We can no longer say open-source AI is months behind frontier models.
GLM-5.2 matches Sonnet-5 in parameter size, but absolutely crushes it in performance, speed, and cost.
Just imagine when GLM drops a 1.6T or 5T modelโOpus and Fable won't even stand a chance.
At this point, it's more accurate to say closed-source AI is months behind open-source.
What people are missing about the US government's AI regulation announcement:
ID verification will now be forced on all accounts to prove citizenship.
Frontier labs will take your data, and your sovereignty is officially dead.
A permanent underclass division and a total control society are beginning right now.
People ignored me when I started saying this last year, but it is happening right in front of our eyes.
Get into Open-Source and Sovereign AI.
Advancing together through collective intelligence is the only way to fight back.
A few years ago, I used to dream about fast cars, beautiful girlfriends, and expensive houses.
Now, the only thing I'm obsessed with is fast, beautiful, and expensive VRAM.
Because that one thing has it all.
If you mainly use local LLMs for Hermes-style agentic loops, this might surprise you:
Qwen 3.6 35B actually *beats* DeepSeek v4 Flash โ especially on tool-heavy & coding-adjacent workflows.
You're not missing out.
Full results ๐
https://t.co/IgfJpWx8U3
This is amazing.
If youโre in offensive security no-brainer to have this deployed on something like @modal or other serverless provider.
Costs nothing when idle and you have an absolutely unhinged model for $10-15 a hour when you need it.
i have no idea how @0xSero manages to tweet this much. trying to match some of that energy so i can share my experiments, codex is going to help. more soon ๐
@thsottiaux can you please show us when resets are going to expire. im trying to get codex to hit the backend rn would much prefer to just see it in UI