{"id":92976,"date":"2026-05-21T16:19:22","date_gmt":"2026-05-21T16:19:22","guid":{"rendered":"https:\/\/www.europesays.com\/cz\/92976\/"},"modified":"2026-05-21T16:19:22","modified_gmt":"2026-05-21T16:19:22","slug":"obecna-umela-inteligence-uz-je-mozna-tady-proc-jen-krcime-rameny","status":"publish","type":"post","link":"https:\/\/www.europesays.com\/cz\/92976\/","title":{"rendered":"Obecn\u00e1 um\u011bl\u00e1 inteligence u\u017e je mo\u017en\u00e1 tady. Pro\u010d jen kr\u010d\u00edme rameny?"},"content":{"rendered":"

\u201eObecn\u00e1 um\u011bl\u00e1 inteligence je te\u010f na obzoru,\u201c prohl\u00e1sil<\/a> v\u00a0\u00fater\u00fd z\u00a0p\u00f3dia v\u00fdvoj\u00e1\u0159sk\u00e9 konference Google I\/O \u0161\u00e9f jej\u00ed AI sekce Demis Hassabis. \u201eStane se tou nejv\u00fdznamn\u011bj\u0161\u00ed technologi\u00ed, jakou kdy \u010dlov\u011bk vynalezl. Pokud ji postav\u00edme dob\u0159e, mohla by posunout pokrok a\u00a0rozkv\u011bt lidstva za hranice na\u0161\u00ed p\u0159edstavivosti.\u201c<\/p>\n

Foto: YouTube\/Google<\/p>\n

\u201eObecn\u00e1 um\u011bl\u00e1 inteligence je na dosah!\u201c T\u00edmhle prohl\u00e1\u0161en\u00edm u\u017e dnes nikoho neohrom\u00edte.<\/p>\n

\u201eTen pokrok je \u00fa\u017easn\u011b rychl\u00fd,\u201c podivoval se nahlas Hassabis, zakladatel firmy DeepMind a\u00a0dr\u017eitel Nobelovy ceny za chemii. \u201eI pro n\u00e1s, kte\u0159\u00ed jsme v\u00fdvoji AI zasv\u011btili cel\u00fd \u017eivot, je to neuv\u011b\u0159iteln\u00e9.\u201c<\/p>\n

\u010ctete uk\u00e1zku z\u00a0newsletteru TechMIX, ve kter\u00e9m Pavel Kas\u00edk a\u00a0Matou\u0161 L\u00e1z\u0148ovsk\u00fd ka\u017edou st\u0159edu p\u0159in\u00e1\u0161ej\u00ed hned n\u011bkolik koment\u00e1\u0159\u016f a\u00a0post\u0159eh\u016f ze sv\u011bta v\u011bdy a\u00a0nov\u00fdch technologi\u00ed. Pokud v\u00e1s TechMIX zaujme, p\u0159ihlaste se k\u00a0jeho odb\u011bru<\/a>!<\/p>\n

\u201eA\u017e se za touto dobou ohl\u00e9dneme, mysl\u00edm, \u017ee si uv\u011bdom\u00edme, \u017ee jsme st\u00e1li na prahu singularity. Bude to hlubok\u00fd okam\u017eik pro lidstvo,\u201c sna\u017eil se znovu Hassabis podtrhnout v\u00fdznam.<\/p>\n

Odezva publika byla ov\u0161em nulov\u00e1. \u017d\u00e1dn\u00fd potlesk nebo \u00fadiv. Pro\u010d tak\u00e9? To je p\u0159ece norm\u00e1lka. V\u011bt\u0161\u00ed potlesk vyvolalo i\u00a0ozn\u00e1men\u00ed, \u017ee nyn\u00ed budete moci zaklapnout v\u00edko notebooku<\/a> a\u00a0AI agent bude d\u00e1l na pozad\u00ed programovat.<\/p>\n

Obecn\u00e1 um\u011bl\u00e1 inteligence (\u010dasto p\u0159ezd\u00edvan\u00e1 zkratkou AGI) nebo siln\u011bj\u0161\u00ed superinteligence (zkratka ASI) jsou dnes pojmy tak b\u011b\u017en\u00e9, \u017ee se nad nimi nikdo nepozastavuje.<\/p>\n

Co se t\u00fd\u010de t\u00e9 AGI, dost mo\u017en\u00e1 u\u017e jsme t\u00e9to mety dos\u00e1hli<\/a> a\u00a0k t\u00e9 superinteligenci \u0161\u00e9fov\u00e9 velk\u00fdch AI laborato\u0159\u00ed zat\u00edm celkem odhodlan\u011b a\u00a0zcela nepokryt\u011b sm\u011b\u0159uj\u00ed<\/a>. V\u00a0Silicon Valley se to te\u010f pova\u017euje za \u00fapln\u011b standardn\u00ed: Po\u010d\u00edtat s\u00a0t\u00edm, \u017ee do n\u011bkolika let tu bude po\u010d\u00edta\u010dov\u00e1 entita, kter\u00e1 je chyt\u0159ej\u0161\u00ed ne\u017e jak\u00fdkoliv \u010dlov\u011bk. Mluv\u00ed se o \u201en\u011b\u017en\u00e9 singularit\u011b\u201c nebo \u201emilionu v\u011bdc\u016f v\u00a0datacentru\u201c, jako by se nechumelilo.<\/p>\n

N\u011bco z\u00a0toho je samoz\u0159ejm\u011b klasick\u00fd trik, kter\u00fd pat\u0159\u00ed do reperto\u00e1ru Silicon Valley u\u017e p\u016fl stolet\u00ed: Fake it till you make it. Nap\u0159ed mus\u00edte o\u00a0t\u00e9 budoucnosti mluvit, jako by tu byla, a\u00a0t\u00edm ji p\u0159ivol\u00e1te. Ob\u010das se to zvrtne<\/a>, kdy\u017e n\u011bkdo sliby p\u0159e\u017eene, a\u00a0pak praskne, \u017ee nebyly ni\u010d\u00edm podlo\u017een\u00e9. Ale u\u00a0um\u011bl\u00e9 inteligence to zat\u00edm nevypad\u00e1 na n\u011bjak\u00e9 prasknut\u00ed nebo n\u00e1raz do zdi. Naopak se zd\u00e1, \u017ee schopnosti stoupaj\u00ed je\u0161t\u011b o\u00a0trochu rychleji, ne\u017e nazna\u010dovaly ty nejoptimisti\u010dt\u011bj\u0161\u00ed predikce (nebo nejpesimisti\u010dt\u011bj\u0161\u00ed, dle \u00fahlu pohledu).<\/p>\n

Z kac\u00ed\u0159sk\u00e9 my\u0161lenky je norma<\/p>\n

Kdy\u017e v\u00a0roce 2015\u00a0Elon Musk a\u00a0Sam Altman zakl\u00e1dali laborato\u0159 OpenAI<\/a>, term\u00edn obecn\u00e1 um\u011bl\u00e1 inteligence p\u016fsobil nejen nev\u011bdecky, ale p\u0159\u00edmo sm\u011b\u0161n\u011b. Ilya Sutskever, jeden z\u00a0nad\u011bjn\u00fdch v\u00fdvoj\u00e1\u0159\u016f AI, se tehdy sv\u011b\u0159il, \u017ee OpenAI se pro n\u011bj stalo \u00fato\u010di\u0161t\u011bm pr\u00e1v\u011b proto, \u017ee tam zkratku AGI brali v\u00e1\u017en\u011b. Jinde, mezi \u201eskute\u010dn\u00fdmi v\u00fdzkumn\u00edky strojov\u00e9ho u\u010den\u00ed\u201c, se takov\u00fdmi sci-fi nesmysly moc nezab\u00fdvali.<\/p>\n

OpenAI se do v\u00fdzkumu AGI pustila, ale definovat ji neum\u011bli. I\u00a0s\u00e1m Sutskever o\u00a0n\u00ed mluvil \u010dasto sp\u00ed\u0161 jako guru a\u00a0mystik ne\u017e jako v\u011bdec. \u201eV Silicon Valley se vyvinulo n\u011bco, \u010demu \u0159\u00edk\u00e1m kvazin\u00e1bo\u017eensk\u00e9 hnut\u00ed,\u201c v\u0161imla si<\/a> novin\u00e1\u0159ka Karen Hao, kter\u00e1 v\u00a0roce 2019\u00a0jako prvn\u00ed napsala d\u016fkladn\u00fd profil firmy OpenAI. \u201ePojem obecn\u00e9 um\u011bl\u00e9 inteligence nen\u00ed v\u011bdecky podlo\u017een\u00fd.\u201c<\/p>\n

<\/a><\/p>\n

\u201eV tomto \u201aAGI n\u00e1bo\u017eenstv\u00ed\u2018 existuj\u00ed dv\u011b frakce. Jedni si mysl\u00ed, \u017ee n\u00e1m AGI p\u0159inese utopii, druz\u00ed si mysl\u00ed, \u017ee zni\u010d\u00ed cel\u00e9 lidstvo. Ob\u011b skupiny ale v\u011b\u0159\u00ed, \u017ee to p\u0159ijde brzy, a\u00a0tud\u00ed\u017e doch\u00e1z\u00ed ke stejn\u00e9mu z\u00e1v\u011bru: Mus\u00edme b\u00fdt t\u011bmi, kdo tu technologii ovl\u00e1dnou, a\u00a0nesm\u00edme ji uvolnit mezi ostatn\u00ed,\u201c dod\u00e1v\u00e1.<\/p>\n

Hao si ve sv\u00e9 (mimochodem velmi zda\u0159il\u00e9) knize Imp\u00e9rium um\u011bl\u00e9 inteligence v\u0161\u00edm\u00e1, \u017ee Altman i\u00a0dal\u0161\u00ed \u0161\u00e9fov\u00e9 um\u00ed s\u00a0t\u00edmto term\u00ednem obratn\u011b pracovat. Kdy\u017e prezentuj\u00ed svou vizi Microsoftu, pou\u017e\u00edvaj\u00ed jin\u00e9 metriky, ne\u017e kdy\u017e stejnou vizi prezentuj\u00ed investor\u016fm, zam\u011bstnanc\u016fm nebo ve\u0159ejnosti. Pro ka\u017ed\u00e9ho nabere zkratka jin\u00fd v\u00fdznam. I\u00a0to je asi d\u016fvod, pro\u010d se firmy OpenAI a\u00a0Microsoft nikdy nedohodly na form\u00e1ln\u00ed definici momentu, kdy AGI nastane.<\/p>\n

Z p\u016fl stolet\u00ed jsou dva roky<\/p>\n

Schv\u00e1ln\u011b jsem se pod\u00edval na to, jak o\u00a0\u201eobecn\u00e9 AI\u201c diskutovali v\u011bdci specializuj\u00edc\u00ed se na strojov\u00e9 u\u010den\u00ed p\u0159ed deseti lety. V\u00a0roce 2016\u00a0administrativa prezidenta Baracka Obamy sestavila anal\u00fdzu<\/a> ohledn\u011b predikc\u00ed o\u00a0mo\u017enostech um\u011bl\u00e9 inteligence: \u201eSou\u010dasn\u00fd konsenzus expertn\u00ed komunity, se kterou v\u00a0tomto souhlas\u00ed i\u00a0vl\u00e1dn\u00ed experti, je, \u017ee obecn\u00e1 um\u011bl\u00e1 inteligence nebude dosa\u017eiteln\u00e1 je\u0161t\u011b minim\u00e1ln\u011b n\u011bkolik desetilet\u00ed.\u201c<\/p>\n

Predik\u010dn\u00ed platforma Metacalculus<\/a> odhadovala v\u00a0roce 2020\u00a0n\u00e1stup \u201eslab\u00e9 obecn\u00e9 inteligence\u201c na rok 2055.\u00a0\u0160lo o\u00a0medi\u00e1n odhad\u016f\u00a0\u2013 polovina lid\u00ed tipovala, \u017ee to bude trvat je\u0161t\u011b d\u00e9le nebo \u017ee takov\u00e9ho momentu nebude dosa\u017eeno do konce stolet\u00ed.<\/p>\n

Aktu\u00e1ln\u00ed medi\u00e1nov\u00e1 predikce na stejn\u00e9 platform\u011b mluv\u00ed o\u00a0dubnu 2028, tedy za dva roky. Experti \u010dasto mluv\u00ed o\u00a0kolapsu \u010dasov\u00fdch odhad\u016f<\/a>. Za zm\u00ednku stoj\u00ed, \u017ee ty odhady, kter\u00e9 v\u00a0roce 2021\u00a0zn\u011bly jako nejv\u00edce p\u0159ita\u017een\u00e9 za vlasy, se ve zp\u011btn\u00e9m pohledu ukazuj\u00ed jako nejp\u0159esn\u011bj\u0161\u00ed.<\/p>\n

Dobr\u00fdm p\u0159\u00edkladem jsou predikce Daniela Kokotajla, uzn\u00e1van\u00e9ho v\u00fdzkumn\u00edka a\u00a0filozofa AI, kter\u00fd v\u00a0roce 2024\u00a0ode\u0161el z\u00a0OpenAI. Ten v\u00a0roce 2021\u00a0napsal svou esej: Jak bude vypadat rok 2026?<\/a><\/p>\n

\u201eV roce 2026\u00a0budou chatboty pou\u017e\u00edvat stamiliony lid\u00ed, budou s\u00a0nimi \u0159e\u0161it r\u016fzn\u00e9 v\u011bci, praktick\u00e9, ale i\u00a0vztahov\u00e9. \u0158ada lid\u00ed bude br\u00e1t chatboty jako sv\u00e9ho druhu kamar\u00e1dy.\u201c Naprost\u00e1 trefa, nejv\u011bt\u0161\u00ed chatboti ChatGPT a\u00a0Gemini maj\u00ed t\u00e9m\u011b\u0159 miliardu aktivn\u00edch u\u017eivatel\u016f. Trefil se<\/a> i\u00a0v dal\u0161\u00edch aspektech, mnoh\u00e9 z\u00a0nich jsou pom\u011brn\u011b technick\u00e9.\u201eKdy\u017e jsem predikci psal, sna\u017eil jsem se domyslet, jak by r\u016fzn\u00e9 p\u0159\u00edb\u011bhy mohly v\u011brohodn\u011b skon\u010dit,\u201c popsal<\/a> svou metodiku v\u00fdzkumn\u00edk. \u201eVy\u0161lo to, mysl\u00edm, l\u00e9pe, ne\u017e jsem \u010dekal. Realita se bl\u00ed\u017e\u00ed t\u00e9to fikci v\u00edce, ne\u017e jsem \u010dekal, a\u00a0rozhodn\u011b v\u00edce, ne\u017e \u010dekala v\u011bt\u0161ina ostatn\u00edch lid\u00ed.\u201c<\/p>\n

Schopnost automaticky neznamen\u00e1 u\u017eite\u010dnost<\/p>\n

Spr\u00e1vn\u011b tak\u00e9 Kokotajlo odhadl n\u011bco, co \u0159ad\u011b lid\u00ed st\u00e1le ani dnes nedoch\u00e1z\u00ed. \u017de vysok\u00e9 schopnosti um\u011bl\u00e9 inteligence samy o\u00a0sob\u011b neznamenaj\u00ed automatickou zm\u011bnu cel\u00e9 lidsk\u00e9 spole\u010dnosti. P\u0159edv\u00eddal tehdy nap\u0159\u00edklad, \u017ee dopad um\u011bl\u00e9 inteligence na produktivitu (m\u011b\u0159enou klasick\u00fdm HDP) bude v\u00a0roce 2026\u00a0nulov\u00fd. A\u00a0i kdy\u017e \u00fapln\u00e1 nula to z\u0159ejm\u011b nen\u00ed, je pravda, \u017ee dopad AI na re\u00e1ln\u011b m\u011b\u0159itelnou produktivitu cel\u00fdch st\u00e1t\u016f je sp\u00ed\u0161e mal\u00fd.<\/p>\n

Jedn\u00edm z\u00a0d\u016fvod\u016f m\u016f\u017ee b\u00fdt, \u017ee sou\u010dasn\u00e1 st\u00e1tn\u00ed, obchodn\u00ed i\u00a0firemn\u00ed infrastruktura je\u00a0\u2013 a\u00a0bu\u010fme za to r\u00e1di\u00a0\u2013 optimalizov\u00e1na pro lidi, pro lidsk\u00e9 vn\u00edm\u00e1n\u00ed, lidsk\u00e9 chyby a\u00a0lidsk\u00e9 schopnosti. Jak u\u017e jsem psal<\/a> v\u00a0lo\u0148sk\u00e9m \u010dl\u00e1nku: \u201eFirmy na novou situaci nedok\u00e1\u017eou rychle nasko\u010dit. Nasazov\u00e1n\u00ed rozs\u00e1hl\u00fdch AI automatizac\u00ed, kter\u00e9 by nahradily cel\u00e9 odd\u011blen\u00ed, je zna\u010dn\u011b nejist\u00e9 a\u00a0rizikov\u00e9. A\u00a0tak se AI pou\u017e\u00edv\u00e1 hlavn\u011b jako podp\u016frn\u00fd prost\u0159edek, ale st\u00e1le funguje v\u00a0infrastruktu\u0159e, kter\u00e1 je d\u011blan\u00e1 pro lidi.\u201c<\/p>\n

Jak si v\u0161iml moder\u00e1tor Dwarkesh Patel<\/a>: \u201eV\u00fdkonnost model\u016f stoup\u00e1 tempem, jak\u00e9 p\u0159edpov\u00eddaj\u00ed zast\u00e1nci rychl\u00fdch \u010dasov\u00fdch odhad\u016f. Ale u\u017eite\u010dnost roste tempem, jak\u00e9 p\u0159edpov\u00eddaj\u00ed zast\u00e1nci pomal\u00fdch \u010dasov\u00fdch odhad\u016f.\u201c<\/p>\n

<\/a><\/p>\n

Znamen\u00e1 to, \u017ee firmy nyn\u00ed masivn\u011b nasazuj\u00ed um\u011blou inteligenci (a utr\u00e1cej\u00ed za tyto tokeny miliardy dolar\u016f), ale jen minimum t\u011bchto token\u016f je sp\u00e1leno smyslupln\u011b. V\u011bt\u0161ina chatbot\u016f je nyn\u00ed nasazena k\u00a0tomu, aby \u201espolupracovala\u201c na materi\u00e1lech, kter\u00e9 jsou ur\u010den\u00e9 pro ty star\u00e9, na lidi zam\u011b\u0159en\u00e9 struktury.<\/p>\n

N\u011bkdo pou\u017eije chatbota, aby pomohl p\u0159ipravit prezentaci nebo napsat e-mail, p\u0159\u00edpadn\u011b naprogramovat n\u011bjak\u00e9 intern\u00ed n\u00e1stroje, vygenerovat obr\u00e1zky pro reklamu na soci\u00e1ln\u00ed s\u00edt\u011b atd. To jsou v\u0161echno \u010dinnosti, kter\u00e9 se neprojev\u00ed na celkov\u00e9 produktivit\u011b.<\/p>\n

\u0158ada firem nav\u00edc po\u0159\u00e1dn\u011b nev\u00ed<\/a>, jak\u00fd je jejich c\u00edl nebo strategie r\u016fstu. Dosud to toti\u017e nemusely moc \u0159e\u0161it. Tak\u017ee kdy\u017e dostanou nov\u00e9 n\u00e1stroje, neum\u00ed je nasadit k\u00a0ni\u010demu jin\u00e9mu ne\u017e k\u00a0tomu, co d\u011blaly dosud. Pokud d\u0159\u00edve chodily v\u00a0kruz\u00edch, tak nyn\u00ed jezd\u00ed ve stejn\u00fdch kruz\u00edch na motorce. Je to intenzivn\u011bj\u0161\u00ed<\/a> tempo, ale re\u00e1ln\u00e1 produktivita nestoup\u00e1.<\/p>\n

Mohli bychom to br\u00e1t jako uchl\u00e1cholen\u00ed: N\u00e1stup AI nakonec a\u017e tolik v\u011bc\u00ed nezm\u011bn\u00ed. Firmy si budou v\u011bci d\u011blat d\u00e1l s\u00a0pomoc\u00ed lidsk\u00fdch zam\u011bstnanc\u016f a\u00a0AI bude jen tu a\u00a0tam p\u0159icmrnd\u00e1vat. Ale to by podle m\u00e9ho byl velmi kr\u00e1tkozrak\u00fd pohled. <\/p>\n

Optimalizace struktur na lidsk\u00e9 procesy n\u00e1m d\u00e1v\u00e1 \u010das nav\u00edc. Nem\u011bli bychom jej ale promrhat. Stoj\u00ed za p\u0159ipomenut\u00ed, \u017ee schopnosti AI agent\u016f rostou exponenci\u00e1ln\u011b<\/a>. A\u00a0kdykoli n\u011bco roste exponenci\u00e1ln\u011b, nesm\u00edme se spol\u00e9hat na svou intuici.<\/p>\n

V pln\u00e9 verzi newsletteru TechMIX toho najdete je\u0161t\u011b mnohem v\u00edc. P\u0159ihlaste se k\u00a0odb\u011bru<\/a> a\u00a0budete ho dost\u00e1vat ka\u017edou st\u0159edu p\u0159\u00edmo do sv\u00e9 e-mailov\u00e9 schr\u00e1nky.<\/p>\n","protected":false},"excerpt":{"rendered":"\u201eObecn\u00e1 um\u011bl\u00e1 inteligence je te\u010f na obzoru,\u201c prohl\u00e1sil v\u00a0\u00fater\u00fd z\u00a0p\u00f3dia v\u00fdvoj\u00e1\u0159sk\u00e9 konference Google I\/O \u0161\u00e9f jej\u00ed AI sekce…\n","protected":false},"author":2,"featured_media":92977,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":"","_share_on_mastodon":"0"},"categories":[10],"tags":[3120,17,21,6563,16,6223,10596,2232,18,19,22,126,20],"class_list":["post-92976","post","type-post","status-publish","format-standard","has-post-thumbnail","category-zpravy","tag-chatgpt","tag-headlines","tag-hlavni-udalosti","tag-llm-velke-jazykove-modely","tag-news","tag-obecna-umela-inteligence-agi","tag-prognoza","tag-techmix","tag-top-news","tag-top-stories","tag-udalosti-dne","tag-umela-inteligence-ai","tag-zpravy"],"share_on_mastodon":{"url":"https:\/\/pubeurope.com\/@cz\/116613478380853775","error":""},"_links":{"self":[{"href":"https:\/\/www.europesays.com\/cz\/wp-json\/wp\/v2\/posts\/92976","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/cz\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/cz\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/cz\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/cz\/wp-json\/wp\/v2\/comments?post=92976"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/cz\/wp-json\/wp\/v2\/posts\/92976\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/cz\/wp-json\/wp\/v2\/media\/92977"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/cz\/wp-json\/wp\/v2\/media?parent=92976"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/cz\/wp-json\/wp\/v2\/categories?post=92976"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/cz\/wp-json\/wp\/v2\/tags?post=92976"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}