{"id":166425,"date":"2026-03-26T17:13:13","date_gmt":"2026-03-26T15:13:13","guid":{"rendered":"https:\/\/prohoster.info\/blog\/novosti-interneta\/intervyu-s-gregom-kroa-hartmanom-ob-otchyotah-ob-oshibkah-sgenerirovannyh-ai"},"modified":"2026-03-26T17:13:13","modified_gmt":"2026-03-26T15:13:13","slug":"intervyu-s-gregom-kroa-hartmanom-ob-otchyotah-ob-oshibkah-sgenerirovannyh-ai","status":"publish","type":"post","link":"https:\/\/prohoster.info\/sq\/blog\/novosti-interneta\/intervyu-s-gregom-kroa-hartmanom-ob-otchyotah-ob-oshibkah-sgenerirovannyh-ai","title":{"rendered":"Intervista me Greg Kroah-Hartman mbi raportet p\u00ebr gabimet, t\u00eb gjeneruara nga AI","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"
N\u00eb konferenc\u00ebn KubeCon Europe, edicioni The Register intervistoi Greg Kroah-Hartman, i cili \u00ebsht\u00eb p\u00ebrgjegj\u00ebs p\u00ebr mb\u00ebshtetje t\u00eb deg\u00ebve stabile dhe \u00abstaging\u00bb t\u00eb b\u00ebrtham\u00ebs Linux, duke mbajtur postin e mbajt\u00ebsit n\u00eb 16 n\u00ebn sistemi t\u00eb b\u00ebrtham\u00ebs. N\u00eb intervist\u00eb diskutohet q\u00ebndrimi i Gregut ndaj raporteve t\u00eb gabimeve q\u00eb d\u00ebrgon AI. AI n\u00eb b\u00ebrtham\u00eb tashm\u00eb p\u00ebrdoret gjat\u00eb rishikimit t\u00eb ndryshimeve n\u00eb n\u00ebn sistemin rrjet\u00ebsor, eBPF dhe DRM, plus se koh\u00ebt e fundit u integrua nj\u00eb mjet i quajtur Sashiko nga Google p\u00ebr t\u00eb verifikuar ndryshimet e dor\u00ebzuara. <\/p>\n
Disa nga deklaratat e Gregut: <\/p>\n
\n
\u00abDisa muaj m\u00eb par\u00eb merrnim at\u00eb q\u00eb e quanim AI-debris \u2014 raporte p\u00ebr sigurin\u00eb t\u00eb gjeneruara nga AI, t\u00eb cilat ishin duksh\u00ebm t\u00eb pasakta ose t\u00eb ul\u00ebta n\u00eb cil\u00ebsi. Kjo ishte madje e qeshur. Nuk shqet\u00ebsoheshim shum\u00eb p\u00ebr k\u00ebt\u00eb... Nj\u00eb muaj m\u00eb par\u00eb ndodhi di\u00e7ka, dhe situata u ndryshua ndjesh\u00ebm. Tani po marrim raporte reale.\u00bb\n
\u00abNj\u00eb situat\u00eb e ngjashme po v\u00ebrehet jo vet\u00ebm n\u00eb Linux \u2014 t\u00eb gjitha projektet me kod t\u00eb hapur po marrin raporte reale, t\u00eb p\u00ebrgatitura me ndihm\u00ebn e AI, dhe tani jan\u00eb cil\u00ebsor\u00eb dhe t\u00eb vlefsh\u00ebm. Ekipet q\u00eb jan\u00eb p\u00ebrgjegj\u00ebse p\u00ebr sigurin\u00eb n\u00eb projektet m\u00eb t\u00eb m\u00ebdha t\u00eb hapura po raportojn\u00eb nj\u00eb tendenc\u00eb t\u00eb nj\u00ebjt\u00eb gjat\u00eb bisedave t\u00eb informuara mes tyre.\u00bb\n
Kur u pyet nga e ka shkaktuar k\u00ebt\u00eb, Gregu u p\u00ebrgjigj: \u00abNuk e dim\u00eb. Duket se askush nuk e di. Ose shum\u00eb mjete jan\u00eb b\u00ebr\u00eb shum\u00eb m\u00eb t\u00eb mira, ose njer\u00ebzit filluan t\u00eb thon\u00eb: \u2018Hej, le t\u00eb merremi me k\u00ebt\u00eb.\u2019 Duket se kjo i p\u00ebrket shum\u00eb grupeve dhe kompanive t\u00eb ndryshme. Sa i p\u00ebrket b\u00ebrtham\u00ebs, ne do ta menaxhojm\u00eb. Ekipi yn\u00eb \u00ebsht\u00eb b\u00ebr\u00eb shum\u00eb m\u00eb i madh, \u00ebsht\u00eb shum\u00eb i shp\u00ebrndara n\u00eb mbar\u00eb bot\u00ebn, dhe rritja jon\u00eb \u00ebsht\u00eb nj\u00eb realitet, dhe nuk po ngadal\u00ebsohet. K\u00ebto jan\u00eb gj\u00ebra t\u00eb vogla, asgj\u00eb serioze, por ndihma n\u00eb k\u00ebt\u00eb \u00e7\u00ebshtje mund t\u00eb nevojitet p\u00ebr t\u00eb gjitha projektet e hapura. N\u00eb projektet e vogla ka shum\u00eb m\u00eb pak mund\u00ebsi p\u00ebr t'u p\u00ebrballur me nj\u00eb fluks t\u00eb papritur t\u00eb raporteve p\u00ebr gabime dhe vulnerabilitetet e gjeneruara nga AI, t\u00eb cilat p\u00ebrmendin gabime reale, e jo mbetje.\u00bb\n
Gregu tregoi se kur i k\u00ebrkoi AI t\u00eb gjente gabime nga nj\u00eb grup ndryshimesh t\u00eb propozuara \u2014 ai gjeti 60 dhe ofroi patches p\u00ebr t'i ndrequr ato. Vet\u00ebm nj\u00eb e treta e gabimeve t\u00eb gjetura ishin t\u00eb tilla, dhe nga patches vet\u00ebm 2\/3 ishin zgjidhje t\u00eb sakta q\u00eb nuk k\u00ebrkonin rregullime, por gjithsesi, kjo nuk ishte aspak e padobishme. Sipas Gregut, mbajt\u00ebsit nuk mund ta injorojn\u00eb k\u00ebt\u00eb, sidomos pasi rezultatet e pun\u00ebs s\u00eb AI po b\u00ebhen gjithnj\u00eb e m\u00eb t\u00eb mira. P\u00ebr t\u00eb sh\u00ebnuar patches t\u00eb krijuara me p\u00ebrdorimin e AI, u shtua etiketa \u00abCo-developed:\u00bb. Megjithat\u00eb, p\u00ebrkund\u00ebr p\u00ebrpjekjeve t\u00eb ve\u00e7anta p\u00ebr t\u00eb p\u00ebrdorur AI p\u00ebr t\u00eb krijuar funksionalitet t\u00eb ri, kryesisht AI n\u00eb b\u00ebrtham\u00eb p\u00ebrdoret p\u00ebr rishikimin e ndryshimeve.\u00bb\n
Nj\u00eb nga p\u00ebrfitimet m\u00eb t\u00eb dukshme t\u00eb AI \u00ebsht\u00eb e quajtur reduktimi i koh\u00ebs s\u00eb p\u00ebrpunimit t\u00eb patches. Kur asistenti AI identifikon probleme evidente, autor\u00ebt e patches marrin feedback shum\u00eb m\u00eb shpejt nga sa do t\u00eb kishte arritur t\u00eb b\u00ebnte mbajt\u00ebsi i v\u00ebrtet\u00eb: \u00abN\u00ebse shoh se sistemi reagon ndaj di\u00e7kaje, ai i jep feedback autorit m\u00eb shpejt se sa do e b\u00ebnte mbajt\u00ebsi, dhe kjo \u00ebsht\u00eb e shk\u00eblqyer. Tani kemi disa bota q\u00eb kontrollojn\u00eb patches. N\u00ebse v\u00eb re se ata japin nj\u00eb gabim, e kuptoj menj\u00ebher\u00eb se si mbajt\u00ebs nuk kam nevoj\u00eb ta shikoj. Dhe zhvilluesi mendon \u2018Oh, mund t\u00eb b\u00ebj nj\u00eb version tjet\u00ebr nes\u00ebr\u2019, q\u00eb ndihmon t\u00eb p\u00ebrmir\u00ebsojm\u00eb pak feedback-un.\u00bb <\/ul>\n