bagxile emisebenzini esezingeni eliphezulu nasekusunguleni izinto ezintsha." ( Ingabe I-AI Izongena Esikhundleni Sonjiniyela ngo-2025: I-Sneak Peek in the Future ) Imisebenzi eminingi yokubhala ikhodi evamile ingase isingathwe i-AI, kodwa onjiniyela bazoba matasa kakhulu ngomklamo wesistimu, ukuhlanganisa amamojula, ukuqinisekisa ikhwalithi, nokubhekana nezinkinga zamanoveli. Unjiniyela omkhulu ovela enkampanini eyodwa ye-AI-forward ukufingqe kahle: I-AI ayithathi indawo yabathuthukisi bethu; kuyabakhulisa .​ Unjiniyela oyedwa ohlome ngamathuluzi e-AI anamandla angenza umsebenzi wabaningana, kodwa lowo mthuthukisi manje usethatha umsebenzi onzima kakhulu futhi onomthelela.

Isibonelo Somhlaba Wangempela: Cabangela isimo esivela enkampanini yesofthiwe ehlanganise i-GitHub Copilot yabo bonke onjiniyela bayo. Umthelela osheshayo kwaba ukuncishiswa okuphawulekayo kwesikhathi esichithwa kubhalwa izivivinyo zeyunithi kanye nekhodi ye-boilerplate. Omunye unjiniyela osemncane uthole ukuthi esebenzisa i-Copilot angakwazi ukukhiqiza u-80% wekhodi yesici esisha ngokushesha, bese echitha isikhathi sakhe enza ngendlela oyifisayo u-20% osele futhi abhale izivivinyo zokuhlanganisa. Ukukhiqiza kwakhe mayelana nokukhishwa kwekhodi kucishe kwaphindeka kabili, kodwa okuthakazelisa nakakhulu, uhlobo lomnikelo wakhe lwashintsha - waba umbuyekezi wekhodi kanye nomklami wokuhlola wekhodi ebhaliwe ye-AI. Ithimba liphinde laqaphela ukuthi ukubuyekezwa kwekhodi kwaqala ukubamba amaphutha e-AI esikhundleni sokuthayipha kwabantu. Isibonelo, u-Copilot ngezikhathi ezithile wayephakamisa ukuqaliswa kokubethela okungavikelekile; abathuthukisi abangabantu kwadingeka babone futhi balungise lezo. Lolu hlobo lwesibonelo lukhombisa ukuthi ngenkathi okukhiphayo kwanda, ukwengamela kanye nobuchule bomuntu kwaba bucayi kakhulu ekuhambeni komsebenzi.

Ukufingqa, i-AI ngokungenakuphikwa ishintsha indlela onjiniyela abasebenza ngayo: ibenza basheshe futhi ibavumele ukuthi babhekane nezinkinga zokuvelela, kodwa futhi idinga ukuthi bathuthukise amakhono (kokubili ekuthuthukiseni i-AI nasekucabangeni kwezinga eliphezulu). Incane indaba yokuthi "i-AI ithatha imisebenzi" futhi indaba "ye-AI eshintsha imisebenzi." Onjiniyela abafunda ukusebenzisa lawa mathuluzi ngempumelelo bangakwazi ukuphindaphinda umthelela wabo - isisho esivame ukuzwa ukuthi, “I-AI ngeke ithathe indawo yonjiniyela, kodwa onjiniyela abasebenzisa i-AI bangangena esikhundleni salabo abangayisebenzisi.” Izigaba ezilandelayo zizohlola ukuthi kungani abathuthukisi abangabantu besabalulekile (lokho i-AI engakwazi ukuyenza kahle), nokuthi onjiniyela bangawajwayela kanjani amakhono abo ukuze baphumelele eduze kwe-AI.

Imikhawulo ye-AI (Kungani Abantu Behlala Bebalulekile)

Naphezu kwamakhono ayo amangalisayo, i-AI yanamuhla inemikhawulo eyivimbelayo ekwenzeni abahleli bezinhlelo abangabantu baphelelwe yisikhathi. Ukuqonda le mikhawulo kubalulekile ekuboneni ukuthi kungani abahleli bezinhlelo besadingeka kakhulu ohlelweni lokuthuthukisa. I-AI iyithuluzi elinamandla, kodwa akuyona inhlamvu yomlingo engangena esikhundleni sokusungula, ukucabanga okujulile, nokuqonda umongo womthuthukisi ongumuntu. Nawa amanye amaphutha ayisisekelo e-AI ezinhlelweni namandla ahambisanayo onjiniyela abangabantu:

• Ukuntuleka Kokuqonda Kweqiniso Nokudala: Amamodeli amanje e-AI awaqondi kahle ikhodi noma izinkinga ngendlela abantu abayenza ngayo; babona amaphethini futhi bavuselele imiphumela okungenzeka isekelwe kudatha yokuqeqeshwa. Lokhu kusho ukuthi i-AI ingazabalaza nemisebenzi edinga izixazululo zangempela, zokudala noma ukuqonda okujulile kwezizinda zezinkinga zenoveli. I-AI ingase ikwazi ukukhiqiza ikhodi ukuze ihlangabezane nencazelo eke yabonwa ngaphambili, kodwa iyicele ukuba idizayine i-algorithm entsha yenkinga engakaze ibonwe noma ihumushe imfuneko engaqondakali, futhi cishe izontengantenga. Njengoba omunye owayebukele akubeka, i-AI namuhla “ayinawo amandla okucabanga okudala nokucabangela alethwa abathuthukisi abangabantu etafuleni.” ( Ingabe I-AI Izongena Esikhundleni Sonjiniyela ngo-2025: I-Sneak Peek in the Future ) Abantu bahamba phambili ekucabangeni ngaphandle kwebhokisi - ukuhlanganisa ulwazi lwesizinda, intuition, kanye nekhono lokusungula ukuklama izakhiwo zesofthiwe noma ukuxazulula izinkinga eziyinkimbinkimbi. I-AI, ngokuphambene, iboshelwe kumaphethini ewafundile; uma inkinga ingahambisani nalawo maphethini kahle, i-AI ingase ikhiqize ikhodi engalungile noma engenangqondo (ngokuvamile ngokuzethemba!). Ukuqanjwa kwesofthiwe - okuza nezici ezintsha, ulwazi olusha lomsebenzisi, noma izindlela zobuchwepheshe ezinoveli - kusalokhu kuwumsebenzi oqhutshwa umuntu.

• Umongo kanye Nokuqonda Kwesithombe Esikhulu: Isofthiwe yokwakha ayikona nje ukubhala imigqa yekhodi. Kubandakanya ukuqonda ukuthi kungani ngemuva kwekhodi - izidingo zebhizinisi, izidingo zabasebenzisi, kanye nomxholo lapho isofthiwe isebenza khona. I-AI inefasitela elincane kakhulu lomongo (imvamisa ikhawulelwe kokokufaka okunikezwa ngesikhathi). Ayiqondi ngempela inhloso enkulu yesistimu noma ukuthi imojuli eyodwa isebenzisana kanjani nenye ngale kwalokho okukukhodi ngokusobala. Njengomphumela, i-AI ingase ikhiqize ikhodi esebenzela umsebenzi omncane ngokobuchwepheshe kodwa engangeni kahle ekwakhiweni kwesistimu enkulu noma ephule imfuneko ethile engacacile. Abathuthukisi abangabantu bayadingeka ukuze kuqinisekiswe ukuthi isofthiwe ihambisana nemigomo yebhizinisi kanye nokulindelwe ngabasebenzisi. Idizayini yezinhlelo eziyinkimbinkimbi - ukuqonda ukuthi uguquko engxenyeni ethile lungase lunyakaze kanjani kwezinye, indlela yokulinganisa ukuhwebelana (njengokusebenza ngokumelene nokufunda), kanye nendlela yokuhlela ukuvela kwe-codebase yesikhathi eside - yinto i-AI engeke iyenze namuhla. Kumaphrojekthi amakhulu anezinkulungwane zezingxenye, i-AI “ibona izihlahla kodwa hhayi ihlathi.” Njengoba kuphawulwe ekuhlaziyweni okukodwa, "I-AI idonsa kanzima ukuqonda umongo ogcwele nobunzima bephrojekthi ye-software enkulu," okufaka izidingo zebhizinisi nokucatshangelwa kokuhlangenwe nakho komsebenzisi ( Ingabe i-AI Izofaka Esikhundleni Sonjiniyela ngo-2025: Ukubheka Kwekusasa ). Abantu bagcina umbono wesithombe esikhulu.

• Ukulungiswa Komqondo Ovamile kanye Nokungaqondakali: Izidingo kumaphrojekthi wangempela ngokuvamile azicacile noma ziyavela. Umthuthukisi ongumuntu angakwazi ukufuna ukucaciswa, enze ukucabangela okunengqondo, noma ahlehlise izicelo ezingenangqondo. I-AI ayinayo ukucabanga okunengqondo noma ikhono lokubuza imibuzo ecacisayo (ngaphandle uma ifakwe ngokucacile ekwazisweni, futhi noma kunjalo ayinaso isiqinisekiso sokuyilungisa). Kungakho ikhodi ekhiqizwe i-AI kwesinye isikhathi ingalunga ngokobuchwepheshe kodwa ingasebenzi kahle - ayinakho ukwahlulela kokwazi ukuthi umsebenzisi ubehloseni ngempela uma imiyalelo ingacacile. Ngokuphambene, umhleli womuntu angatolika isicelo sezinga eliphezulu (“enze le UI ibe enembile” noma “uhlelo lokusebenza kufanele luphathe okokufaka okungajwayelekile kahle”) futhi lithole ukuthi yini okufanele yenziwe ngekhodi. I-AI izodinga ukucaciswa okuningiliziwe, okucacile ukuze kungene unjiniyela ngempela, futhi nokubhala leyo ncazelo ngempumelelo kunzima njengokubhala ikhodi ngokwayo. Njengoba isihloko soMkhandlu we-Forbes Tech siphawulwe ngokufanelekile, ukuze i-AI ingene esikhundleni sonjiniyela, kuzodingeka iqonde imiyalo engacacile futhi ivumelane nomuntu - izinga lokucabanga lamanje i-AI ayinalo ( I-Sergii Kuzin's Post - LinkedIn ).

• Ukwethembeka kanye “Nemibono”: Amamodeli e-AI akhiqizayo anamuhla anephutha elaziwayo: angakhiqiza imiphumela engalungile noma eqanjwe ngokuphelele, into evame ukubizwa ngokuthi i-allucination . Ekubhaleni amakhodi, lokhu kungase kusho ukuthi i-AI ibhala ikhodi ebukeka izwakala kodwa ingalungile noma ingavikelekile. Onjiniyela abakwazi ukwethemba ngokungaboni iziphakamiso ze-AI. Empeleni, yonke ingxenye yekhodi ebhaliwe ye-AI idinga ukubuyekezwa nokuhlolwa ngokucophelela ngumuntu . Idatha yenhlolovo ye-Stack Overflow ibonisa lokhu - kulabo abasebenzisa amathuluzi e-AI, u-3% kuphela othembela kakhulu ukunemba kokukhiphayo kwe-AI, futhi ngempela iphesenti elincane alikwethembi ( 70% wonjiniyela basebenzisa amathuluzi okubhala amakhodi e-AI, 3% bathemba kakhulu ukunemba kwabo - ShiftMag ). Iningi lonjiniyela liphatha iziphakamiso ze-AI njengamacebo awusizo, hhayi ivangeli. Lokhu kuthembana okuphansi kuyafaneleka ngoba i-AI ingenza amaphutha ayinqaba okungekho muntu onekhono ongayenza (njengamaphutha angekho-ngalinye, ukusebenzisa imisebenzi eyehlisiwe, noma ukukhiqiza izixazululo ezingasebenzi kahle) ngoba ayicabangi ngempela ngenkinga. Njengoba ukuphawula okukodwa kwenkundla kuphawulwe kabi, “Ba (ama-AI) abona izinto ezingekho emthethweni futhi benza izinqumo zokuklama ezixakile umuntu angeke aze azenze” ( Ingabe abahleli bohlelo bazophelelwa yisikhathi ngenxa ye-AI? - Iseluleko Somsebenzi ). Ukwengamela komuntu kubalulekile ukuze sibambe la maphutha. I-AI ingase ikutholele u-90% wesici ngokushesha, kodwa uma u-10% osele unesiphazamisi esicashile, isawela kunjiniyela ongumuntu ukuze asixilonge futhi asilungise. Futhi uma kukhona okungahambi kahle ekukhiqizweni, onjiniyela abangabantu okufanele balungise - i-AI ayikwazi okwamanje ukuziphendulela ngamaphutha ayo.

• Ukugcina kanye Nokuguqula I-Codebases: Amaphrojekthi we-software aphila futhi akhule ngokuhamba kweminyaka. Zidinga isitayela esingaguquki, ukucaciseleka kwabanakekeli besikhathi esizayo, nezibuyekezo njengoba izidingo zishintsha. I-AI namuhla ayinayo inkumbulo yezinqumo zesikhathi esidlule (ngaphandle kokwaziswa okulinganiselwe), ngakho-ke ingase ingagcini ikhodi ihambisana kuyo yonke iphrojekthi enkulu ngaphandle uma iqondiswa. Abathuthukisi abangabantu baqinisekisa ukugcinwa kwekhodi - ukubhala imibhalo ecacile, ukukhetha izixazululo ezifundekayo kunezihlakaniphile-kodwa-ezicashile, kanye nekhodi yokufaka kabusha njengoba kudingeka lapho isakhiwo sishintsha. I-AI ingasiza kule misebenzi (njengokuphakamisa ukwenza kabusha), kodwa ukunquma ukuthi yini okufanele yenziwe kabusha noma ukuthi yiziphi izingxenye zesistimu ezidinga ukuklanywa kabusha kuwucingo lokwahlulela lomuntu. Ngaphezu kwalokho, lapho uhlanganisa izingxenye, ukuqonda umthelela wesici esisha kumamojula akhona (ukuqinisekisa ukuhambisana okubuyela emuva, njll.) yinto abantu abayiphathayo. Ikhodi ekhiqizwe yi-AI kumele ihlanganiswe futhi ivunyelaniswe ngabantu. Njengokuhlola, abanye onjiniyela bazamile ukuvumela i-ChatGPT yakhe zonke izinhlelo zokusebenza ezincane; umphumela uvamise ukusebenza ekuqaleni kodwa kuba nzima kakhulu ukuwugcina noma ukunweba ngenxa yokuthi i-AI ayisebenzisi njalo ukwakheka okucatshangelwayo - yenza izinqumo zendawo umakhi ongumuntu angagwema.

• Ukucatshangelwa Kwezimiso Zokuhle Nokuvikeleka: Njengoba i-AI ibhala ikhodi eyengeziwe, futhi iphakamisa imibuzo yokuchema, ukuphepha, kanye nokuziphatha. I-AI ingase yethule ingaqondile ubungozi bokuphepha (isibonelo, ukungahlanzeki kahle okokufaka, noma ukusebenzisa izinqubo ze-cryptographic ezingavikelekile) ezingabanjwa unjiniyela ongumuntu ongumakadebona. Futhi, i-AI ayinawo umuzwa wemvelo wezimiso zokuziphatha noma ukukhathazeka ngokungenzeleli - ngokwesibonelo, ingase iqeqeshe ngedatha echemile futhi iphakamise ama-algorithms abandlulula ngokungenhloso (esicini esiqhutshwa yi-AI njengekhodi yokugunyaza imalimboleko noma i-algorithm yokuqasha). Abathuthukisi abangabantu bayadingeka ukuze bahlole imiphumela ye-AI yalezi zinkinga, baqinisekise ukuthotshelwa kwemithetho, futhi bafake isoftware ngokucatshangelwa kokuziphatha. Ingxenye yezenhlalo yesofthiwe - ukuqonda ukwethembeka komsebenzisi, ukukhathazeka kobumfihlo, kanye nokwenza ukukhetha kwemiklamo ehambisana nezindinganiso zomuntu - "akunakuzitshwa. Lezi zici zentuthuko ezigxile kumuntu zingaphezu kokufinyelelwa yi-AI, okungenani esikhathini esizayo esibonakalayo." ( Ingabe I-AI Izongena Esikhundleni Sonjiniyela ngo-2025: I-Sneak Peek into the Future ) Onjiniyela kufanele basebenze njengonembeza nesango lekhwalithi leminikelo ye-AI.

Ngokucabangela le mikhawulo, ukuvumelana kwamanje ukuthi i-AI iyithuluzi, hhayi ukumiselela . Njengoba u-Satya Nadella eshilo, kumayelana amandla , hhayi ukubashintsha ( Ingabe I-AI Izothatha Indawo Yabahleli Bezinhlelo? Iqiniso Elingemuva Kwe-Hype | nge-PyCoach | Ikhona Lokwenziwa | Mar, 2025 | Medium ). I-AI ingacatshangwa njengomsizi omncane: iyashesha, ayikhathali, futhi ingathatha iphasi yokuqala emisebenzini eminingi, kodwa idinga isiqondiso sonjiniyela ophezulu kanye nobuchwepheshe ukuze ikhiqize umkhiqizo wokugcina opholishiwe. Kuyasho ukuthi ngisho nezinhlelo ezithuthuke kakhulu zokufaka amakhodi ze-AI zisetshenziswa njengabasizi ekusetshenzisweni komhlaba wangempela (Copilot, CodeWhisperer, njll.) hhayi njengamakhodi azimele. Izinkampani aziwaxoshi amaqembu azo wokuhlela futhi zivumela i-AI ukuthi isebenze ngendlela exakile; esikhundleni salokho, bashumeka i-AI ekuhambeni komsebenzi konjiniyela ukuze babasize.

Isicaphuna esisodwa esingokomfanekiso sivela ku-OpenAI ka-Sam Altman, owaphawula ukuthi njengoba ama-ejenti e-AI ethuthuka, “lawa ma-ejenti e-AI ngeke athathe indawo yabantu ngokuphelele” ekuthuthukisweni kwesoftware ( u-Sam Altman uthi ama-agent e-AI maduzane azokwenza imisebenzi eyenziwa onjiniyela be-software: Indaba ephelele ngamaphuzu angu-5 - India Today ). Bazosebenza “njengabasebenza ndawonye” abasingatha imisebenzi echazwe kahle yonjiniyela abangabantu, ikakhulukazi leyo misebenzi ejwayelekile yonjiniyela wesoftware wezinga eliphansi onolwazi lweminyaka embalwa. Ngamanye amazwi, i-AI ingase yenze umsebenzi womthuthukisi omncane kwezinye izindawo, kodwa lowo njiniyela osemncane akaphelelwa umsebenzi - aguqukela endimeni yokuqondisa i-AI kanye nokubhekana nemisebenzi yezinga eliphezulu i-AI engakwazi ukuyenza. Ngisho nokubheka esikhathini esizayo, lapho abanye abacwaningi bebikezela ukuthi ngo-2040 i-AI ingase ibhale amakhodi ayo amaningi ( Ingabe Likhona Ikusasa Lonjiniyela Besofthiwe? Umthelela we-AI [2024] ), kuvunyelwana ngokuthi abahleli bezinhlelo abangabantu basazodingeka ukuze baqondise, baqondise, futhi banikeze inhlansi yokudala nokucabanga okubucayi okuntula imishini .

Kubalulekile futhi ukuqaphela ukuthi ukuthuthukiswa kwesofthiwe kungaphezu nje kokubhala ikhodi . Kubandakanya ukuxhumana nababambiqhaza, ukuqonda izindaba zabasebenzisi, ukubambisana emaqenjini, nokwakheka okuphindaphindwayo - zonke izindawo lapho amakhono abantu ebaluleke kakhulu. I-AI ayikwazi ukuhlala emhlanganweni neklayenti ukuze isheshe iveze ukuthi ifunani ngempela, futhi ayikwazi ukuxoxisana ngezinto ezibalulekile noma ukukhuthaza iqembu ngombono womkhiqizo. Isici somuntu sihlala simaphakathi.

Kafushane, i-AI inobuthakathaka obubalulekile: abukho ubuhlakani bangempela, ukuqonda okulinganiselwe komongo, ukuthambekela kwamaphutha, akukho sibopho sokuziphendulela, kanye nokungaqondi imithelela ebanzi yezinqumo zesoftware. Lezi zikhala yilapho kanye abathuthukisi abangabantu bekhanya khona. Kunokuba ubuke i-AI njengosongo, kungase kunembe kakhudlwana ukuyibuka njenge -amplifier enamandla yabathuthukisi abangabantu - ukuphatha okujwayelekile ukuze abantu bakwazi ukugxila ekujuleni. Isigaba esilandelayo sizodingida ukuthi onjiniyela bangasebenzisa kanjani lokhu kukhuliswa ngokulungisa amakhono nezindima zabo ukuze zihlale zibalulekile futhi zibalulekile emhlabeni othuthukayo ongathandwa yi-AI.

Ukuzijwayeza kanye Nokuchuma Ngenkathi ye-AI

Kubahleli bohlelo nabathuthukisi, ukukhuphuka kwe-AI ekubhaleni amakhodi akufanele kube usongo olukhulu - kungaba yithuba. Okubalulekile ukuzivumelanisa nokuziphendukela kwemvelo kanye nobuchwepheshe. Labo abafunda ukusebenzisa i-AI bazozithola kakhulu futhi bedingeka kakhulu, kuyilapho labo abangayinaki bangathola ukuthi basalele emuva. Kulesi sigaba, sigxile ezinyathelweni ezisebenzayo namasu okuba onjiniyela bahlale behambisana futhi bechuma njengoba amathuluzi e-AI eba yingxenye yentuthuko yansuku zonke. Umqondo ongawusebenzisa owokufunda okuqhubekayo nokusebenzisana ne-AI, kunokuncintisana. Nansi indlela onjiniyela abangalungisa ngayo nokuthi yimaphi amakhono amasha nezindima okufanele bazicabangele:

1. Yamukela i-AI Njengethuluzi (Funda Ukusebenzisa Abasizi Bekhodi Be-AI Ngokuphumelelayo): Okokuqala nokubaluleke kakhulu, onjiniyela kufanele bakhululeke ngamathuluzi e-AI atholakalayo. Phatha i-Copilot, i-ChatGPT, noma amanye ama-AI wokubhala amakhodi njengozakwethu omusha wokuhlela. Lokhu kusho ukufunda indlela yokubhala ukwaziswa okuhle noma ukuphawula ukuze uthole iziphakamiso zekhodi eziwusizo, kanye nokwazi ukuthi ungaqinisekisa kanjani ngokushesha noma ususe iphutha ikhodi ekhiqizwe yi-AI. Njengoba nje unjiniyela bekumele afunde i-IDE yakhe noma ukulawula inguqulo, ukufunda izinto ezingezinhle zomsizi we-AI kuba yingxenye yesethi yamakhono. Isibonelo, unjiniyela angazijayeza ngokuthatha ucezu lwekhodi alubhalile futhi acele i-AI ukuthi iyithuthukise, bese ehlaziya izinguquko. Noma, lapho uqala umsebenzi, wuveze kumazwana futhi ubone ukuthi i-AI inikeza ini, bese ulungisa kusukela lapho. Ngokuhamba kwesikhathi, uzothuthukisa intuition yalokho i-AI ekwaziyo nokuthi ungakha kanjani ngokubambisana nayo. Kucabange “njengentuthuko esizwa yi-AI” – ikhono elisha ongalengeza ebhokisini lakho lamathuluzi. Ngempela, abathuthukisi manje bakhuluma "ngobunjiniyela obusheshayo" njengekhono - ukwazi ukubuza i-AI imibuzo efanele. Labo abakwaziyo bangathola imiphumela engcono kakhulu emathuluzi afanayo. Khumbula, “onjiniyela abasebenzisa i-AI bangangena esikhundleni salabo abangayisebenzisi” – ngakho-ke yamukela ubuchwepheshe futhi ubenze umngane wakho.

2. Gxila Kumakhono Ezinga Eliphezulu (Ukuxazulula Inkinga, Idizayini Yesistimu, Izakhiwo): Njengoba i-AI ikwazi ukuphatha ikhodi yezinga eliphansi, onjiniyela kufanele bakhuphule ileli yokukhipha . Lokhu kusho ukugcizelela kakhulu ukuqonda ukwakheka kwesistimu nezakhiwo. Hlakulela amakhono okudiliza izinkinga eziyinkimbinkimbi, ukuklama amasistimu angalinganisa, nokwenza izinqumo zezakhiwo - izindawo lapho ukuqonda komuntu kubaluleke kakhulu. Gxila ekutheni kungani futhi kanjani isixazululo, hhayi nje ukuthi yini. Isibonelo, esikhundleni sokuchitha sonke isikhathi sakho uphelelisa umsebenzi wokuhlunga (uma i-AI ingakubhalela eyodwa), chitha isikhathi uqonda ukuthi iyiphi indlela yokuhlunga elungele umongo wohlelo lwakho lokusebenza nokuthi ingena kanjani ekugelezeni kwedatha yohlelo lwakho. Ukucabanga kokuklama - kucatshangelwa izidingo zabasebenzisi, ukuhamba kwedatha, nokusebenzisana kwengxenye - kuzokwaziswa kakhulu. I-AI ingakwazi ukukhiqiza ikhodi, kodwa ngunjiniyela onquma isakhiwo sonke sesofthiwe futhi aqinisekise ukuthi zonke izingxenye zisebenza ngokuvumelana. Ngokulola ukucabanga kwakho kwesithombe esikhulu, uzenza ubaluleke kakhulu njengomuntu oqondisa i-AI (kanye neqembu lonke) ekwakheni into efanele. Njengoba omunye umbiko obheka isikhathi esizayo waphawula, abathuthukisi kufanele “bagxile ezindaweni lapho ukuqonda kwabantu kungenakuthatheka, njengokuxazulula izinkinga, ukucabanga komklamo, nokuqonda izidingo zabasebenzisi.” ( Ingabe i-AI izongena esikhundleni sonjiniyela ngo-2025: Ukulunguza Kwekusasa )

3. Thuthukisa Ulwazi Lwakho lwe-AI ne-ML: Ukuze usebenze eduze ne-AI, kuyasiza ukuqonda i-AI . Onjiniyela akudingekile ukuthi bonke babe abacwaningi bokufunda ngomshini, kodwa ukuba nokuqonda okuqinile kokuthi la mamodeli asebenza kanjani kuzoba yinzuzo. Funda izinto eziyisisekelo zokufunda ngomshini nokufunda okujulile - lokhu akukwazi nje ukuvula izindlela ezintsha zemisebenzi (njengoba imisebenzi ehlobene ne-AI idlondlobala ( Ingabe Likhona Ikusasa Lonjiniyela Besofthiwe? Umthelela we-AI [2024] ), kodwa futhi kuzokusiza ukuthi usebenzise amathuluzi e-AI ngokuphumelelayo. Uma wazi, ngokwesibonelo, imikhawulo yemodeli yolimi enkulu nokuthi yaqeqeshwa kanjani, ungabikezela ukuthi ingase yehlule nini futhi uklame ukwaziswa kwakho noma izivivinyo ngokufanele. Ukwengeza, imikhiqizo eminingi yesofthiwe manje isihlanganisa izici ze-AI (isibonelo, uhlelo lokusebenza olunenjini yokuncoma noma i-chatbot). Umthuthukisi wesofthiwe onolwazi oluthile lwe-ML angaba negalelo kulezo zici noma okungenani ahlanganyele ngobuhlakani nososayensi bedatha. Izindawo ezibalulekile okufanele zicatshangelwe ukufunda zihlanganisa: izisekelo zesayensi yedatha , indlela yokucubungula idatha kusengaphambili, ukuqeqeshwa uma kuqhathaniswa nokucatshangwayo, kanye nezimiso zokuziphatha ze-AI. Zijwayeze nezinhlaka ze-AI (TensorFlow, PyTorch) kanye nezinsizakalo ze-AI zamafu; ngisho noma ungawakhi amamodeli kusukela ekuqaleni, ukwazi ukuthi ungayihlanganisa kanjani i-AI API ohlelweni lokusebenza kuyikhono elibalulekile. Ngamafuphi, ukuba "i-AI literate" ngokushesha kubaluleka njengokukwazi ukufunda kuwebhu noma ku-database technologies. Onjiniyela abangakwazi ukugxambukela emhlabeni wobunjiniyela besofthiwe bendabuko kanye ne-AI bazoba sesimweni esihle sokuhola amaphrojekthi azayo.

4. Thuthukisa Amakhono Athambile Aqinile kanye Nolwazi Lwesizinda: Njengoba i-AI ithatha imisebenzi yemishini, amakhono omuntu ayingqayizivele aba abaluleke kakhulu. Ukuxhumana, ukusebenzisana, kanye nobungcweti besizinda yizindawo okufanele sibhekane nazo kabili. Ukuthuthukiswa kwesofthiwe ngokuvamile kumayelana nokuqonda isizinda senkinga - noma ngabe ezezimali, ukunakekelwa kwezempilo, imfundo, nanoma yimuphi omunye umkhakha - nokuhumusha lokho kube yizixazululo. I-AI ngeke ibe nalowo mongo noma ikhono lokuxhumana nababambe iqhaza, kodwa unalo. Ukuba nolwazi oluthe xaxa esizindeni osebenza kuso kukwenza ube umuntu oya kuye ukuze uqinisekise ukuthi isofthiwe ihlangabezana nezidingo zomhlaba wangempela. Ngokufanayo, gxila kumakhono akho okubambisana: ukuqeqeshwa, ubuholi, nokuxhumana. Amaqembu asazodinga abathuthukisi abaphezulu ukuze babuyekeze ikhodi (kuhlanganise nekhodi ebhaliwe ye-AI), ukuze bacebise abancane ngemikhuba engcono kakhulu, futhi baxhumanise amaphrojekthi ayinkimbinkimbi. I-AI ayisusi isidingo sokusebenzelana kwabantu kumaphrojekthi. Eqinisweni, ngekhodi ekhiqizayo ye-AI, ukwelulekwa komthuthukisi ophezulu kungase kushintshele ekufundiseni abancane ukuthi bangasebenza kanjani ne-AI futhi baqinisekise ukuphuma kwayo , esikhundleni sokubhala iluphu. Ukwazi ukuqondisa abanye kule paradigm entsha kuyikhono elibalulekile. Futhi, zijwayeze ukucabanga okujulile - buza futhi uhlole imiphumela ye-AI, futhi ukhuthaze abanye ukuthi benze okufanayo. Ukuhlakulela ukungabaza okunempilo kanye nengqondo yokuqinisekisa kuzovimbela ukuthembela ngobumpumputhe ku-AI futhi kunciphise amaphutha. Empeleni, thuthukisa amakhono i-AI engenawo: ukuqonda abantu nomongo, ukuhlaziya okubucayi, kanye nokucabanga kwemikhakha eyahlukene.

5. Ukufunda Impilo Yonke Nokuzivumelanisa Nezimo: Ijubane loshintsho ku-AI lishesha kakhulu. Okuzwakala kuyi-Cutting-Edge namuhla kungase kuphelelwe yisikhathi eminyakeni embalwa. Onjiniyela kufanele bamukele ukufunda impilo yonke kunangaphambili. Lokhu kungase kusho ukuzama njalo abasizi bekhodi be-AI abasha, ukuthatha izifundo ze-inthanethi noma izitifiketi ku-AI/ML, ukufunda amabhulogi ocwaningo ukuze uhlale unolwazi lwakamuva ngalokho okuzayo, noma ukubamba iqhaza emiphakathini yonjiniyela egxile ku-AI. Ukuzivumelanisa nezimo kubalulekile - lungele ukuzungeza kumathuluzi amasha nokugeleza komsebenzi njengoba kuvela. Isibonelo, uma ithuluzi elisha le-AI lifika elingakwazi ukwenza ukwakheka kwe-UI ngokuzenzakalelayo kusuka ezidwetshweni, umthuthukisi ongaphambili kufanele alungele ukufunda futhi akuhlanganise lokho, aguqule ukugxila kwakhe mhlawumbe ekucwengiseni i-UI ekhiqiziwe noma ukuthuthukisa imininingwane yomsebenzisi egejiwe ngokuzenzakalela. Labo abaphatha ukufunda njengengxenye eqhubekayo yomsebenzi wabo (okuyinto onjiniyela abaningi asebevele beyenza) bayokuthola kulula ukuhlanganisa ukuthuthukiswa kwe-AI. Elinye isu wukunikela ingxenye encane yesonto lakho ekufundeni nasekuhloleni - kuphathe njengokutshala imali ekusaseni lakho. Izinkampani nazo seziqala ukuhlinzeka ngoqeqesho kubathuthukisi bazo ekusebenziseni amathuluzi e-AI ngempumelelo; ukuwasebenzisa ngokunenzuzo amathuba anjalo kuyokubeka phambili. Onjiniyela abaphumelelayo kuzoba yilabo ababona i-AI njengozakwethu othuthukayo futhi baqhubeke bephucula indlela yabo yokusebenza nalowo zakwethu.

6. Hlola Izindima Ezisafufusa kanye Nezindlela Zomsebenzi: Njengoba i-AI ingena ekuthuthukisweni, amathuba amasha emisebenzi ayavela. Isibonelo, i-Prompt Engineer noma i-AI Integration Specialist izindima ezigxile ekudaleni ukwaziswa okufanele, ukuhamba komsebenzi, nengqalasizinda yokusebenzisa i-AI emikhiqizweni. Esinye isibonelo i-AI Ethics Engineer noma i-AI Auditor - izindima ezigxile ekubuyekezeni imiphumela ye-AI ngokuchema, ukuthobelana, nokunemba. Uma unentshisekelo kulezo zindawo, ukuzibeka endaweni ngolwazi olulungile kungavula lezi zindlela ezintsha. Ngisho nangaphakathi kwendima yakudala, ungase uthole izindawo ezifana “nonjiniyela osizwa yi-AI” vs “unjiniyela osizwa yi-AI” lapho ngamunye esebenzisa amathuluzi akhethekile. Bheka ukuthi izinhlangano ziwakha kanjani amaqembu azungeze i-AI. Ezinye izinkampani “zinezinhlangano ze-AI” noma izikhungo ezisezingeni eliphezulu zokuqondisa ukwamukelwa kwe-AI kumaphrojekthi - ukusebenza emaqenjini anjalo kungakubeka phambili. Ngaphezu kwalokho, cabanga ukufaka isandla ekuthuthukisweni kwamathuluzi e-AI ngokwawo: isibonelo, ukusebenza kumaphrojekthi womthombo ovulekile athuthukisa amathuluzi kanjiniyela (mhlawumbe athuthukise ikhono le-AI lokuchaza ikhodi, njll.). Lokhu akugcini nje ngokujulisa ukuqonda kwakho ubuchwepheshe kodwa futhi kukubeka emphakathini ohola izinguquko. Okubalulekile wukukhuthala mayelana nokuba bukhali emsebenzini . Uma izingxenye zomsebenzi wakho wamanje zizenzakalela, lungele ukushintsha ezindimeni ezidizayina, zengamele, noma ezengeza lezo zingxenye ezizenzakalelayo.

7. Gcina futhi Ubonise Ikhwalithi Yomuntu: Emhlabeni lapho i-AI ingakwazi ukukhiqiza ikhodi emaphakathi yenkinga emaphakathi, abathuthukisi abangabantu kufanele balwele ukukhiqiza ezihlukile nezinozwela - AI engeke ikwazi. Lokhu kungase kusho ukugxila ku-finesse yokuhlangenwe nakho komsebenzisi, ukuthuthukiswa kokusebenza kwezimo ezingajwayelekile, noma ukubhala nje ikhodi ehlanzekile nebhalwe kahle (i-AI ayilungile ekubhaleni imibhalo enengqondo noma ukuphawula kwekhodi okuqondakalayo - ungangeza inani lapho!). Kwenze iphuzu ukuhlanganisa ukuqonda komuntu emsebenzini: isibonelo, uma i-AI ikhiqiza ucezu lwekhodi, wengeza amazwana achaza isizathu ngendlela omunye umuntu angayiqonda kamuva, noma uyilungise ukuze ifundeke kakhudlwana. Ngokwenza kanjalo, ungeza ungqimba lobuchwepheshe kanye nekhwalithi engenawo umsebenzi owenziwe ngomshini. Ngokuhamba kwesikhathi, ukwakha idumela lesofthiwe yekhwalithi ephezulu “esebenza nje” emhlabeni wangempela kuzokuhlukanisa. Amaklayenti nabaqashi bazokwazisa onjiniyela abakwazi ukuhlanganisa ukusebenza kahle kwe-AI nobuciko bomuntu .

Ake sicabangele nokuthi izindlela zemfundo zingavumelana kanjani. Onjiniyela abasha abangena kulo mkhakha akufanele bagweme amathuluzi e-AI ohlelweni lwabo lokufunda. Ngokuphambene, ukufunda nge -AI (isb, ukusebenzisa i-AI ukusiza ngomsebenzi wesikole noma amaphrojekthi, bese uhlaziya imiphumela) kungasheshisa ukuqonda kwabo. Kodwa-ke, kubalulekile ukuthi futhi ufunde okuyisisekelo ngokujulile - ama-algorithms, izakhiwo zedatha, nemiqondo yokuhlela ewumongo - ukuze ube nesisekelo esiqinile futhi ukwazi ukubona lapho i-AI iduka. Njengoba i-AI isingatha izivivinyo zokubhala amakhodi ezilula, ikharikhulamu ingase ibeke isisindo esiningi kumaphrojekthi adinga ukuklanywa nokuhlanganiswa. Uma usanda kufika, gxila ekwakheni iphothifoliyo ebonisa ikhono lakho lokuxazulula izinkinga eziyinkimbinkimbi nokusebenzisa i-AI njengelinye lamathuluzi amaningi.

Ukuhlanganisa isu lokuzijwayeza: ube umshayeli wendiza, hhayi umgibeli. Sebenzisa amathuluzi e-AI, kodwa ungathembeli ngokweqile kuwo noma ukhululeke. Qhubeka ucije izici zentuthuko eziyingqayizivele yomuntu. UGrady Booch, iphayona elihlonishwayo lobunjiniyela bezinhlelo zesoftware, ukusho kahle: “I-AI izoshintsha ngokuyisisekelo ukuthi kusho ukuthini ukuba umdidiyeli. Ngeke ibaqede abahleli bohlelo, kodwa izodinga ukuthi bathuthukise amakhono amasha futhi basebenze ngezindlela ezintsha.” ( Ingabe Likhona Ikusasa Lonjiniyela Besofthiwe? Umthelela we-AI [2024] ). Ngokuthuthukisa ngokuqhubekayo lawo makhono amasha kanye nezindlela zokusebenza, onjiniyela bangaqinisekisa ukuthi bahlala beyisihlalo somshayeli emisebenzini yabo.

Ukufingqa lesi sigaba, nalu uhlu lokuhlola olusheshayo lwezithenjwa zonjiniyela abafuna ukufakazela imisebenzi yabo esikhathini esizayo eminyakeni ye-AI:

Ngokulandela lawa maqhinga, abathuthukisi bangaguqula ukuguquguquka kwe-AI kube inzuzo yabo. Labo abazivumelanisa nezimo bazothola ukuthi i-AI ithuthukisa amakhono abo futhi ibavumele ukuthi bakhiqize isoftware engcono kunangaphambili, kunokuba ibenze baphelelwe yisikhathi.

I-Future Outlook: Ukusebenzisana Phakathi Kwe-AI Nonjiniyela

Liyini ikusasa lezinhlelo emhlabeni oqhutshwa i-AI? Ngokusekelwe kumathrendi amanje, singalindela ikusasa lapho i-AI kanye nonjiniyela abangabantu besebenzisana eduze nakakhulu . Iqhaza lomhleli cishe lizoqhubeka nokushintshela esikhundleni sokuqondisa nesokudala, i-AI ibambe ngaphezulu "kokuphakamisa okunzima" ngaphansi kokuqondisa komuntu. Kulesi sigaba sokuphetha, senza ezinye zezimo zesikhathi esizayo futhi siqinisekise ukuthi umbono wonjiniyela ungahlala ulungile - inqobo nje uma siqhubeka nokuzivumelanisa nezimo.

Esikhathini esizayo esiseduze (eminyakeni emi-5-10 ezayo), maningi amathuba okuthi i-AI ibe yonke indawo enqubweni yokuthuthukiswa njengamakhompyutha ngokwawo. Njengoba nje engekho unjiniyela namuhla obhala ikhodi ngaphandle komhleli noma ngaphandle kwe-Google/StackOverflow ezandleni zabo, maduzane akekho unjiniyela ozobhala ikhodi ngaphandle kohlobo oluthile losizo lwe-AI olusebenza ngemuva. Izimo Zokuthuthukisa Ezididiyelwe (ama-IDE) sezivele ziyavela ukuze zifake izici ezinikwe amandla e-AI emnyombweni wazo (ngokwesibonelo, abahleli bekhodi abangakuchazela ikhodi noma baphakamise yonke izinguquko zekhodi kuyo yonke iphrojekthi). Singase sifinyelele endaweni lapho umsebenzi oyinhloko kanjiniyela kuwukwenza izinkinga nezithiyo ngendlela i-AI engayiqonda, bese ikhetha futhi icwengisise izixazululo ezinikezwa yi-AI . Lokhu kufana nohlobo lwezinga eliphezulu lokuhlela, ngezinye izikhathi olubizwa ngokuthi “uhlelo olusheshayo” noma “i-AI yomculo.”

Kodwa-ke, ingqikithi yalokho okumele kwenziwe - ukuxazulula izinkinga kubantu - ihlala ingashintshile. I-AI yesikhathi esizayo ingase ikwazi ukwenza lonke uhlelo lokusebenza ngencazelo (“ngakhele uhlelo lokusebenza lweselula ukuze ngibhukhe izikhathi zokubona udokotela”), kodwa umsebenzi wokucacisa leyo ncazelo, ukuqinisekisa ukuthi ilungile, nokulungisa kahle umphumela ukuze ujabulise abasebenzisi kuzobandakanya onjiniyela (kanye nabaklami, abaphathi bemikhiqizo, njll.). Eqinisweni, uma ukukhiqizwa kohlelo lokusebenza oluyisisekelo kuba lula, ubuhlakani bomuntu kanye nokuqanjwa kabusha kwesoftware kuzoba okubaluleke kakhulu ukuhlukanisa imikhiqizo. Singabona ukuchuma kwesoftware, lapho izinhlelo eziningi ezijwayelekile zenziwa yi-AI, kuyilapho abathuthukisi abangabantu begxile kumaphrojekthi aphambili, ayinkimbinkimbi, noma ahlakaniphile aphusha imingcele.

Kungenzeka futhi ukuthi umgoqo wokungena ezinhlelweni wehliswe - okusho ukuthi abantu abaningi abangebona onjiniyela bezinhlelo ze-software (ake bathi, umhlaziyi webhizinisi noma usosayensi noma umdayisi) bangakha isofthiwe besebenzisa amathuluzi e-AI (ukuqhubeka kokunyakaza "okungenayo ikhodi/ikhodi ephansi" ekhokhiswa kakhulu i-AI). Lokhu akusiqedi isidingo sabathuthukisi abangochwepheshe; kunalokho, kuyayishintsha. Onjiniyela bangase bathathe indima eyengeziwe yokubonisana noma yokuqondisa ezimweni ezinjalo, baqinisekise ukuthi lezi zinhlelo zokusebenza ezithuthukiswe isakhamuzi zivikelekile, ziyasebenza, futhi ziyanakekeleka. Abahleli bohlelo abangochwepheshe bangase bagxile ekwakhiweni kwamapulatifomu nama-API asetshenziswa yi-AI “abangebona abezinhlelo”.

Ngokombono wemisebenzi, izindima ezithile zokuhlela zingancipha kuyilapho ezinye zikhula. Isibonelo, ezinye izikhundla zokufaka amakhodi ezingeni lokungena zingaba mbalwa ngenani uma izinkampani zithembele ku-AI ngemisebenzi elula. Umuntu angacabanga ukuthi isiqalo esincane esikhathini esizayo sidinga ingxenye yenombolo yonjiniyela abancane ngoba ama-dev abo aphezulu, ahlome nge-AI, angathola umsebenzi omningi oyisisekelo. Kodwa ngesikhathi esifanayo, imisebenzi emisha ngokuphelele (njengoba sixoxile esigabeni sokuzivumelanisa) izovela. Ngaphezu kwalokho, njengoba isofthiwe ingena ngisho nangokwengeziwe emnothweni (ngesofthiwe ekhiqiza i-AI yezidingo ze-niche), isidingo sisonke semisebenzi ehlobene nesofthiwe singaqhubeka sikhuphuka. Umlando ukhombisa ukuthi i-automation ivamise ukuholela eminingi ngokuhamba kwesikhathi , nakuba kuyimisebenzi ehlukene - ngokwesibonelo, ukuzenzela kwemisebenzi ethile yokukhiqiza kuholele ekukhuleni kwemisebenzi yokuklama, ukunakekela, kanye nokwenza ngcono izinhlelo ezizenzakalelayo. Ngomongo we-AI nohlelo, kuyilapho eminye imisebenzi i-dev encane ejwayele ukuyenza izenzakalela, ububanzi bayo yonke i-software esifuna ukuyidala iyakhula (ngoba manje ishibhile/kushesha ukuyidala), okungaholela kumaphrojekthi amaningi futhi ngaleyo ndlela isidingo sokwengamela komuntu okwengeziwe, ukuphathwa kwephrojekthi, ukwakha, njll. Umbiko we-World Economic Forum ku-World Economic Forum uyaphakanyiswa ku-World Economic Forum mayelana nokuthuthukiswa kwe-software yesikhathi esizayo . isidingo, hhayi ukuncipha, ngenxa yenguquko yedijithali.

Kufanele futhi sicabangele ukubikezela kwe-2040 okukhulunywe ngakho ngaphambili: abacwaningi base-Oak Ridge National Lab baphakamise ukuthi ngo-2040, "imishini ... izobhala iningi lamakhodi abo" ( Ingabe Likhona Ikusasa Lonjiniyela Besofthiwe? Umthelela we-AI [2024] ). Uma lokho kubonakala kunembile, yini esele kubahleli bezinhlelo abangabantu? Ngokunokwenzeka, ukugxila kuzoba ekuqondisweni kwezinga eliphezulu kakhulu (ukutshela imishini ukuthi yini esifuna ukuthi iyifeze ngamagalelo abanzi) nasezindaweni ezibandakanya ukuhlanganiswa okuyinkimbinkimbi kwezinhlelo, ukuqonda kwengqondo yomuntu, noma izizinda zezinkinga zenoveli. Ngisho nakusimo esinjalo, abantu bebengathatha izindima ezifana nabaklami bemikhiqizo, onjiniyela abayizimfuneko, kanye nabaqeqeshi/abaqinisekisayo be-AI . Ikhodi ingase izibhale ngokwayo, kodwa othile kufanele anqume ukuthi iyiphi ikhodi okufanele ibhalwe nokuthi kungani , bese eqinisekisa ukuthi umphumela ulungile futhi uhambisana nemigomo. Kuyafana nokuthi izimoto ezizishayelayo zingase zizishayele kanjani ngelinye ilanga, kodwa usatshela imoto ukuthi kufanele iye kuphi futhi ingenelele ezimeni ezinzima - kanye nabantu abaklama imigwaqo, imithetho yomgwaqo, nazo zonke izingqalasizinda ezizungezile.

Ngakho ochwepheshe abaningi bacabanga ngekusasa lokubambisana, hhayi ukumiselela . Njengoba omunye umxhumanisi wezobuchwepheshe akubeka, "ikusasa lentuthuko alikona ukukhetha phakathi kwabantu noma i-AI kodwa ukubambisana okunikeza okungcono kakhulu kwakho kokubili." ( Ingabe I-AI Izongena Esikhundleni Sonjiniyela ngo-2025: I-Sneak Peek into the Future ) I-AI ngokungangabazeki izoguqula ukuthuthukiswa kwesofthiwe, kodwa iwukuvela kwendima yonjiniyela kunokushabalala. Onjiniyela "abamukela izinguquko, bajwayelane namakhono abo, futhi bagxile ezicini ezihlukile zobuntu zomsebenzi wabo" bazothola ukuthi i-AI ithuthukisa amakhono abo kunokunciphisa inani labo.

Singadweba ukufana nomunye umkhakha: cabanga ngokukhuphuka komklamo osizwa ngekhompyutha (CAD) kwezobunjiniyela nezakhiwo. Ingabe lawo mathuluzi athathela indawo onjiniyela nabadwebi bezakhiwo? Cha - babenze bakhiqiza kakhulu futhi babavumela ukuthi benze imiklamo eyinkimbinkimbi. Kodwa ubuhlakani bomuntu kanye nokuthatha izinqumo kwahlala kuyingqikithi. Ngokufanayo, i-AI ingabonwa njenge-Computer-Assisted Coding - izosiza ukuphatha inkimbinkimbi nomsebenzi we-grunt, kodwa unjiniyela uhlala engumklami nomenzi wezinqumo.

Ngokuhamba kwesikhathi, uma sicabanga nge-AI ethuthuke ngempela (ake sithi, uhlobo oluthile lwe-AI oluvamile olungase lukwazi ukwenza okuningi okungenziwa umuntu), ukushintsha kwezenhlalo nezomnotho kungaba banzi kakhulu kunokuhlela nje. Asikabi lapho, futhi sinokulawula okubalulekile kokuthi siyihlanganisa kanjani i-AI emsebenzini wethu. Indlela ehlakaniphile iwukuqhubeka nokuhlanganisa i-AI ngezindlela ezandisa amandla omuntu . Lokho kusho ukutshala imali kumathuluzi nezinqubo (nezinqubomgomo) ezigcina abantu benolwazi. Kakade, sibona izinkampani ezisungula ukubusa kwe-AI – imihlahlandlela yokuthi i-AI kufanele isetshenziswe kanjani ekuthuthukisweni ukuze kuqinisekiswe imiphumela yezimiso zokuziphatha nephumelelayo ( Inhlolovo iveza umthelela we-AI kokuhlangenwe nakho konjiniyela - I-GitHub Blog ). Lesi simo cishe sizokhula, siqinisekise ukuthi ukwengamela komuntu ngokomthetho kuyingxenye yepayipi lokuthuthukisa i-AI.

Sengiphetha, umbuzo othi "Ingabe i-AI izothatha indawo yabahleli bohlelo?" ingaphendulwa: Cha - kodwa izoshintsha kakhulu okwenziwa abahleli bohlelo. Izingxenye ezijwayelekile zohlelo zisendleleni yokuthi zizenzekele. Izingxenye zokudala, eziyinselelo, nezigxile kumuntu zizohlala lapha, futhi zizogqama kakhulu. Ikusasa lizobona abahleli bezinhlelo besebenzisana nabasizi be-AI abahlale behlakaniphe, njengelungu leqembu. Cabanga nje unozakwethu we-AI ongakwazi ukukhipha ikhodi 24/7 - kuyisikhuthazo esihle sokukhiqiza, kodwa kusadinga othile ozoyitshela ukuthi yimiphi imisebenzi okufanele isebenze futhi ahlole umsebenzi wayo.

Imiphumela emihle kakhulu izotholwa yilabo abaphatha i-AI njengomhlanganyeli. Njengoba omunye oyi-CEO ekubeka, “I-AI ngeke ithathe indawo yabahleli bezinhlelo, kodwa abahleli bezinhlelo abasebenzisa i-AI bazongena esikhundleni salabo abangayisebenzisi.” Ngamagama asebenzayo, lokhu kusho ukuthi isibopho sikubathuthukisi ukuthi bathuthuke ngobuchwepheshe. Umsebenzi wokuhlela awufi – uyazivumelanisa nezimo . Kuzoba nenqwaba ye-software ezokwakhiwa kanye nezinkinga ezizoxazululwa ngekusasa elibonakalayo, mhlawumbe nangaphezulu kunanamuhla. Ngokuhlala ufundile, uhlale uvumelana nezimo, futhi ugxile kulokho abantu abakwenza kangcono kakhulu, abathuthukisi bangakwazi ukuvikela umsebenzi ophumelelayo nogculisayo ngokubambisana ne-AI .

Okokugcina, kufanelekile ukugubha iqiniso lokuthi singena enkathini lapho abathuthukisi benamandla amakhulu. Isizukulwane esilandelayo sabahleli bohlelo sizofeza ngamahora lokho obekuthatha izinsuku, futhi sibhekane nezinkinga ebezingafinyeleleki ngaphambilini, ngokusebenzisa i-AI. Kunokuba wesabe, umuzwa wokuqhubekela phambili ungase ube owokuba nethemba kanye nelukuluku . Uma nje sisondela ku-AI sivule amehlo - siqaphela imikhawulo yayo futhi sikhumbula umthwalo wethu wemfanelo - singabumba ikusasa lapho i-AI nabahleli bezinhlelo bendawonye bakhe amasistimu e-software amangalisayo, okungaphezu kwalokho okungenziwa yedwa. Ubuhlakani bomuntu obuhlanganiswe nokusebenza kahle komshini kuyinhlanganisela enamandla. Ekugcineni, akuphathelene nokushintshwa , kodwa mayelana ne-synergy. Indaba ye-AI nabahleli bezinhlelo isabhalwa - futhi izobhalwa ngumuntu nomshini, ndawonye.

Imithombo:

1. Brainhub, “Ingabe Likhona Ikusasa Lonjiniyela Besoftware? Umthelela We-AI [2024]” ( Ingabe Likhona Ikusasa Lonjiniyela Besoftware? Umthelela We-AI [2024] ).

2. U-Brainhub, uchwepheshe ucaphuna u-Satya Nadella noJeff Dean ku-AI njengethuluzi, hhayi ukushintshwa ( Ingabe Likhona Ikusasa Lonjiniyela Besofthiwe? Umthelela We-AI [2024] ) ( Ingabe Likhona Ikusasa Lonjiniyela Besofthiwe? Umthelela we-AI [2024] ).

3. Medium (PyCoach), “Ingabe I-AI Izothatha Indawo Yabahleli Bezinhlelo? Iqiniso Elisemuva Kwe-Hype” , iphawula okungokoqobo okungaqondakali vs hype ( Ingabe I-AI Izothatha Indawo Yabahleli Bezinhlelo? Iqiniso Elingemuva Kwe-Hype | nge-PyCoach | Ikhona Lokwenziwa | Mar, 2025 | Medium ) kanye nokusho okugcwele kuka-Sam Altman emsebenzini kodwa hhayi ukucaphuna okuhle emsebenzini we-AI.

4. I-DesignGurus, “Ingabe i-AI izothatha indawo yabathuthukisi… (2025)” , igcizelela ukuthi i-AI izokhulisa futhi iphakamise abathuthukisi kunokuba ibenze bangabi namsebenzi ( Ingabe I-AI Izongena Esikhundleni Sonjiniyela ngo-2025: Ukubheka Okuzayo ) kanye nokufakwa kuhlu kwezindawo ze-AI ezisalele (ubuciko, umongo, izimiso zokuziphatha).

5. I-Stack Overflow Developer Survey 2023, ukusetshenziswa kwamathuluzi e-AI ngo-70% wama-devs, ukwethenjwa okuphansi kokunemba (3% ukwethenjwa okukhulu) ( 70% wonjiniyela basebenzisa amathuluzi okubhala amakhodi e-AI, 3% bathemba kakhulu ukunemba kwabo - ShiftMag ).

6. Inhlolovo ye-GitHub yango-2023, ebonisa u-92% wonjiniyela bazamile amathuluzi okubhala amakhodi e-AI futhi u-70% babona izinzuzo ( Inhlolovo iveza umthelela we-AI kulwazi lonjiniyela - I-GitHub Blog ).

7. Ucwaningo lwe-GitHub Copilot, ithola ukuqedwa komsebenzi osheshayo okungu-55% ngosizo lwe-AI ( Ucwaningo: ukulinganisa umthelela we-GitHub Copilot ekukhiqizeni nasekujabuleni konjiniyela - I-GitHub Blog ).

8. I-GeekWire, ku-AlphaCode ye-DeepMind esebenza ngokwesilinganiso ngeleveli yamakhodi womuntu (phezulu 54%) kodwa kude nabadlali abaphezulu ( I-DeepMind's AlphaCode ifana nekhono elimaphakathi lomklami ).

9. I-IndiaToday (Feb 2025), isifinyezo sombono ka-Sam Altman we-AI “asebenzisana nabo” abenza imisebenzi yonjiniyela abancane kodwa “ngeke bathathe indawo yabantu ngokuphelele” ( u-Sam Altman uthi ama-agent e-AI maduzane azokwenza imisebenzi eyenziwa onjiniyela be-software: Indaba ephelele ngamaphuzu angu-5 - India Today ).

10. McKinsey & Company, balinganisela ukuthi ~ 80% wemisebenzi yokuhlela izohlala igxile kumuntu naphezu kokuzenzakalela ( Ingabe Likhona Ikusasa Lonjiniyela Besofthiwe? Umthelela we-AI [2024] ).

Izindatshana ongathanda ukuzifunda ngemva kwalesi:

🔗 Amathuluzi Ohlelo Ahamba Phambili e-AI
Hlola amathuluzi e-AI ahamba phambili angahlanganyela nawe njengozakwethu wokubhala amakhodi ukuze uthuthukise ukuthuthuka komsebenzi wakho.

🔗 Iyiphi I-AI Engcono Kakhulu Ekubhaleni Ikhodi - Abasizi Bekhodi Be-AI Abaphezulu
Umhlahlandlela wamathuluzi e-AI asebenza kahle kakhulu okukhiqiza amakhodi, ukulungisa iphutha, kanye nokusheshisa amaphrojekthi esofthiwe.

🔗 I-Artificial Intelligence Software Development - Ukuguqula Ikusasa Lezobuchwepheshe
Qonda ukuthi i-AI iyishintsha kanjani indlela isofthiwe eyakhiwe ngayo, evivinywayo, nesetshenziswa ngayo.