🦞 Umsunguli we-OpenClaw uSteinberger ujoyina i-OpenAI, i-bot yomthombo ovulekile iba yisisekelo ↗
UPeter Steinberger uya e-OpenAI ukuyocindezela “ama-ejenti omuntu siqu,” kuyilapho i-OpenClaw ngokwayo ipakwe esisekelweni ukuze ihlale ivulekile (futhi isekelwa). Lokho kuhlukana… kuhlakaniphile ngandlela thile - qasha umakhi, gcina iphrojekthi isobala.
Inkulumo ye-OpenClaw iwusizo ngendlela evuselelayo: ukuthunyelwa kwe-imeyili, amaphepha omshwalense, ukungena ngendiza, imisebenzi yokuphatha impilo ephikelelayo. Iphinde yaduma ku-GitHub, futhi lokho kuthandwa kuye kwabangela ukukhathazeka ngezokuphepha, ikakhulukazi uma abantu beyisebenzisa ngokunganaki.
🪖 IPentagon 'ikhathele' ukugxekwa kwe-Anthropic mayelana nokusetshenziswa kwemodeli ye-Claude AI ngamasosha, ingase inqume ubudlelwano, kusho umbiko ↗
Impi eyinhloko: iPentagon ifuna ukufinyelela okubanzi, “kuzo zonke izinhloso ezisemthethweni”, kanti i-Anthropic isazama ukugcina imingcele eqinile mayelana nezikhali ezizimele ngokuphelele kanye nokuqapha okukhulu. Lolo uhlobo lokungavumelani okuzwakala kuyifilosofi kuze kube yilapho othile ethi, “singase sikuthathele indawo.”
Umbono owodwa ongahlonishwa kakhulu - izikhulu azifuni ukuthi imodeli ivimbe ngokuzumayo imisebenzi phakathi, futhi azifuni ukuxoxisana ngamacala onqenqema unomphela (kulungile… kodwa futhi yebo). Kukhona ukungezwani kwangempela “kobani ophethe okhiye” lapha, futhi akucashile.
🧠 Imodeli yokwakha isiqalo ukubikezela ukuziphatha kwabantu ↗
I-Simile ikhiphe umjikelezo we-$100M ukwakha imodeli "yokufunda okulinganiselwe" ehlose ukubikezela ukuthi abantu bangenzani - okuhlanganisa, ikakhulukazi, ukulindela imibuzo engaba khona ezintweni ezifana nezingcingo zenzuzo. Inhloso encane, isifiso esikhulu, inhlanganisela eyethusa kancane.
Le ndlela incike ezingxoxweni nabantu bangempela kanye nedatha yocwaningo lokuziphatha, bese isebenzisa ukulingisa ngama-ejenti e-AI ahloselwe ukulingisa izintandokazi zangempela. Kufana nokwenza imodeli yesimo sezulu yezinqumo zabantu... okuzwakala kungenakwenzeka kuze kube yilapho kungenzeki.
🧑⚖️ I-Scoop: I-White House icindezela ummeli wase-Utah ukuthi aqede umthethosivivinywa wokubonisana nge-AI ↗
Isiphakamiso sobusobala be-AI ezingeni lesifundazwe e-Utah sithola ukugxekwa okuqondile yi-White House, lapho izikhulu zinxusa umxhasi womthethosivivinywa ukuba angawuqhubeki phambili. Uhlaka lomthethosivivinywa lumayelana nokucaca kanye nokuphepha kwezingane - okunzima ukuphikisana nakho nge-optical emsulwa.
Kodwa impi enkulu imayelana nomthetho: ukuthi ubani obeka imithetho, izifundazwe noma uhulumeni wobumbano. Futhi yebo, kuyinto exakile - njengabantu ababili ababambe isiteringi esifanayo futhi baphikelela ngokuthi bangabantu abazolile.
🎬 I-ByteDance ithembisa ukuvimbela ukusetshenziswa kwe-IP okungagunyaziwe kuthuluzi levidiyo le-AI ngemuva kokusongela kwe-Disney ↗
I-Disney ithumele umyalelo wokuyeka nokuyeka ukukhiqiza ividiyo ye-AI ye-ByteDance, kanti i-ByteDance ithi iqinisa izindlela zokuvikela ukuvimbela ukusetshenziswa okungagunyaziwe kwe-IP kanye nokufana. Isikhalazo - okusolwa ukuthi - ukuthi ithuluzi lingakhipha izinhlamvu ze-franchise ezijwayelekile njengokungathi nje... ziyizitikha zomphakathi.
Kungukungqubuzana wonke umuntu akubona kuza: amathuluzi evidiyo e-AI asakazekayo ahamba ngokushesha, ama-studio ahamba ephikisana, futhi "sizokwengeza izivikelo" kuba ulimi lokuxolisa oluzenzakalelayo. Ngendlela ehlukile, ubuchwepheshe bubukeka njengomlingo - kanti uhlangothi lwezomthetho lubukeka njengokudonsa phansi.
Imibuzo Evame Ukubuzwa
Kusho ukuthini ukuthi umsunguli we-OpenClaw wajoyina i-OpenAI ngenkathi i-OpenClaw ithuthela esisekelweni?
Kubonisa ukuhlukana phakathi komuntu owakha “ama-ejenti omuntu siqu” kanye nephrojekthi esabuswa ngumphakathi. USteinberger ojoyina i-OpenAI uphakamisa ukuthi uzogxila ekuthuthukiseni imikhiqizo efana neyama-ejenti lapho. Ukubeka i-OpenClaw esisekelweni kuhloswe ngayo ukuyigcina ivulekile futhi isekelwa ngendlela eqhubekayo. Empeleni, lesi sinyathelo sihlose ukugcina ukwethenjwa komphakathi ngenkathi umakhi eya lapho kukhona khona izinsiza.
Kungani ama-ejenti e-AI esitayela se-OpenClaw egxile emisebenzini efana ne-imeyili kanye nephepha?
Ngoba umsebenzi "wokuphatha impilo yonke" uphindaphindwa, usekelwe emithethweni, futhi udla isikhathi, okwenza kube yinto ewusizo yokwenza izinto ngokuzenzakalela. Izibonelo lapha - ukuhlolwa kwe-imeyili, amaphepha omshwalense, kanye nokungena ngendiza - yimisebenzi emincane enezindlela ezicacile zempumelelo. Lokho kugxila kungenza ama-ejenti azizwe ewusizo ngokushesha kunabasizi abaningi abavulekile. Kugcizelela nokuthi kungani ukulawulwa kokufinyelela okucophelela kubalulekile lapho ama-ejenti ethinta ama-akhawunti omuntu siqu.
Ungayisebenzisa kanjani i-ejenti ye-AI yomthombo ovulekile njenge-OpenClaw ngaphandle kokudala izinkinga zokuphepha?
Yiphathe njengesofthiwe engabona idatha ebucayi, hhayi njengesikripthi sethoyizi. Indlela evamile ukuvala iziqinisekiso, ukukhawulela izimvume okungenani okudingekayo, nokugcina izingodo kanye nemizila yokuhlola. Yisebenzise endaweni evinjiwe bese uyihlukanisa nezinhlelo zenani eliphezulu. Izinkathazo eziningi zokuphepha zivela ekusetshenzisweni ngokunganaki, ikakhulukazi lapho abantu beveza ama-endpoints noma amathokheni ngaphandle kwezivikelo eziqinile.
Kungani iPentagon ingajabule ngemikhawulo ka-Anthropic yokusetshenziswa kwezempi kukaClaude?
Impikiswano igxile ebubanzini nasekulawuleni: iPentagon ifuna ukufinyelela okubanzi “kwazo zonke izinjongo ezisemthethweni”, kuyilapho i-Anthropic ichazwa njengegcina imingcele eqinile ezungeze izikhali ezizimele ngokuphelele kanye nokuqapha okukhulu. Izikhulu azifuni futhi ukuthi amamodeli avimbe ukuhamba komsebenzi phakathi noma adinge izingxoxo ezingapheli zecala elingapheli. Lokho kungezwani akucaci njengoba kuzwakala - kumayelana nokuthi ubani onquma ukuthi imodeli ingenzani emisebenzini yangempela.
Izinkampani ezintsha zizama kanjani ukubikezela ukuziphatha kwabantu nge-AI, futhi kungani kuzwakala kuyimpikiswano?
Isibonelo lapha, i-Simile, silandela imodeli "yokufunda okulinganiselwe" ehlose ukubikezela lokho abantu abangase bakwenze, okuhlanganisa nokubikezela imibuzo engaba khona ezimweni ezifana nezingcingo zenzuzo. Indlela echazwe ihlanganisa izingxoxo nedatha yocwaningo lokuziphatha kanye nokulingisa kusetshenziswa ama-ejenti e-AI ahloselwe ukulingisa izintandokazi zangempela. Kuzwakala kuyinqaba ngoba ishintsha i-AI isuke ekuphenduleni abantu iye ekubikezeleni. Inselele ukugcina izimangalo zilinganiselwe nokugwema ukuzethemba ngokweqile.
Kwenzekani uma amathuluzi evidiyo e-AI ekhiqiza abalingiswa abanamalungelo okushicilela, njengasekungqubuzaneni kwe-ByteDance-Disney?
Iphethini ebikiwe ijwayelekile: isitudiyo sikhipha umyalelo wokuyeka nokuyeka, futhi ipulatifomu isabela ngokuqinisa izindlela zokuvikela ukuvimbela ukusetshenziswa kwe-IP okungagunyaziwe noma ukufana. Kumathuluzi amaningi, izindlela zokuvikela zisho izihlungi zokuqukethwe eziqinile, ukutholakala okuthuthukisiwe kwezinhlamvu ezibonakalayo, kanye nokuqiniswa kwenqubomgomo yomsebenzisi okucacile. Ingxabano eyisisekelo ijubane uma kuqhathaniswa nesibopho - ukukhiqizwa kwegciwane kuhamba ngokushesha, futhi ukuqiniswa kwamalungelo kusebenza njengokudonsa phansi. Lindela okuningi kwalokhu kushayisana njengoba abakhiqizi bevidiyo besakazeka.