OpenAI t\u00eb enjten njoftoi GPT-5, modelin e saj m\u00eb t\u00eb fundit dhe m\u00eb t\u00eb p\u00ebrparuar t\u00eb inteligjenc\u00ebs artificiale n\u00eb shkall\u00eb t\u00eb gjer\u00eb.<\/p>\n\n\n\n
Kompania po e b\u00ebn GPT-5 t\u00eb disponuesh\u00ebm p\u00ebr t\u00eb gjith\u00eb, p\u00ebrfshir\u00eb p\u00ebrdoruesit e saj falas. OpenAI tha se modeli \u00ebsht\u00eb m\u00eb i zgjuar, m\u00eb i shpejt\u00eb dhe “shum\u00eb m\u00eb i dobish\u00ebm”, ve\u00e7an\u00ebrisht n\u00eb fusha si shkrimi, kodimi dhe kujdesi sh\u00ebndet\u00ebsor.<\/p>\n\n\n\n
“U p\u00ebrpoqa t\u00eb kthehesha te GPT-4, dhe ishte mjaft i mjeruesh\u00ebm,” tha CEO i OpenAI, Sam Altman, n\u00eb nj\u00eb takim me gazetar\u00ebt.<\/p>\n\n\n\n
Q\u00eb nga lan\u00e7imi i chatbot-it t\u00eb saj t\u00eb inteligjenc\u00ebs artificiale ChatGPT n\u00eb vitin 2022, OpenAI \u00ebsht\u00eb b\u00ebr\u00eb shum\u00eb i p\u00ebrhapur. Kompania tha se pret t\u00eb arrij\u00eb 700 milion\u00eb p\u00ebrdorues aktiv\u00eb javor\u00eb n\u00eb ChatGPT k\u00ebt\u00eb jav\u00eb, dhe \u00ebsht\u00eb n\u00eb bisedime me investitor\u00ebt p\u00ebr nj\u00eb shitje t\u00eb mundshme t\u00eb aksioneve me nj\u00eb vler\u00ebsim prej af\u00ebrsisht 500 miliard\u00eb dollar\u00ebsh, si\u00e7 raportoi m\u00eb par\u00eb CNBC.<\/p>\n\n\n\n
OpenAI tha se shkalla e halucinacioneve t\u00eb GPT-5 \u00ebsht\u00eb m\u00eb e ul\u00ebt, q\u00eb do t\u00eb thot\u00eb se modeli prodhon p\u00ebrgjigje m\u00eb rrall\u00eb. Kompania tha se gjithashtu kreu vler\u00ebsime t\u00eb gjera sigurie gjat\u00eb zhvillimit t\u00eb GPT-5, duke p\u00ebrfshir\u00eb 5,000 or\u00eb testime.<\/p>\n\n\n\n
N\u00eb vend q\u00eb t\u00eb refuzoj\u00eb plot\u00ebsisht t’u p\u00ebrgjigjet pyetjeve t\u00eb p\u00ebrdoruesve n\u00ebse ato jan\u00eb potencialisht t\u00eb rrezikshme, GPT-5 do t\u00eb p\u00ebrdor\u00eb “p\u00ebrfundime t\u00eb sigurta”, tha OpenAI. Kjo do t\u00eb thot\u00eb q\u00eb modeli do t\u00eb jap\u00eb p\u00ebrgjigje t\u00eb nivelit t\u00eb lart\u00eb brenda kufizimeve t\u00eb siguris\u00eb q\u00eb nuk mund t\u00eb p\u00ebrdoren p\u00ebr t\u00eb shkaktuar d\u00ebm.<\/p>\n\n\n\n
“GPT-5 \u00ebsht\u00eb trajnuar p\u00ebr t\u00eb dalluar kur nj\u00eb detyr\u00eb nuk mund t\u00eb p\u00ebrfundoj\u00eb, p\u00ebr t\u00eb shmangur spekulimet dhe p\u00ebr t\u00eb shpjeguar kufizimet m\u00eb qart\u00eb, gj\u00eb q\u00eb zvog\u00eblon pretendimet e pamb\u00ebshtetura krahasuar me modelet e m\u00ebparshme”, tha Michelle Pokrass, nj\u00eb drejtuese pas trajnimit n\u00eb OpenAI.<\/p>\n","protected":false},"excerpt":{"rendered":"
OpenAI t\u00eb enjten njoftoi GPT-5, modelin e saj m\u00eb t\u00eb fundit dhe m\u00eb t\u00eb p\u00ebrparuar t\u00eb inteligjenc\u00ebs artificiale n\u00eb shkall\u00eb t\u00eb gjer\u00eb. Kompania po e b\u00ebn GPT-5 t\u00eb disponuesh\u00ebm p\u00ebr t\u00eb gjith\u00eb, p\u00ebrfshir\u00eb p\u00ebrdoruesit e saj falas. OpenAI tha se modeli \u00ebsht\u00eb m\u00eb i zgjuar, m\u00eb i shpejt\u00eb dhe “shum\u00eb m\u00eb i dobish\u00ebm”, ve\u00e7an\u00ebrisht n\u00eb<\/p>\n","protected":false},"author":4,"featured_media":24301,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[349],"tags":[40],"class_list":{"0":"post-24300","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","7":"category-artificial-intelligence","8":"tag-featured"},"yoast_head":"\n