Ingabe onjiniyela besofthiwe bazothathelwa indawo yi-AI?

Ingabe Onjiniyela Besofthiwe Bazothathelwa Indawo Yi-AI?

Lona omunye wemibuzo ekhathazayo, ephazamisayo kancane engena ezingxoxweni zasebusuku ze-Slack kanye nezingxoxo eziqhutshwa yikhofi phakathi kwabakhi bekhodi, abasunguli, kanye nanoma ubani owake wabheka iphutha eliyimfihlakalo. Ngakolunye uhlangothi, amathuluzi e-AI aqhubeka eshesha, ebukhali, cishe ngendlela exakile endleleni akhipha ngayo ikhodi. Ngakolunye uhlangothi, ubunjiniyela besofthiwe abuzange bube nje ngokuchaza indlela yokubhala. Ake siphinde sibheke emuva - ngaphandle kokungena kuskripthi se-sci-fi esivamile esibizwa ngokuthi “imishini izothatha izintambo”.

Izihloko ongase uthande ukuzifunda ngemva kwalesi:

🔗 Amathuluzi e-AI aphezulu okuhlola isofthiwe
Thola amathuluzi okuhlola asebenzisa i-AI enza i-QA ibe hlakaniphe futhi isheshe.

🔗 Ungaba kanjani unjiniyela we-AI
Umhlahlandlela wesinyathelo ngesinyathelo wokwakha umsebenzi ophumelelayo ku-AI.

🔗 Amathuluzi e-AI angcono kakhulu angenakhodi
Dala kalula izixazululo ze-AI ngaphandle kokufaka ikhodi usebenzisa amapulatifomu aphezulu.

Onjiniyela Besofthiwe Babalulekile 🧠✨

Ngaphansi kwawo wonke amakhibhodi kanye nezindlela zokulandelela izitaki, ubunjiniyela bulokhu buxazulula izinkinga, ubuhlakani, kanye nokwahlulela ezingeni lesistimu. Impela, i-AI ingakhipha izingcezu noma ngisho nokwakha uhlelo lokusebenza ngemizuzwana, kodwa onjiniyela bangempela baletha izinto imishini engazithinti nhlobo:

  • Ikhono lokuqonda umongo.

  • Ukwenza ukuhweba (isivinini vs. izindleko vs. ukuphepha… njalo kuyisenzo sokuxubana).

  • Ukusebenza nabantu, hhayi nje ikhodi.

  • Ukubamba amakesi angavamile angalingani nephethini elihle.

Cabanga nge-AI njengomfundi oqeqeshwayo osheshayo futhi ongakhathali. Uwusizo? Yebo. Ukuqondisa ukwakheka kwezakhiwo? Cha.

Cabanga ngalokhu: ithimba lokukhula lifuna isici esihambisana nemithetho yamanani, i-logic endala yokukhokha, kanye nemikhawulo yamanani. I-AI ingabhala izingxenye zayo, kodwa inqume ukuthi izoyibeka kuphi i-logic, ukuthi izoyithatha siphi isikhathi, nokuthi izoyichitha kanjani ama-invoyisi phakathi nokufuduka - ukuthi isinqumo sokwahlulela singesomuntu. Yilowo umehluko.

Lokho Okuboniswa Yidatha Ngempela 📊

Izinombolo ziyamangaza. Ezifundweni ezihlelekile, abathuthukisi abasebenzisa i-GitHub Copilot baqede imisebenzi ngokushesha okungu-55% kunalabo ababhala amakhodi bodwa [1]. Imibiko ebanzi yensimu? Ngezinye izikhathi kufika ku-2× ngokushesha nge-gen-AI efakwe emisebenzini yokusebenza [2]. Ukwamukelwa nakho kukhulu: ama-84% onjiniyela basebenzisa noma bahlela ukusebenzisa amathuluzi e-AI, futhi ochwepheshe abangaphezu kwengxenye bawasebenzisa nsuku zonke [3].

Kodwa kukhona ukushwabana. Umsebenzi obuyekezwe ontanga uphakamisa ukuthi ababhali bekhodi abasizwa yi-AI babenamathuba amaningi okubhala ikhodi engaphephile - futhi bavame ukuhamba bezethemba ngokweqile ngayo [5]. Yingakho izinhlaka zivikela ukucindezeleka: ukuqapha, ukuhlola, ukubuyekezwa kwabantu, ikakhulukazi ezizindeni ezibucayi [4].

Ukuxhumana Okusheshayo: I-AI vs. Onjiniyela

Isici Amathuluzi e-AI 🛠️ Onjiniyela Besofthiwe 👩💻👨💻 Kungani Kubalulekile
Isivinini Umbani ezingcezwini zokubhonga [1][2] Kancane kancane, ngokucophelela kakhudlwana Isivinini esiluhlaza akuyona umklomelo
Ubuciko Ihlanganiswe nedatha yayo yokuqeqeshwa Ngingasungula ngempela Ukusungula izinto ezintsha akusikho ukukopisha iphethini
Ukulungisa amaphutha Iphakamisa ukulungiswa kwendawo Uyaqonda ukuthi kungani yaphuka Imbangela ebalulekile
Ukubambisana Umqhubi wedwa Ufundisa, uxoxisana, uxhumana Isofthiwe = ukusebenzisana
Izindleko 💵 Kushibhile ngomsebenzi ngamunye Ibiza kakhulu (umholo + izinzuzo) Izindleko eziphansi ≠ umphumela ongcono
Ukuthembeka Ama-hallucinates, ukuphepha okuyingozi [5] Ukwethembana kukhula ngokuhlangenwe nakho Inani lokuphepha kanye nokwethenjwa
Ukuthobela imithetho Idinga ukuhlolwa nokuphathwa [4] Imiklamo yemithetho nokuhlolwa kwezimali Akuxoxiswana ngazo emikhakheni eminingi

Ukwanda Kwabasizi Bokubhala Ikhodi Ye-AI 🚀

Amathuluzi afana nama-Copilot nama-IDE asebenzisa i-LLM ashintsha indlela yokusebenza

  • I-boilerplate ephrintiwe ngokushesha.

  • Nikeza amacebiso okulungisa kabusha.

  • Chaza ama-API ongakaze uwathinte.

  • Ngisho nokukhipha izivivinyo (ngezinye izikhathi ziqhekekile, ngezinye izikhathi ziqinile).

Ushintsho? Imisebenzi yesigaba esincane manje isiyinto elula. Lokho kushintsha indlela abaqalayo abafunda ngayo. Ukugaya ngokusebenzisa izihibe ezingapheli akubalulekile kangako. Indlela ehlakaniphile: vumela i-AI ibhale phansi, bese uqinisekisa: bhala iziqinisekiso, sebenzisa ama-linters, uhlole ngobudlova, bese ubuyekeza amaphutha okuphepha angacacile ngaphambi kokuhlanganisa [5].

Kungani i-AI ingeyona indawo ephelele

Masikhulume ngokungagwegwesi: I-AI inamandla kodwa futhi… ayinalwazi. Ayinayo:

  • Ukuqonda - ukubamba izidingo ezingenangqondo.

  • Izimiso zokuziphatha - ukulinganisa ubulungisa, ukucwasa, ingozi.

  • Umongo - ukwazi ukuthi kungani isici kufanele noma kungafanele sibe khona.

Ngesofthiwe ebalulekile emsebenzini - ezezimali, ezempilo, ezokuhamba ngezindiza - awugembuli ohlelweni lwe-black-box. Izinhlaka zikwenza kucace: abantu bahlala benesibopho, kusukela ekuhlolweni kuya ekuqapheni [4].

Umphumela "Wokuphuthuma Okuphakathi" Emisebenzini 📉📈

I-AI ishaya kakhulu phakathi kweleveli yamakhono:

  • Onjiniyela bezinga lokungena: Ingabe isengozini - ukufaka ikhodi okuyisisekelo kuyazenzakalela. Indlela yokukhula? Ukuhlola, ukusebenzisa amathuluzi, ukuhlolwa kwedatha, ukubuyekezwa kokuphepha.

  • Onjiniyela/abakhi bezakhiwo abaphezulu: Kuphephile - ukuba nomklamo, ubuholi, ubunzima, kanye nokuhlela i-AI.

  • Ochwepheshe be-Niche: Kuphephile nakakhulu - ukuphepha, izinhlelo ezifakiwe, i-ML infra, izinto lapho izici zesizinda zibalulekile khona.

Cabanga ngama-calculator: awazange asuse izibalo. Ashintshe ukuthi yimaphi amakhono aba yinto ebalulekile.

Izici Zobuntu Ze-AI Ziyawa

Amanye amandla amakhulu onjiniyela i-AI asantula:

  • Ukulwa ngekhodi yefa le-spaghetti elibi kakhulu.

  • Ukufunda ukukhungatheka komsebenzisi nokufaka uzwela ekwakhiweni.

  • Ukuzulazula kwezepolitiki zehhovisi kanye nezingxoxo zamakhasimende.

  • Ukuzivumelanisa nezimo ezingakasungulwa.

Ngokumangalisayo, izinto zabantu ziba yinzuzo ebukhali kakhulu.

Indlela Yokugcina Umsebenzi Wakho Uyisiqinisekiso Sekusasa 🔧

  • Hlela, ungancintisani: Phatha i-AI njengomuntu osebenza naye.

  • Ukubuyekezwa okuphindwe kabili: Ukumodela okusongelayo, imininingwane-njengoba-kuhlolwa, ukubonwa.

  • Funda ukujula kwesizinda: Izinkokhelo, impilo, izindiza, isimo sezulu - umongo uyikho konke.

  • Yakha ithuluzi lomuntu siqu: Ama-Linter, ama-fuzzers, ama-API athayishiwe, izakhiwo eziphindaphindwayo.

  • Izinqumo zedokhumenti: Ama-ADR kanye nohlu lokuhlola lugcina izinguquko ze-AI zilandeleka [4].

Ikusasa Elingenzeka: Ukubambisana, Hhayi Ukushintsha 👫🤖

Isithombe sangempela akusikho ukuthi “i-AI vs. onjiniyela.” Kuyi -AI nonjiniyela. Labo abancika kuyo bazohamba ngokushesha, bacabange kakhulu, futhi bakhiphe umsebenzi wokubhonga. Labo abamelana nayo basengozini yokusala ngemuva.

Ukuhlolwa kweqiniso:

  • Ikhodi yenqubo → i-AI.

  • Isu + izingcingo ezibucayi → Abantu.

  • Imiphumela emihle kakhulu → onjiniyela abathuthukiswe yi-AI [1][2][3].

Ukuyisonga 📝

Ngakho-ke, ingabe onjiniyela bazoshintshwa? Cha. Imisebenzi yabo izoshintshashintsha. Akuyona "indlela yokubhala amakhodi" kodwa "ukubhala amakhodi kuyashintsha." Abawinile kuzoba yilabo abafunda ukuqhuba i -AI, hhayi ukulwa nayo.

Kungumbuso omusha onamandla, hhayi ukushelela okupinki.

Izinkomba

[1] GitHub. “Ucwaningo: ukulinganisa umthelela we-GitHub Copilot ekukhiqizeni nasekujabuleni konjiniyela.” (2022). https://github.blog/news-insights/research/research-quantifying-github-copilots-impact-on-developer-productivity-and-happiness/

[2] McKinsey & Company. “Ukukhulula umkhiqizo wonjiniyela nge-AI ekhiqizayo.” (Juni 27, 2023). https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/unleashing-developer-productivity-with-generative-ai

[3] Ukugcwala Kwezitaki. “Ucwaningo Lonjiniyela Lwango-2025 — I-AI.” (2025). https://survey.stackoverflow.co/2025/ai

[4] I-NIST. “Uhlaka Lokuphathwa Kwengozi ye-AI (i-AI RMF).” (2023–). https://www.nist.gov/itl/ai-risk-management-framework

[5] UPerry, N., uSrivastava, M., Kumar, D., kanye noBoneh, D. “Ingabe Abasebenzisi Babhala Ikhodi Engavikelekile Kakhulu Ngabasizi Be-AI?” I-ACM CCS (2023). https://dl.acm.org/doi/10.1145/3576915.3623157

Thola i-AI Yakamuva Esitolo Esisemthethweni Somsizi we-AI

Mayelana NATHI

Buyela kubhulogi