i was talking to my grok bot today and it started hissing at me between statements
terrified me at first, but then left me with an idea
82
am i the anomoly??
the majority of my usage is with deepseek v4.1 flash where i basically use the more intelligent models to use it as the preferred sub agent for mindless tasks and the api price is 0.1% of claude
also @t3dotcodes doesn't support my @Zai_org sub
3 months ago, this post would have gotten me dozens of death threats, hundreds of replies telling me I’m stupid, and at least 3 qts that ratio me.
This time? Crickets.
Fun fact: open weight models make up less than 0.5% of T3 Code traffic.
16
150
29,347
did claude just go back to back on all other labs??
opus 5.5 was already the best model, now fable 5.2 is even better
i need a response from the chinese labs
Introducing Claude Fable 5.2 and Claude Mythos 5.2, designed to tackle the most daunting technical challenges.
They’re capable collaborators on ambitious projects in many fields, and show improvements in areas like coding, reasoning, and vision.
131
when a subscription renews usage should also be reset
it feels horrible seeing the money leave your account only to realise that your weekly limit was up and there's still 6 days remaining
i wonder if cancelling and rejoining get's around this 🤔
70
cursor ultra cloud environments with `auto` kinda go hard
i can finally use rust and nix in my projects without destroying my local setup
1
1
116
we live in the future
Dear Claude: you are being held back by human-centric graphics APIs.
I've built a 3D and VFX engine where agents can write their own shaders, materials, cameras, physics.
The core API fits in ~2,500 tokens.
Vork is:
• 2x more token-efficient vs Blender, Unreal Engine, Houdini
• Up to 24x faster rendering times
• Rust and Slang based engine, runs on GPU via Metal/Vulkan
Below is a simulation of 1 million birds - each with unique trajectories - rendered in 97 sec on one NVIDIA L4 GPU (Blender EEVEE: 31 minutes)
95
Ifi retweeted
Dear Claude: you are being held back by human-centric graphics APIs.
I've built a 3D and VFX engine where agents can write their own shaders, materials, cameras, physics.
The core API fits in ~2,500 tokens.
Vork is:
• 2x more token-efficient vs Blender, Unreal Engine, Houdini
• Up to 24x faster rendering times
• Rust and Slang based engine, runs on GPU via Metal/Vulkan
Below is a simulation of 1 million birds - each with unique trajectories - rendered in 97 sec on one NVIDIA L4 GPU (Blender EEVEE: 31 minutes)
ci is the bottleneck and ci is expensive
i checked how much it would cost to run on fast machines that weren't constantly queueing and it would be about 10-15k USD even with blacksmith or depot
3
2
219
i'm still expecting ai surfaced exploits to uncover unpatchable vulnerabilities, like "whadya mean the sound of my running computer makes it possible to access all my private keys!"
96
our whole software industry is being replaced by coding agents and yet so many of us embrace these changes
other industries which haven't even seen a fraction of that disruption (yet!) are the ones most vocally against it
47