Amaqembu ocwaningo avela eNyuvesi yaseYork naseNyuvesi yaseCalgary ahlaziye okuthunyelwe ku-Reddit okuphathelene nezinkinga ze-AI yokuhlela, athola ukuthi ama-agent e-AI, ngenxa yezimvume ezibanzi kakhulu, abhala phezu kwamafayela, asusa idatha ebalulekile, aze akhiqize nekhodi enonya.
Iqembu locwaningo lasebenzisa i-web crawler laqoqa okuthunyelwe okudumile okungu-3801 okumakwe njengezihlobene namamodeli amakhulu olimi (LLM) phakathi kukaFebhuwari 2023 noMashi 2026, lase likhetha okuthunyelwe okungu-446 okuphathelene nokuphepha kwe-AI yokuhlela, lahlaziya amazwana abasebenzisi angaphezu kuka-6000 ngaphansi kwalezo zikhala.
Ngokwesikhathi, ucwaningo luveza ukuthi inani lokuthunyelwe kwe-Reddit mayelana nezinkinga ze-AI yokuhlela laliphezulu kakhulu cishe ehlobo lika-2025, kanti uJulayi 2025 waba neningi lezinkinga.
Uma kuhlukaniswa ngamathuluzi, ithuluzi okubikwe ngalo izinkinga eziningi kakhulu yi-Cursor, lilandelwa yi-Claude, i-Codex, i-Copilot, i-Windsurf, i-VSCode, i-Replit, i-Cline namanye.
Ishadi lokuhlukanisa izinhlobo zezinkinga libonisa ukuthi i-Cursor ihlangabezane nezinkinga eziningi “zokuphepha ngesikhathi sokusebenza” kanye “nokufinyelela idatha okungagunyaziwe”, okube nomthelela omubi ezindaweni zokukhiqiza. U-Claude yena uhlangabezane nezingozi eziningi “ezihlobene nokuhlanganiswa namathuluzi ezinkampani zangaphandle”. Ulwazi oluhlobene luboniswe kanje:
