{"id":783490,"date":"2026-01-27T22:10:55","date_gmt":"2026-01-27T21:10:55","guid":{"rendered":"https:\/\/www.hashtag.al\/?p=783490"},"modified":"2026-01-27T22:10:55","modified_gmt":"2026-01-27T21:10:55","slug":"kina-mund-te-lere-ne-bisht-perendimin-ne-garen-per-inteligjencen-e-vertete-kumbari-i-ai-se","status":"publish","type":"post","link":"https:\/\/www.hashtag.al\/index.php\/2026\/01\/27\/kina-mund-te-lere-ne-bisht-perendimin-ne-garen-per-inteligjencen-e-vertete-kumbari-i-ai-se\/","title":{"rendered":"\u201cKina mund t\u00eb l\u00ebr\u00eb n\u00eb bisht Per\u00ebndimin n\u00eb gar\u00ebn p\u00ebr inteligjenc\u00ebn e v\u00ebrtet\u00eb\u201d – \u201cKumbari\u201d i AI-s\u00eb","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"

\"\"<\/p>\n

Nj\u00eb nga figurat themelore t\u00eb inteligjenc\u00ebs artificiale \u00ebsht\u00eb larguar n\u00eb m\u00ebnyr\u00eb t\u00eb papritur nga META, duke dh\u00ebn\u00eb nj\u00eb paralajm\u00ebrim t\u00eb ashp\u00ebr se orientimi pothuajse i verb\u00ebr i industris\u00eb teknologjike drejt modeleve gjigante t\u00eb gjuh\u00ebs \u2014 si ChatGPT i OpenAI dhe Gemini i Google \u2014 po e \u00e7on k\u00ebt\u00eb fush\u00eb drejt nj\u00eb ng\u00ebr\u00e7i pa perspektiv\u00eb.<\/span><\/p>\n

Dr. Yann LeCun, fitues i \u00c7mimit Turing dhe ish-drejtues shkencor i AI n\u00eb Meta, u shk\u00ebput nga kompania disa muaj m\u00eb par\u00eb, n\u00eb at\u00eb q\u00eb burime t\u00eb brendshme e p\u00ebrshkruajn\u00eb si nj\u00eb ndarje profesionale t\u00eb fort\u00eb, jehona e s\u00eb cil\u00ebs u ndje nga Kalifornia deri n\u00eb Pekin. Sipas LeCun, p\u00ebrqendrimi aktual te modelet e m\u00ebdha t\u00eb gjuh\u00ebs nuk do t\u00eb jet\u00eb kurr\u00eb i mjaftuesh\u00ebm p\u00ebr t\u00eb krijuar nj\u00eb inteligjenc\u00eb artificiale t\u00eb krahasueshme me at\u00eb njer\u00ebzore.<\/span><\/p>\n

\u201cIndustria \u00ebsht\u00eb b\u00ebr\u00eb \u2018e dehur\u2019 nga LLM-t\u00eb (LLM-pilled),\u201d \u00ebsht\u00eb shprehur LeCun n\u00eb intervista, duke p\u00ebrdorur nj\u00eb term zhargon p\u00ebr t\u00eb p\u00ebrshkruar fiksimin e ngusht\u00eb t\u00eb sektorit. Nd\u00ebrsa Silicon Valley, sipas tij, vuan nga nj\u00eb \u201ckompleks superioriteti\u201d, ai argumenton se kompanit\u00eb inovative kineze mund ta kalojn\u00eb p\u00ebrfundimisht Per\u00ebndimin n\u00eb nd\u00ebrtimin e inteligjenc\u00ebs artificiale t\u00eb v\u00ebrtet\u00eb.<\/span><\/p>\n

Kritikat e LeCun vijn\u00eb n\u00eb nj\u00eb periudh\u00eb delikate p\u00ebr Meta-n. Kompania \u00ebsht\u00eb p\u00ebrballur me reagime negative p\u00ebr modelin Llama 4, p\u00ebr t\u00eb cilin studiuesit theksojn\u00eb se nuk ka p\u00ebrmbushur pritshm\u00ebrit\u00eb n\u00eb raport me investimet shum\u00eb t\u00eb m\u00ebdha t\u00eb b\u00ebra. Si kund\u00ebrp\u00ebrgjigje, CEO Mark Zuckerberg krijoi nj\u00eb laborator t\u00eb ri t\u00eb quajtur \u201csuper-inteligjenc\u00eb\u201d dhe alokoi fonde t\u00eb m\u00ebdha p\u00ebr harduer\u00eb me performanc\u00eb t\u00eb lart\u00eb, kryesisht t\u00eb furnizuar nga Nvidia.<\/span><\/p>\n

LeCun, prej koh\u00ebsh mb\u00ebshtet\u00ebs i zhvillimit me burim t\u00eb hapur (open-source), e konsideron k\u00ebt\u00eb ndryshim strategjik t\u00eb Meta-s si nj\u00eb gabim. Ai thekson se modelet aktuale t\u00eb gjuh\u00ebs jan\u00eb n\u00eb thelb sisteme statistikore q\u00eb parashikojn\u00eb fjal\u00ebn pasuese dhe se u mungon kuptimi themelor i fizik\u00ebs, planifikimit dhe kauzalitetit (shkak-pasoj\u00eb).<\/span><\/p>\n

Pas largimit nga Meta, LeCun themeloi n\u00eb Paris laboratorin AMI Labs (Advanced Machine Intelligence). Ky startup p\u00ebrqendrohet te t\u00eb ashtuquajturat \u201cmodele t\u00eb bot\u00ebs\u201d (world models), nj\u00eb qasje q\u00eb synon t\u00eb imitoj\u00eb m\u00ebnyr\u00ebn se si njer\u00ebzit dhe kafsh\u00ebt m\u00ebsojn\u00eb \u2014 jo duke p\u00ebrpunuar miliarda faqe teksti, por duke v\u00ebzhguar mjedisin dhe duke kuptuar nd\u00ebrveprimet mes objekteve. LeCun thekson se nd\u00ebrsa nj\u00eb model tipik i madh gjuhe k\u00ebrkon ekuivalentin e rreth 400 mij\u00eb viteve leximi njer\u00ebzor p\u00ebr t\u2019u trajnuar, f\u00ebmij\u00ebt e kuptojn\u00eb bot\u00ebn p\u00ebrmes vet\u00ebm mij\u00ebra or\u00ebve p\u00ebrvoj\u00eb vizuale.<\/span><\/p>\n

V\u00ebrejtjet e LeCun p\u00ebr inovacionin kinez p\u00ebrkojn\u00eb me shqet\u00ebsime m\u00eb t\u00eb gjera gjeopolitike, t\u00eb theksuara edhe n\u00eb Forumin Ekonomik Bot\u00ebror t\u00eb fundit n\u00eb Davos, ku analist\u00ebt vun\u00eb n\u00eb dukje se hendeku teknologjik mes Per\u00ebndimit dhe Kin\u00ebs \u00ebsht\u00eb ngushtuar n\u00eb rreth gjasht\u00eb muaj. P\u00ebrpjekjet e SHBA-s\u00eb p\u00ebr t\u2019i kufizuar Kin\u00ebs aksesin n\u00eb \u00e7ipat m\u00eb t\u00eb avancuar t\u00eb Nvidia-s, me synimin p\u00ebr t\u00eb frenuar p\u00ebrparimin e AI-s\u00eb n\u00eb Pekin, n\u00eb praktik\u00eb kan\u00eb nxitur kompanit\u00eb kineze t\u00eb inovojn\u00eb duke kap\u00ebrcyer kufizimet e harduerit.<\/span><\/p>\n

Kompani si DeepSeek kan\u00eb t\u00ebrhequr v\u00ebmendjen me modele si R1, t\u00eb cilat p\u00ebrdorin arkitektura efikase si \u201cmixture-of-experts\u201d (MoE). K\u00ebto sisteme aktivizojn\u00eb vet\u00ebm pjes\u00eb t\u00eb caktuara t\u00eb rrjetit neural p\u00ebr \u00e7do detyr\u00eb, duke arritur shpejt\u00ebsi dhe efikasitet m\u00eb t\u00eb lart\u00eb edhe mbi \u00e7ipa m\u00eb pak t\u00eb avancuar \u2014 nj\u00eb zhvillim q\u00eb ka befasuar disa studiues n\u00eb Silicon Valley, t\u00eb m\u00ebsuar me disponueshm\u00ebri t\u00eb pakufizuar harduerike.<\/span><\/p>\n

Q\u00ebndrimi kritik i LeCun ndaj rrymave dominuese nuk \u00ebsht\u00eb i ri. N\u00eb vitet 1980, kur shumica e komunitetit shkencor i kishte hedhur posht\u00eb rrjetet neurale, ai vazhdoi pun\u00ebn mbi rrjetet neurale konvolucionale (CNN). Rezultatet e tij n\u00eb Bell Labs provuan se kompjuter\u00ebt mund t\u00eb lexonin shkrimin me dor\u00eb n\u00eb \u00e7eqe bankare \u2014 nj\u00eb teknologji q\u00eb m\u00eb pas u b\u00eb themelore p\u00ebr njohjen e fytyr\u00ebs dhe automjetet autonome.<\/span><\/p>\n

Debati i sot\u00ebm pasqyron nj\u00eb ndarje t\u00eb vjet\u00ebr n\u00eb k\u00ebrkimin e AI-s\u00eb: n\u00ebse p\u00ebrparimi do t\u00eb vij\u00eb nga forca brutale llogarit\u00ebse \u2014 m\u00eb shum\u00eb procesor\u00eb dhe m\u00eb shum\u00eb t\u00eb dh\u00ebna \u2014 apo nga arkitektura m\u00eb inteligjente dhe m\u00eb t\u00eb sofistikuara. Udh\u00ebheq\u00ebs t\u00eb Silicon Valley-t, si Demis Hassabis i Google dhe Sam Altman i OpenAI, po p\u00ebrshpejtojn\u00eb nd\u00ebrtimin e infrastrukturave nga frika se mos lihen pas nga zhvillimet kineze. LeCun, megjithat\u00eb, e krahason k\u00ebt\u00eb qasje me nd\u00ebrtimin e nj\u00eb Kulle moderne t\u00eb Babelit, t\u00eb destinuar t\u00eb shembet.<\/span><\/p>\n

N\u00eb Kin\u00eb, zhvillimi i inteligjenc\u00ebs artificiale po orientohet gjithnj\u00eb e m\u00eb shum\u00eb drejt integrimit praktik n\u00eb industri reale, si prodhimi dhe sektori automobilistik, n\u00eb vend t\u00eb kopjimit t\u00eb agjent\u00ebve bisedues n\u00eb stilin e ChatGPT. P\u00ebr shum\u00eb v\u00ebzhgues, fakti q\u00eb LeCun parashikoi revolucionin e \u201cdeep learning\u201d n\u00eb nj\u00eb koh\u00eb kur t\u00eb tjer\u00ebt e konsideronin fantashkenc\u00eb, i b\u00ebn paralajm\u00ebrimet e tij ve\u00e7an\u00ebrisht t\u00eb v\u00ebshtira p\u00ebr t\u2019u shp\u00ebrfillur, shkruan Ynet<\/a><\/span>.<\/span><\/p>\n

Sot, gara p\u00ebr inteligjenc\u00ebn artificiale ka kaluar p\u00ebrtej teknologjis\u00eb: sipas ekspert\u00ebve, ajo \u00ebsht\u00eb shnd\u00ebrruar n\u00eb nj\u00eb betej\u00eb strategjike p\u00ebr t\u00eb ardhmen e udh\u00ebheqjes ekonomike dhe t\u00eb siguris\u00eb globale.<\/span><\/p>\n","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"excerpt":{"rendered":"

Nj\u00eb nga figurat themelore t\u00eb inteligjenc\u00ebs artificiale \u00ebsht\u00eb larguar n\u00eb m\u00ebnyr\u00eb t\u00eb papritur nga META, duke dh\u00ebn\u00eb nj\u00eb paralajm\u00ebrim t\u00eb ashp\u00ebr se orientimi pothuajse i verb\u00ebr i industris\u00eb teknologjike drejt modeleve gjigante t\u00eb gjuh\u00ebs \u2014 si ChatGPT i OpenAI dhe Gemini i Google \u2014 po e \u00e7on k\u00ebt\u00eb fush\u00eb drejt nj\u00eb ng\u00ebr\u00e7i pa perspektiv\u00eb. Dr. […]<\/p>\n","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"author":51,"featured_media":783491,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[53],"tags":[],"class_list":["post-783490","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-fast-foodi"],"gt_translate_keys":[{"key":"link","format":"url"}],"_links":{"self":[{"href":"https:\/\/www.hashtag.al\/index.php\/wp-json\/wp\/v2\/posts\/783490","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.hashtag.al\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.hashtag.al\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.hashtag.al\/index.php\/wp-json\/wp\/v2\/users\/51"}],"replies":[{"embeddable":true,"href":"https:\/\/www.hashtag.al\/index.php\/wp-json\/wp\/v2\/comments?post=783490"}],"version-history":[{"count":1,"href":"https:\/\/www.hashtag.al\/index.php\/wp-json\/wp\/v2\/posts\/783490\/revisions"}],"predecessor-version":[{"id":783492,"href":"https:\/\/www.hashtag.al\/index.php\/wp-json\/wp\/v2\/posts\/783490\/revisions\/783492"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.hashtag.al\/index.php\/wp-json\/wp\/v2\/media\/783491"}],"wp:attachment":[{"href":"https:\/\/www.hashtag.al\/index.php\/wp-json\/wp\/v2\/media?parent=783490"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.hashtag.al\/index.php\/wp-json\/wp\/v2\/categories?post=783490"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.hashtag.al\/index.php\/wp-json\/wp\/v2\/tags?post=783490"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}