Kwenzekeni
Abacwaningi abathathu—u-Harsh Jaiswal, u-Mohan Pedhapati, kanye no-Rahul Maini we-Hacktron AI—baphule ngempumelelo izinhlelo zangaphakathi ze-OpenAI phakathi nenkathi yamahora angu-72. Ngokusho kwe-NewsBytes, ithimba lisebenzise i-Anthropic's Claude AI ukusiza ekubhaleni nasekulungiseni amaphutha ekhodi edingekayo ukuze kusetshenziswe ubungozi kungxenyekazi ye-Discourse. Ngokuhlanganisa iphutha lesofthiwe yokucubungula izithombe ngobuthaka obuhlobene nobunikazi, abacwaningi bathole ukufinyelela okungagunyaziwe kuma-akhawunti wesisebenzi ChatGPT kanye ne-Codex, kanye nekhosombe eliyimfihlo le-GitHub.
Ithimba le-Hacktron AI liqale ucwaningo lwalo kusukela kunkundla yomphakathi, ekugcineni lahlonza imbobo yokuphepha kuplathifomu ye-Discourse. Basebenzise ukuba sengozini kwesofthiwe yenkundla yokucubungula izithombe futhi bayibopha ngobuthakathaka obuhlobene nobunikazi obuhlukile ukuze badlule ukufakazela ubuqiniso.
Abacwaningi basebenzise i-Anthropic's Claude AI ukusiza ukuthuthukiswa nokulungisa iphutha lekhodi yabo yokuxhaphaza. Lonke lolu hlelo, kusukela ocwaningweni lokuqala kuya ekwephuleni umthetho ngempumelelo, lwaqedwa ngaphansi kwamahora angama-72.
Ithimba lithole ukufinyelela ezimpahleni zangaphakathi ezibucayi, okuhlanganisa nama-akhawunti ezisebenzi ze-ChatGPT ne-Codex, kanye nekhosombe eliyimfihlo le-GitHub. Ukukhombisa ukwephulwa komthetho ngaphandle kokulimaza, abacwaningi bathumele isicelo sokudonsa esingenabungozi.
Abacwaningi basebenzise imali engaphansi kuka-$3,000 kumathokheni wemodeli ye-AI ngesikhathi sokusebenza. I-OpenAI ngemuva kwalokho ibhekane nobungozi obuhlobene nobunikazi futhi yaklomelisa iqembu ibhonasi ka-$6,500 yeziphazamisi ngokudalula kwabo okuwumthwalo wemfanelo.
Imininingwane yomthombo: newsbytesapp.com ↗
Kungani kubalulekile
Lesi sigameko sibonisa ukusetshenziswa okukabili kwemvelo ye-AI ekhiqizayo ekuvikelekeni kwe-inthanethi, okubonisa ukuthi amamodeli e-AI angasetshenziswa kanjani ngempumelelo ukuze kusheshiswe ukutholakala kobungozi besofthiwe obuyinkimbinkimbi. Ikhono labacwaningi lokusebenzisa izigaba eziningi ngaphansi kwezinsuku ezintathu besebenzisa usizo lwe-AI ligcizelela isimo esishintshayo sosongo lapho imisebenzi yokuvikela evikelayo nehlaselayo yandiswa kakhulu ngamamodeli ezilimi ezinkulu. Ukuhlonzwa ngempumelelo kanye nokubikwa kwalawa maphutha, okuzuzele iqembu inzuzo engu-$6,500 evela ku-OpenAI, kuphinde kufanekise ukusetshenziswa okungokoqobo kwezinhlelo zenzuzo yeziphazamisi ekuvikeleni izindawo zebhizinisi ezihlanganiswe ne-AI.
Ukusetshenziswa kwe-Claude AI ukwenza ngokuzenzakalelayo noma ukusiza ekudalweni kwamaketango okuxhashazwa kumele inguquko ebalulekile yokuthi ucwaningo lwezokuphepha lwenziwa kanjani. Lokhu kukhombisa ukuthi i-AI ingakwazi ukwehlisa umgoqo wokungena ukuze kuhlonzwe ubungozi obuyinkimbinkimbi, obunezinyathelo eziningi kusofthiwe yebhizinisi.
Lesi sigameko sisebenza njengesibonelo sokusebenza kwezinhlelo zenzuzo yeziphazamisi. Ngokugqugquzela izigebengu zesimilo ukuthi zibike ubungozi, i-OpenAI ikwazile ukulungisa iphutha elibalulekile elihlobene nobunikazi ngaphambi kokuthi lisetshenziswe abadlali abanonya.
Umcimbi uqokomisa isidingo sokuthi izinkampani ziphathe ukugebenga okusizwa yi-AI njengento esemthethweni eyingozi, edinga ukuqapha okuyinkimbinkimbi kwezinhlelo zangaphakathi kanye nemijikelezo eqinile yokuchibiyela isofthiwe yomuntu wesithathu njenge-Discourse.
I-Interactive Mechanism: Indlela Esebenza Ngayo Ngempela
Hlola ubuchwepheshe obuyisisekelo ngemuva kwalokhu kuthuthukiswa ngokuhlanganyela.
Why can ethical evaluation not be reduced to one model score?
Ongakubuka ngokulandelayo
Okubalulekile okugxilwe kakhulu ekutheni izinhlangano zizozilungisa kanjani izivumelwano zazo zokuphepha ukuze zivikele ekutholeni kabusha okusizwa yi-AI nokuxhashazwa. Njengoba abacwaningi beqhubeka nokusebenzisa amamodeli afana ne-Claude ukuhlonza ubungozi, izinkampani zingase zidinge ukuthi zenze ukuhlola okuzenzakalelayo okuzenzakalelayo kokuhlola kanye nezindlela zokuqinisekisa ubunikazi. Okwamanje akwaziwa ukuthi i-OpenAI isebenzise ukuvikela okuthile okusha ukuvimbela uchungechunge olufanayo losizo lwe-AI lokuba sengozini, noma uma ezinye izinhlangano zibuyekeza indlela yazo yenzuzo yeziphazamisi ukuze zibhekelele izindlela zocwaningo eziqhutshwa yi-AI.
Izibukeli kufanele ziqaphe ukuthi ingabe ezinye izinkampani ezinkulu ze-AI zibika ukwephulwa okufanayo kosizo lwe-AI, okungase kubonise ukuthambekela okubanzi ezindleleni zocwaningo lwezokuphepha.
Kusazobonakala ukuthi i-OpenAI noma abanye abaholi bemboni bazokhipha yini isiqondiso esithile noma izinhlaka zokuphepha ezibuyekeziwe mayelana nokusetshenziswa kwamathuluzi e-AI ocwaningweni lokuba sengozini.
Umthelela wesikhathi eside ezinhlakeni zohlelo lwenzuzo yeziphazamisi uyindawo ebalulekile ethakaselwayo, ikakhulukazi ukuthi izinkampani zizolungisa yini izinkokhelo noma izimfuneko ukuze ziphendule isivinini esithe xaxa nokusebenza kahle okuhlinzekwa ucwaningo olusizwa yi-AI.