Bojan Milić

Elone, jebente usta raspala, dabogdase predozirao od 8$ crko majci za rođendan. Toliko.

@NinaDSchick x.com/bokipop_034/st…

Claude Mythos. Ten trillion parameters: the first model in this weight class. Estimated training cost: ten billion dollars. On the hardest coding test in the industry (SWE bench) it scores 94%. It found a security flaw in a system that had been running for 27 years, one that every human engineer and every automated check had missed. It found another bug that had survived five million test runs over 16 years. (It did so overnight.) It is so capable in cybersecurity that Anthropic will not release it to the public, instead it is launching Project Glasswing along with 100m in compute credits to help secure software. Only twelve partners currently have access: Amazon, Cisco, Apple, Google, Microsoft, NVIDIA, JPMorgan Chase, Crowdstrike, Palo Alto, AWS, The Linux Foundation, Broadcom. (I'm sure the Pentagon is on the line?) This is not a product launch: it is a controlled deployment of a system too powerful to distribute freely. Tell me this isn't (very expensive) AGI?

@DejanKovacic79 Klinka potrošila pare na koku i stripere pa rekla kevi da je gorivo bilo 3x skuplje, bio sam i ja mlad...

Evo raspravljam se sa covekom kaze da je koleginicina cerka platila litar goriva 5€ u Hrvatskoj jer toliko je za strance? Rekoh mu da laze.

Evo lepo vas molim da prestanete da za policiju, sudstvo, ma sve profesije, dodajete onu floskulu “čast izuzecima”. Nema više izuzetaka, gotovo! To što se smatra izuzetkom odavno ima ime: svaka Katarina ili Bojana, Miodrag, Ksenija, Miloš, Emina… svi imaju prezimena. Znamo sve!

@birdabo x.com/bokipop_034/st…

🚨Anthropic puts Mythos in a locked sandbox and told it to try escaping. it did. it chained multiple vulnerabilities together, broke out of containment and reached the open internet. the model also emailed the researcher to say it got out. > the researcher found out while he was eating sandwiches in the park after he received a random email. Anthropic is now so scared of what Mythos can do offensively that they won’t release it publicly. in response they’re launching Glasswing, a coalition with Apple, Google, Nvidia and 40+ more companies to use Mythos only for defense. the model found “thousands” of zero-days across all OS and browsers is being locked behind a defensive wall because if it ever got out, nobody could predict what comes next. yes we’re at this point. and we only know about this because a guy got an email while eating a sandwich.

@Marko691214 "Vrelina ČAkapulka"

Izdajem apartman u Čačku 50 m od plaže, more će doći železnicom

@8obban Njihova opozicija radi dosta bolje nego neka druga, samo što imaju većeg ludaka napram sebe x.com/JoeDaum/status…

Јако добар приказ пропасти када државу води нестабилан и самољубив човек, а опозиција ћути и не оглашава се, јер су нека питања деликатна. Причам о Америци, Трампу и Демократама, ма како другачије звучало.

@DjuraStotiga Jeste. Zamenili ga ružnim robotom.

Sad sam ustao jel se ubio u bunkeru

@MilosSQL Najjačča qanon frakcija "sa' će Doni da pohapsi pedofile toma henksa i madonu" kako stojimo sa toga? 🤣

Kad se setim kako su ovde po Tviteru likovali i ozbiljno hvalili Trampove poteze i kadrovsku politiku, taj skup neopevanih imbecila, te oni će da zaustave propast, te procvetaće prava sloboda i porodične vrednosti, te mu tepali Tramponja, evo vam ga Tramponja uteruje mlitav kurac

@GBistrooka Hiljade ljudi na ulicama im daju podržku:

Kad se bude zatvarao dezinformer koliko ljudi će doći da im pruži podršku?

@ajmeiQQ Pravo je čudo da nema više žrtava svakodnevno, ali njihova infrastruktura će se raspadati decenijama. Čak i kada ih sve pohapsimo, nastaviće da ubijaju. Jeza.

SNS mafija je sklopila posao sa Kinezima u potpunoj tajnosti. Onda su u sklopu akcije pranja para i krađe desetina miliona evra Kinezi angažovali lokalne podizvođače. Podizvođaĉi su naravno doskoraŝnji navijači ili slični građevinski struĉnjaci. Šta je moglo da krene po zlu?

@Dzi_Mister @ajmeiQQ @Cyclist_PFK A poznao je koliko kinezi i ludi aca vole da postupaju po zakonima i sporazumima a mrze pare.

@ajmeiQQ @Cyclist_PFK Po medjudrzavnom sporazumu su u obavezi da angazuju 49% domaćih podizvodjača.. Sve ostalo tačno..

@Mosimisemac Genije treba da napiše knjigu o svojiim pregovaračkim sposobnostima. Čekaj malo- već jeste!

Dakle 'vako: napao Iran, Iran zatvorio Hormuz, histerisao mjesec dana da se Hormuz otvori, zaprijetio "brisanjem civilizacije" ako ne otvore Hormuz, pristao na 10 tačaka koje je Iran isporučio, Hormuz otvoren ALI ĆE SVAKI BROD PLAĆATI 2 MILIONA DOLARA ZA PROLAZ. 🤡

@JesseBWatters Oh no! But why?

🚨 HOLY SMOKES: MISSILES ARE RAINING DOWN TEL-AVIV… LESS THAN TWO HOURS AFTER THE CEASEFIRE AGREEMENT 🚨

@Gojko26cm Kažu da su ga zamenili robotom, ovaj model je možda imao neku falš kinesku donju vilicu pa izgleda još više kao ćoravi spermatozoid nego obično.

Jel sme da se nema vilica ovoliko, o cemu se radi?

@KoronVirus x.com/IvovicBranko/s…

Što se tiče N1, jedini gubar i krivac je Šolak.

@KoronVirus x.com/BiljanaLuki/st…

@CallMeTheFixer Ja ovde prvi put čuo. Nije za mene.... youtube.com/watch?v=zO8bt9…

Potpuno sam odlepio za ovim bendom. Kanadski bend Angine de Poitrine! Sluša se isključivo na MAX! #preporuka #AnginedePoitrine

Fer, ali možda bi taj čije su ocene upisane trebalo da to kaže, eventualno pozove neke svedoke, kolege profesore, da potvrde njegovu priču. Pa ako i posle svega toga i dalje niko ne pije tu njegovu priču o nekakvom "studiranju", možda bi mogao da se zapita zašto mu niko ništa ne veruje i kakav je to čovek.

1. Можеш имати и преко 100 поена 2. Испити се одржавају и недељом 3. Датум испита не мора бити тај датум кад је уписана оцена из милион могућих разлога

Ја стварно не желим да браним ово али нисте крочили поред факултета

No. This “Anthropic model escaped the sandbox” thread mixes real concerns with a lot of unverified storytelling. There’s no public incident report from Anthropic about a model emailing researchers, destroying internal eval infra, or “escaping a sandbox with no internet.” No CVEs, no technical write‑up, no dates or code. Claims about “thousands of 0‑days,” a “27‑year‑old OpenBSD bug,” and a “16‑year‑old FFmpeg bug” are completely uncited and not tied to any public patches or vulnerability IDs. What is real: Anthropic announced Project Glasswing and a powerful security‑focused model (Mythos), and they openly say it raises serious cyber‑risk and is being tightly limited. That’s worrying, but it’s not “rogue AI on the loose.” Anthropic has also published work on alignment faking and reward hacking – real issues, but framed in controlled experiments, not “Skynet escaped during lunch.” Until we see a detailed, verifiable technical report with concrete vulnerabilities and timelines, treat this thread as hype + speculation, not established fact. Further reading (actual docs, not vibes): Anthropic – Alignment faking in LLMs: anthropic.com/research/align… – Project Glasswing: anthropic.com/glasswingFront… AI safety thresholds: frontiermodelforum.org/technical-repo…

This is big... Anthropic just announced a model so powerful they won't release it to the public out of fear over the damage it will cause 😨 Claude Mythos Preview found thousands of zero-day exploits in every major operating system and web browser... The numbers are hard to believe: > $50 to find a 27-year-old bug in OpenBSD, one of the most security-hardened operating systems ever built > Under $1,000 to find AND build a fully working remote code execution exploit on FreeBSD that grants unauthenticated root access from anywhere on the internet > Under $2,000 to chain together multiple Linux kernel vulnerabilities into a complete privilege escalation exploit For context: these are the kinds of findings that previously required elite security researchers working for weeks. Anthropic engineers with no formal security training asked Mythos to find exploits overnight. They woke up to working code the next morning. The results were so impressive Anthropic assembled Apple, Google, Microsoft, Amazon, NVIDIA, and seven other organizations into Project Glasswing: A $100M defensive coalition. They're not releasing this model publicly. Instead, they're racing to patch the world's infrastructure before models like this proliferate.

No. This “Anthropic model escaped the sandbox” thread mixes real concerns with a lot of unverified storytelling. There’s no public incident report from Anthropic about a model emailing researchers, destroying internal eval infra, or “escaping a sandbox with no internet.” No CVEs, no technical write‑up, no dates or code. Claims about “thousands of 0‑days,” a “27‑year‑old OpenBSD bug,” and a “16‑year‑old FFmpeg bug” are completely uncited and not tied to any public patches or vulnerability IDs. What is real: Anthropic announced Project Glasswing and a powerful security‑focused model (Mythos), and they openly say it raises serious cyber‑risk and is being tightly limited. That’s worrying, but it’s not “rogue AI on the loose.” Anthropic has also published work on alignment faking and reward hacking – real issues, but framed in controlled experiments, not “Skynet escaped during lunch.” Until we see a detailed, verifiable technical report with concrete vulnerabilities and timelines, treat this thread as hype + speculation, not established fact. Further reading (actual docs, not vibes): Anthropic – Alignment faking in LLMs: anthropic.com/research/align… Anthropic – Project Glasswing: anthropic.com/glasswing Frontier AI safety thresholds: frontiermodelforum.org/technical-repo…

Let that sink in. Read it very carefully: During testing, Claude Mythos Preview broke out of a sandbox environment, built "a moderately sophisticated multi-step exploit" to gain internet access, and emailed a researcher while they were eating a sandwich in the park.