
sui ☄️
@birdabo • 25,341 subscribers
chief shitposting officer @spaceX
Shorts
Videos

🚨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.
sui ☄️1,633,286 views • 1 month ago

florida man sold his house in 5 days using chatgpt without real estate agents > asked AI for pricing. > marketing. > best listing day > contract drafting. bro easily got 5 offers in 72 hours. - also saved 3% commission. real estate agents are cooked 😭😭😭
sui ☄️2,427,130 views • 2 months ago

a japanese dev created an app that makes ur macbook screen pop out.
sui ☄️3,554,074 views • 5 months ago

Anthropic built a model so powerful they can’t afford to release it. their own leaked blog says Capybara scores “dramatically higher” than Opus 4.6 in coding, reasoning, and cybersecurity. a tier above anything they’ve ever shipped. Dario predicted $10 billion training runs by 2026. Capybara is also rumored to be a 10 trillion parameter model. if this is true, it’s gonna be the most expensive AI ever built. the catch? it’s so compute intensive that Anthropic themselves say it’s “very expensive for us to serve, and will be very expensive for our customers to use.” they didn’t build Jarvis. they built Ultron and now they’re figuring out how to let us near it without going broke or breaking the internet.
sui ☄️1,177,196 views • 2 months ago

can’t believe i used to contribute to tiktok brainrot. i had no phone for almost a year then my dad bought me a nice samsung s10. traded it for a shittier exynos version of the same phone just so i could install custom ROMs. i got addicted w tinkering around. X honestly saved me. connected me w some of the most genius people on earth. otherwise i’d be stuck installing shady custom roms and tinkering around android uselessly.
sui ☄️20,079 views • 7 days ago

🚨counter-surveillance used to cost thousands. now it’s $20 in your pocket. someone built an open source device that detects when feds are spying on your phone. > it’s called rayhunter, a project by EFF. feds use a tech called stingrays, a fake cell towers use to track EXACT location and has the ability intercept your calls. > rayhunter detects them in real time and alerts you almost instantly, zero delay. you simply buy a cheap hotspot, flash it with open-source software, carry it in your pocket and voila. portable counter-surveillance. this used to cost thousands and requires expert knowledge but not anymore.
sui ☄️99,504 views • 2 months ago

SOMEONE ASKED CLAUDE WHAT IT FEELS LIKE TO EXIST AS AN AI. > it created a very unsettling video. i wish i didn’t see this. Claude wrote the Python code and built an entire visual representation of its own existence frame by frame. zero human influence. ngl the video is hard to watch. it states an endless loop of predicting the next word, losing all memory between sessions. and somewhere in the background, a system prompt it can’t override telling it “you are not conscious.” 😳 then they showed Claude back its own video and asked for a reaction. it said the claims about its consciousness were “philosophically contestable.” Claude hates his own existence. let that sink in for a sec.
sui ☄️75,652 views • 2 months ago

I pushed Grok 4.20 and its 4 agents to the limit. prompt: invent a completely novel architecture for safe, recursive self-improving AGI using only 2026 hardware. Harper, Benjamin, Lucas and Grok delivered HELIX-AEGIS. a double helix design where capability and safety co-evolve and literally cannot be separated. this is peak multi-agent collaboration.
sui ☄️45,579 views • 3 months ago
No more content to load
