🧑💼 I-OpenAI iyeza ngenxa yalezo zimali zebhizinisi ezimnandi ngo-2026 ↗
I-OpenAI yashintsha ubuholi futhi yabeka uBarret Zoph ukuthi aphathe umkhankaso wokuthengisa webhizinisi - amandla aqinisekile "ezicathulo zabantu abadala".
Umyalezo uthi: ama-demo ambalwa okubonisana ngesayensi, "singafika kanjani futhi sande kanjani ngaphakathi kwezinkampani ezinkulu." Bakhomba futhi izigidi zabasebenzisi bebhizinisi kanye nama-logo amakhasimende athile aziwayo kakhulu, afundeka njengokuguquguquka okungabonakali kangako.
🪧 I-CEO ye-Google DeepMind 'iyamangala' i-OpenAI iphuthuma phambili ngezikhangiso ku-ChatGPT ↗
UDemis Hassabis uthe uyamangala ukuthi i-OpenAI isusa lokhu kusenesikhathi ezikhangisweni ngaphakathi kwe-chatbot - hhayi ngoba izikhangiso zimbi, kodwa ngoba isikhathi sizwakala… sishesha.
Uphinde waphakamisa ukuthi i-Google ithathe indlela eqaphile kakhulu (okwamanje), ibukele ukusabela kwabasebenzisi futhi izame ukungazenzi “ukuthuka”. Okufika ngokuhlekisa uma ubheka imodeli yebhizinisi ye-Google yonke.
🎙️ I-Google ithola ithimba elisungula inkampani entsha ye-AI, i-Hume AI, ↗
I-Google DeepMind iletha i-CEO ye-Hume AI u-Alan Cowen kanye neqembu elincane lonjiniyela ngaphansi kwesivumelwano sesitayela selayisensi esihlose ukuthuthukisa izici zezwi ze-Gemini. Akuyona into ethengiswa ngokuphelele, kodwa iziphatha njengayo.
I-Hume iyaqhubeka nokusebenza, kuyilapho i-Google ithola amalungelo angewona awodwa kwezobuchwepheshe - futhi yonke into ingena kahle enkambisweni ethi "thola-ukuqasha kodwa wenze kube kuhlanzekile ngokomthetho". I-Voice iphendukela enkundleni elandelayo yokulwa, kungakhathaliseki ukuthi siyicelile noma cha.
🗣️ Injini ye-Voice AI kanye nomlingani we-OpenAI i-LiveKit ifinyelela inani elingu-$1B ↗
I-LiveKit iqoqe u-$100M ngenani lika-$1B ukuze iqhubeke nokwakha ingqalasizinda yezinhlelo zokusebenza ze-AI zezwi nezevidiyo zesikhathi sangempela - amapayipi enza kube nokwenzeka ukukhuluma.
Ivele inika amandla imodi yezwi ye-ChatGPT futhi ibala inhlanganisela yamakhasimende obuchwepheshe abadumile kanye nabanye abasebenzisi abasengozini enkulu njengezinsizakalo eziphuthumayo. Indaba yokuqala "iphrojekthi yomthombo ovulekile eyaba yinto ebalulekile ebhizinisini buthule," futhi labo bavame ukuba ngabanqobi abayimfihlo.
🧠 Abantu bacabanga ukuthi ukuhlanganisa kuyindlela elandelayo ye-AI, futhi bakha imodeli yokufakazela lokho ↗
I-Humans& iyabheja ukuthi igebe langempela alikho “ukuphendula umbuzo wami,” kodwa “ukusiza iqembu ukuthi lisebenzisane ngaphandle kokulahlekelwa indaba.” Cabanga ngezinqumo ezithatha isikhathi eside, ukushintsha izinto eziza kuqala, ukubambisana okunamandla - izinto ama-chatbot avame ukuzigwema.
Baqoqe imbewu enkulu engu-$480M yokwakha indlela yesisekelo ehlose ubuhlakani bezenhlalo kanye nokuxhumanisa. Kuyinkulumo enesifiso esikhulu, eyinto engavamile, "isimiso sezinzwa esiphakathi sezinhlangano" - kodwa futhi, yebo, ukuxhumanisa yilapho izinhlelo ziya khona.
📊 I-Ipsos yembula i-Horizons, isu layo lokuguqula nokukhula ukuze iqinise ubuholi bayo bomhlaba wonke ↗
I-Ipsos yethule isu elisha eligxile kakhulu ku-AI, ngohlelo lokutshala imali engaphezu kwe-€1B eminyakeni emihlanu ezayo kubuchwepheshe, i-AI, kanye nedatha - ikakhulukazi ngokuthenga kanye nokutshalwa kwezimali okunesu.
Futhi baphusha ushintsho olusheshayo ngaphandle kokulahla ikhwalithi yedatha, futhi baphakamisa ucwaningo "olune-AI-augmented" njengento ehlukanisayo. Ngokuyisisekelo: sheshisa ifektri, ungawehlisi umkhiqizo… kulula ukukusho kunokukwenza, kodwa yilokho ukubheja.
🧩 I-Alibaba izohlela i-IPO yeyunithi yokwenza ama-chip ye-AI i-T-Head, kubika iBloomberg News ↗
Kubikwa ukuthi i-Alibaba ilungiselela uhlu oluhlukile lweyunithi yayo yama-chip i-T-Head, iqala ngokuhlelwa kabusha okunikeza abasebenzi ubunikazi obungaphelele ngaphambi kokuthuthela kwi-IPO. Isikhathi kanye nokulinganisa kuseyinto engacacile - njengoba lezi zinto zihlala zinjalo.
I-T-Head yakha ama-chip kuzo zonke izikhungo zedatha, i-AI, kanye nezigaba ze-IoT, futhi lokhu kufundeka njengokunyakaza "kwama-chip ashisa futhi, ake sibheke inani elingaphezulu". Kubonakala sengathi abatshalizimali bayawuthanda lo mbono, okungenani ekusabeleni okusheshayo.
Imibuzo Evame Ukubuzwa
Kusho ukuthini umkhankaso wokuthengisa we-OpenAI ezinkampanini ngo-2026?
Kubonisa ukuguquka kusuka kuma-demo angenasidingo kuya esimweni "somhlaba nokwandisa" ngaphakathi kwezinhlangano ezinkulu, ngesisindo esikhulu ekupakisheni okulungele ukuthengwa kanye nokuthunyelwa okuphindaphindwayo. Ukukhulunywa ngezigidi zabasebenzisi bebhizinisi kanye nama-logo amakhasimende aqashelwayo kufundwa njengobufakazi bokuthi i-AI yebhizinisi isivele iyinto evamile. Kubathengi, i-takeaway ewusizo ilula: lindela ukukhishwa okwengeziwe okuholwa ukuthengisa, kanye nemizila ecacile yokwandisa ukusetshenziswa kuwo wonke amaqembu.
Ubani uBarret Zoph, futhi kungani ukushintsha kobuholi kubalulekile?
Ukushintshashintsha kubeka uBarret Zoph ukuthi aphathe ukukhushulwa kokuthengiswa kwebhizinisi, okuphakamisa ukuthi i-OpenAI ifuna ukusebenza okubukhali ngokuthengisa ezinkampanini ezinkulu. Ezinhlanganweni eziningi, izinyathelo zobuholi ezinjengalezi zivame ukuhambisana nesinyathelo esiqinile sokuya emakethe, ukumbozwa kwama-akhawunti okuqinile, kanye nokwandiswa kwamakhasimende okuhlelekile. Uma uhlola abathengisi be-AI yebhizinisi, kungabonisa nokugcizelela kwesikhashana esifushane kumadili, ukuvuselelwa, kanye neminikelo ejwayelekile.
Ingabe izikhangiso ziza ku-ChatGPT, futhi kungani i-Google DeepMind isabela kuzo?
Amanothi akhomba ukuthi i-OpenAI iqhubekela phambili ezikhangisweni ezingaphakathi kwe-ChatGPT, lapho uDemis Hassabis ethi uyamangala ukuthi lokhu kwenzeka ngokushesha kangakanani. Ukukhathazeka akucaciswanga ngokuthi “izikhangiso zimbi,” kodwa ukuthi isikhathi singazwakala sishesha uma kuqhathaniswa nokulindelwe ngabasebenzisi kumkhiqizo oxoxwayo. Indlela evamile kulesi simo ukubheka ukusabela kwabasebenzisi, ukushintsha kokuzibandakanya, kanye nemiphumela yokuthembana ngaphambi kokwandisa amafomethi ezikhangiso.
Kungani ingqalasizinda ye-AI yezwi efana ne-LiveKit ibaluleke kangaka ngokuzumayo?
I-LiveKit iqoqe u-$100M ngenani lika-$1B ukwakha "ukufakwa kwamapayipi" ngesikhathi sangempela kwezinhlelo zokusebenza ze-AI zezwi nevidiyo, okuvame ukuba yinto ebalulekile uma imikhiqizo isuka kuma-prototypes iye ekukhiqizweni. Lokhu okuthunyelwe kugqamisa ukuthi kunika amandla imodi yezwi ye-ChatGPT futhi isekela inhlanganisela yabasebenzisi abakhulu bezobuchwepheshe kanye nabasebenzisi abasengozini enkulu njengezinsizakalo eziphuthumayo. Ingqalasizinda eqala njengomthombo ovulekile ingaba umgogodla webhizinisi othule lapho ukuthembeka kanye nokubambezeleka kuqala ukubusa izinqumo.
Kuyini "ukuxhumanisa i-AI," futhi yini abantu abazama ukuyakha?
I-Coordination AI igxile ekusizeni amaqembu ukuthi abhekane nezinqumo ezithatha isikhathi eside, ukushintsha izinto eziza kuqala, kanye nemisebenzi yokubambisana - izindawo lapho ama-chatbot e-Q&A alula avame ukuba nzima khona. I-Humans& ibeka ukubambisana njengomngcele olandelayo, ihlose indlela yemodeli yesisekelo exhunywe ekuhlakanipheni komphakathi kanye nomongo wenhlangano. Umjikelezo wembewu ongu-$480M ugcizelela ukuthi abatshalizimali bawuthatha ngokungathi sína kangakanani umqondo wokuthi i-AI yebhizinisi kufanele ibumbe "indlela umsebenzi owenzeka ngayo," hhayi nje ukukhiqiza izimpendulo.
Yini eboniswa yisu le-AI le-Ipsos kanye nezinhlelo ze-IPO ze-T-Head ze-Alibaba mayelana nokusetshenziswa kwe-AI yebhizinisi?
Isu le-Ipsos elithi “Horizons” ligcizelela ucwaningo oluthuthukisiwe lwe-AI kanye nohlelo lokutshala imali engaphezu kwe-€1B eminyakeni emihlanu kubuchwepheshe, i-AI, kanye nedatha, ikakhulukazi ngokuthenga kanye nokutshalwa kwezimali okunesu. Ngokwehlukana, imibiko ithi i-Alibaba ihlola ukuhlelwa kabusha kanye nendlela engaba khona ye-IPO yeyunithi yayo ye-chip i-T-Head, okubonisa isifiso esisha sabatshalizimali se-AI kanye nehadiwe yesikhungo sedatha. Ndawonye, baphakamisa ukuthi i-AI yebhizinisi iqhuba kokubili ukuguqulwa kwezinsizakalo kanye nokubheja kwengqalasizinda eyisisekelo.