🧒 I-OpenAI ikhipha ukubikezela iminyaka ku-ChatGPT ↗
I-OpenAI isiqalile ukusunduza ungqimba "lokubikezela ubudala" ngaphakathi kwe-ChatGPT - uhlelo oluthule oluzama ukunquma ukuthi i-akhawunti kungenzeka ukuthi ingeyomntwana osemncane, bese lungena ezindleleni zokuphepha ezengeziwe. Uma icabanga ukuthi ungaphansi kweminyaka engu-18, inciphisa ngokuzenzakalelayo ukufinyelela kokuqukethwe okubucayi kakhulu.
Uma iqagela kabi (okuyinto... ezokwenza, ngezinye izikhathi), ikhambi ukuqinisekiswa kobunikazi nge-Persona, kusetshenziswa ukugeleza okusekelwe ku-selfie. Iphinde ifinyelele ezinguqukweni ezibanzi zesikhulumi njengokuhlolwa "kwemodi yabantu abadala" kanye nezinyathelo zokwenza imali ezizwakala, empeleni, njengezitsha ezimbalwa ezijikelezayo ngesikhathi esisodwa.
🏦 IBrithani idinga 'izivivinyo zokucindezeleka ze-AI' zezinsizakalo zezezimali, kusho abenzi bomthetho ↗
Abameli bezomthetho base-UK bathi ngokuyisisekelo: yekani ukungazinaki nge-AI kwezezimali. Impikiswano yabo iwukuthi abalawuli kanye nohulumeni kufanele bahlole ingcindezi yezinhlelo ze-AI njengamabhange bahlolwe ingcindezi - ngoba "hawu" ezweni lemali kuvame ukusabalala ngokushesha.
Ukukhathazeka uhlobo lwenkinga oluziveza kuphela ngobukhulu: ukucwasa, ukuvela kwamaphutha amaningi, izindlela zokukhwabanisa, ukusabela okuhlukahlukene. Kufana nokuvumela ingane encane ishayele ibhasi ngoba umgwaqo uqondile... kuze kube yilapho kungazelelwe ingabi njalo.
🏛️ Abanqobi be-AI baqokwe ukusiza iDolobha ukuthi lisebenzise ngokuphephile amathuba e-AI ↗
Uhulumeni wase-UK uqoke “Ompetha Bobuchwepheshe Bokwenza Izinto Nge-AI” ababili ukusiza ukusheshisa ukwamukelwa kwe-AI ezinkonzweni zezimali - ngezithembiso ezivamile zokukhula, ukukhiqiza, ukutshalwa kwezimali, konke lokho okuzwakala kukuhle.
Okuphawulekayo isenzo sokulinganisela: amandla “okuhamba ngokushesha”, agxivizwe “ngokuphephile” phezulu. Okuqinisekisayo… noma okuhlekisayo kancane, kuye ngesimo sakho sengqondo ngalolo suku.
💰 I-Humans&, inkampani entsha ye-AI 'egxile kubantu' eyasungulwa yi-Anthropic, i-xAI, ne-Google alumni, iqoqe imali engu-$480M nxazonke ↗
Abantu kanye nabafika bedlala ngochungechunge olukhulu kanye nethimba elifundeka njengo-“AI lab starter pack” - abantu abavela ezitolo ezinkulu zamamodeli kanye nobuchwepheshe obukhulu. Inhloso eshiwoyo ukuthi ayidingi i-chatbot-y kodwa “isiza abantu ukuthi baxhumane futhi basebenzisane,” ngezinhlelo ezingasebenza kubantu kanye nama-ejenti e-AI.
Kuyimali eningi ngento esasekusenesikhathi, okuyisibonakaliso sokuthi imakethe iyaqiniseka… noma ukuthi sonke sithenga iziqhumane ngoba siyayithanda imisindo. Eqinisweni, kungaba kokubili.
⚖️ I-Ivo, inkampani entsha ye-AI yezomthetho, iqoqe imali eyizigidi ezingama-55 zamaRandi emzuliswaneni wakamuva wokuxhasa ngezimali ↗
I-Ivo, inkampani entsha ye-AI yezomthetho, ikhiphe imali entsha yokwakha ipulatifomu yayo nokwandisa ukuthengisa, idlula inqwaba yezinkampani zomthetho kanye namaqembu ezomthetho angaphakathi enza imisebenzi eminingi yamaphepha ngokuzenzakalela.
Lesi sigaba siyaqhubeka nokuwina ngoba inani licacile futhi lilula: amahora ambalwa asetshenziswa emsebenzini wokubhala ophindaphindwayo, izinto ezimbalwa ezidlula emaphusheni. Akuyona into ekhangayo, kodwa ikhokhiswa kakhulu.
🧠 I-Inventec yaseTaiwan ithi isinqumo nge-chip ye-H200 yeNvidia 'sibonakala sibambekile' ohlangothini lwaseShayina ↗
Umenzi weseva waseTaiwan uthe udaba lokuthi iNvidia ithengise i-chip yayo ye-H200 AI eShayina lubonakala “lubambekile” - okusho ukuthi isithiyo akuyona nje imithetho yangaphandle, kodwa nokuvunyelwa yiShayina noma amandla angaphakathi.
Leli iqiniso elingathandeki lehadiwe ye-AI: wonke umuntu ukhuluma ngamamodeli nama-ejenti, kodwa ikusasa lonke lingancika ezintweni ezimbalwa ezibangela ukungezwani, izimvume, kanye nezepolitiki ze-supply-chain. Kufana nokuhlela ukwethulwa kwerokhethi... bese uqaphela ukuthi ukhohlwe ukhiye we-valve kaphethiloli.
Imibuzo Evame Ukubuzwa
Kuyini ukubikezela kweminyaka kwe-OpenAI ku-ChatGPT, futhi yini eshintshayo uma uqashelwa njengongaphansi kweminyaka engu-18?
I-OpenAI yethule ungqimba "lokubikezela ubudala" olungabonakali oluzama ukunquma ukuthi i-akhawunti kungenzeka ukuthi ingeyomntwana ongaphansi kweminyaka engu-18. Uma uhlelo luphetha ngokuthi ungaphansi kweminyaka engu-18, i-ChatGPT isebenzisa ngokuzenzakalelayo izindlela zokuphepha ezengeziwe. Empeleni, lokho kuvame ukusho imikhawulo eqinile kokuqukethwe okubhekwa njengokubucayi kakhulu. Inhloso ukunciphisa ubungozi ngaphandle kokuphoqa wonke umsebenzisi ukuthi aveze ubudala bakhe kusengaphambili.
Kwenzekani uma ukubikezela kweminyaka kwe-ChatGPT kuphutha?
Uma ukubikezela iminyaka ku-ChatGPT kuphazamisa futhi kubize umuntu omdala njengongaphansi kweminyaka engu-18, ikhambi elishiwoyo ukuqinisekiswa kobunikazi. Kulokhu kukhishwa, lokho kuqinisekiswa kuphathwa nge-Persona kusetshenziswa ukugeleza okusekelwe ku-selfie. Ngemuva kokuqinisekiswa, i-akhawunti ingabuyela ekufinyeleleni okuvamile kunokuhlala ivaliwe ngemuva kwezithiyo eziqinile. Kuyi-tradeoff: amacala ambalwa onqenqemeni oluyingozi, kanye nokungqubuzana okwengeziwe lapho uhlelo luqagela ngokungalungile.
Ukuqinisekiswa kwe-selfie ye-Persona kuhambisana kanjani nendlela yokuphepha ye-OpenAI?
I-Persona isebenza njengendlela yokuxazulula ukungaqiniseki okuhlobene nobudala lapho ukuphetha kwesistimu kuthathwa njengokungathembeki. Ukuhamba kokuqinisekiswa kobunikazi okusekelwe ku-selfie kubekwa njengesinyathelo sokulungisa ngemuva kokubikezela iminyaka okungalungile. Ezinkundleni eziningi, lolu hlobo lwesinyathelo lusetshenziselwa ukulinganisela ukuphepha nokufinyelela ngaphandle kokuvimbela unomphela abasebenzisi abasemthethweni. Kuphinde kubhekisele ekuphoqelelweni okwengeziwe kwezinga le-akhawunti, hhayi nje ekuhlungeni okusheshayo.
Kungani i-OpenAI yengeza ukuguga manje, futhi ihlobene kanjani "nemodi yabantu abadala"?
Lesi sibuyekezo sifika kanye nokushintsha kwepulatifomu ebanzi, okuhlanganisa ukuhlolwa "kwemodi yabantu abadala" kanye nezinyathelo zokwenza imali okukhulunywe ngazo engxoxweni efanayo. Ukubikezela ubudala ku-ChatGPT kuba yisishintshi esinquma ukuthi ubani othola okuzenzakalelayo okuqinile uma kuqhathaniswa nokufinyelela okubanzi. Lokho kuyasiza epulatifomu ehlola okuhlangenwe nakho okuhlukile kuzo zonke izithameli ezahlukene. Kunganciphisa futhi ukuchayeka kokuqukethwe okuyingozi ngenkathi kuvumela ukunwetshwa kwenye indawo.
Ziyini "izivivinyo zokucindezeleka ze-AI" zezinsizakalo zezimali zase-UK, futhi kungani abenzi bomthetho bezifuna?
Abameli bezomthetho base-UK bafuna "izivivinyo zokucindezeleka ze-AI" eziphatha izinhlelo ze-AI njengezibhange: zihlole ngaphambi kokuba izinkinga zisabalale. Ukukhathazeka ukuthi ukucwasa, amaphutha amaningi, amaphethini okukhwabanisa, kanye nezindlela zokuphendula kungase kubonakale kuphela ngezinga elikhulu. Kwezezimali, ukwehluleka okuncane kungasakazeka ngokushesha ngezinhlelo ezixhumene kakhulu. Ukuhlolwa kokucindezeleka kuhloselwe ukuveza ubuthakathaka kusenesikhathi, ngaphambi kokuba imali yangempela namakhasimende angempela athinteke.
Obani “abaQaphi be-AI” base-UK beDolobha, futhi yini abahloselwe ukuyenza?
Uhulumeni wase-UK uqoke “Ompetha be-AI” ababili ukusiza ukusheshisa ukwamukelwa kwe-AI ezinkonzweni zezezimali. Indima yabo ihlelwe njengokusiza iDolobha ukuthi libambe amathuba e-AI ngenkathi lenza lokho “ngokuphephile.” Empeleni, kufundeka njengokuxhumanisa kanye nokumela: ukukhuthaza ukwamukelwa, ukusekela imikhuba emihle, kanye nokuhlanganisa ababambiqhaza. Kubonisa ukuthi uhulumeni ufuna ukuthunyelwa okusheshayo, kanye nephepha lokuphepha elibonakalayo elizungezile.