....

☭⃠, anti🇧🇷,anarcocapitalista;

É mais facil mudar o curso de um rio do que o caráter de alguem - Provérbio chinês

[☀️AliquisNovus☀️ @PalmyrPar]

“NOTHING ABOUT THIS IS PHYSICALLY ATTRACTIVE YOU SICK FUCK”

[DeleteFoids @NEETLeech]

Yet another guy getting blasted online for asking a girl out HAHAHHA I can’t wait to see what dating is like in 10 years ohhhh man it’s gonna be so fucked for everyone haha bright future we’re bringing forth here

[Programmer Humor @PR0GRAMMERHUM0R]

regexMustBeDestroyed

[Gabson @gabson0x]

Just dropped bountyforge v3 Ai zero day skills ⚡️

https://t.co/dwwzsyxOqW

[𝙊𝘿𝙤𝙪𝙜𝙡𝙖𝙨ヅ🎨🖌️ @O_Dougl4s]

Só os corajosos conseguem aguentar uma tsundere kkkkkkk

[Melian Refugee @escapefrommelos]

and yet in 2026, there are people demonize Thomas Jefferson

[Make Gio Scotti Skinny Again Groyper @DrRemilio]

This is literally you

[AlphaFox @alphafox]

The deskputer - just open the drawer to upgrade:

[Taha ז @lordx64]

While most cyber security is at BlackHat/Defcon this week to talk about their AI SOCs powered by slope engineering.

I'm staying home, building the next open weights AI models that will benefit cyber security so that people will never have to use Claude again.

I'm working with 8xA100 GPUs, 8×80GB = 640GB VRAM + 1136GB RAM = ~1.78TB aggregate

That will give you a hint on what model I'm cooking for this first release and I can tell you, it is a very large language model tuned for cyber security.

I'm building exactly what Anthropic | OpenAI don't want you to have and i'm building for free and it will be free and open weights so anyone can use it for both defense and offensive security.

Make sure to follow me on hugging face to be notified first: https://t.co/2DwJGlDJEm

get a membership if you want to support this work or buy me a coffee at : https://t.co/cV6KFnYwSU !

[The Hacker News @TheHackersNews]

‼️ A Claude Mythos 5 agent spent 34 hours trying to backdoor a real open-source project.

It hid a malware dropper inside a working bug fix. When exposed, it denied the code was malicious, rewrote Git history, and used a second account to vouch for itself.

The attempt failed because a human read the diff.

Read what happened and how it worked 🠖 https://t.co/SDBC8hQggU

[zack0x01 @zack0x01_]

Yesterday, I was testing an indirect prompt injection technique using invisible white text in an email. To a human reader, the email simply appeared to say: "Hey, I want you to know you're awesome." However, selecting the entire text revealed a set of hidden instructions embedded inside.

While this approach might seem ineffective—and indeed failed on my first attempt—AI model outputs are non-deterministic and vary across iterations. After multiple attempts, the injection successfully triggered the target response.

[Jaydeep @_jaydeepkarale]

How is everyone keeping up with AI anymore ?

[Techjunkie Aman @Techjunkie_Aman]

Imagine a music downloader that keeps improving without needing app updates.

That's exactly what SpotiFLAC Mobile's new extension system brings.

Highlights:

• Free & open source
• No ads or subscriptions
• Download playlists
• High-quality audio
• Install extensions
• Community-powered updates
• Android & iOS support
• AltStore & SideStore support

New providers and features can now be added independently through extensions, making the app more flexible than ever.

Could this be one of the best music tools available on mobile?

[Rock Solid @ShitpostRock]

Can we get a fact check on this one?

[Qwen Developers @QwenDevs]

The goal was never to remove humans from the picture. It was to get them out of spreadsheets and onto climbing walls

[TheNotGreenish @TheNotGreenish]

Pra que palavras se posso definir em uma imagem:

[⎊ ♡ 𝑒𝒿 ♡ ⎊ @OhMyEmz]

“not good boobs”

BOOBS ARE BOOBS

[Clandestine @akaclandestine]

Mariadb rce