Me ya faru
Masu binciken tsaro daga Hacktron AI sun yi amfani da Anthropic's Claude Opus 5 don hacking na OpenAI asusu na ma'aikacin ChatGPT kuma su kai ga mahallin GitHub na kamfanin. Masu binciken sun gano wata lahani a cikin aiwatar da sa hannun guda ɗaya na OpenAI wanda ya ba su damar juyar da yanayin dandalin zuwa ga asusun ChatGPT da Codex na masu amfani waɗanda suka inganta ta hanyar sabis ɗin. Daga nan sai suka yi amfani da damar samun damar asusun ma'aikaci da aka lalata zuwa Codex don ƙirƙirar buƙatun ja mara lahani a cikin monorepo na cikin gida na OpenAI, yana nuna cewa sasantawar asusu na iya wuce ChatGPT kanta zuwa abubuwan haɓakawa masu alaƙa.
Masu binciken sun yi amfani da Anthropic's Claude Opus 5 don bincika fakitin libheif mai rauni da haɓaka amfani.
Da farko sun yi amfani da Claude Opus 4.8 amma sun canza zuwa Opus 5, wanda ya haifar da amfani da ARM64 mai aiki a cikin sa'o'i.
Masu binciken sun daidaita amfani da yanayin x86-64 da jemalloc da Discourse ke amfani da shi.
Cikakkar hanyar daga ganowa zuwa samun damar zuwa wurin ma'ajiya ta OpenAI ta ɗauki ƙasa da sa'o'i 72.
Masu binciken sun ba da rahoton binciken ta hanyar shirin Bugcrowd na OpenAI a ranar 25 ga Yuli kuma sun sami kyautar $6,500.
Bayanan tushe: venturebeat.com ↗
Me ya sa yake da mahimmanci
Wannan lamarin yana nuna haɗarin da ke da alaƙa da wakilan AI da ke da alaƙa da tsarin kasuwanci da mahimmancin ware bututun sarrafa fayil ɗin da ba a amince da su ba, kiyaye ƙarancin matakan dogaro da ƙarfi, ƙuntata amincin haɗin gwiwa, da kula da takaddun shaida na AI tare da bincike iri ɗaya kamar asusun ɗan adam masu gata.
Lamarin ya nuna haɗarin da ke tattare da wakilan AI da ke da alaƙa da tsarin kasuwanci.
Yana nuna yadda masu ba da lambar AI ke canza tattalin arziƙin cin gajiyar rashin ƙarfi.
Masu binciken suna jayayya cewa masu ba da lambar iyaka na iya ƙara yin yawancin ayyukan ci gaba na ci gaba a ƙarƙashin jagorancin ɗan adam.
Faɗin yanayin muhallin libheif yana ƙarfafa matsalar kulawa, tare da ɗimbin shawarwarin tsaro da aka buga yayin 2026.
Lamarin ya jaddada bukatar kungiyoyi su dauki matakin da ya dace kan tsaro, gami da ware bututun sarrafa fayil marasa amana, kiyaye masu dogaro da kai da karfi, hana amincewar tarayya, da kuma kula da bayanan wakilai na AI tare da bincike iri daya da asusun mutane masu gata.
Ingantacciyar hanyar sadarwa: Yadda A zahiri yake Aiki
Bincika fasahar da ke bayan wannan ci gaban ta hanyar mu'amala.
crm_get_transaction(id='4092').An agent must create a draft calendar event for Tuesday at 2 p.m. Which evidence would establish the requested result?
Abin kallo na gaba
Lamarin ya nuna yadda wakilai na AI ke canza tattalin arziƙin ƙwaƙƙwaran rashin ƙarfi da kuma buƙatar ƙungiyoyi su ɗauki matakan da suka dace don tsaro.
Lamarin ya nuna tasirin tasirin AI akan tsaro.
Hakan ya nuna bukatar kungiyoyi su dauki matakan da suka dace na tsaro.
Masu binciken suna jayayya cewa masu ba da lambar AI suna canza tattalin arziƙin cin gajiyar rashin ƙarfi.
Lamarin ya jaddada mahimmancin ware bututun da ba a amince da su ba da kuma kiyaye masu dogaro da ƙasa da ƙarfi.
The researchers note that the same class of models is making technically difficult exploit development faster and less expensive.