so much more fun to use a computer via codex
COMPUTING
-
Power limiting reduces inference performance by 10% but saves 50% electricity
By
–
Power limited* not throttled Performance loss is negligible, around 10% in inference but I save 50% in electricity
-
Kernels Are the Actual Work in Model Inference
By
–
You don’t “run a model”
You run Kernels The model is just a graph The Inference Engine is scheduler / optimizer / executor But the actual work? That happens in the Kernels – MatMul Kernels
– Attention Kernels
– RMSNorm Kernels
– KV cache Kernels
– Quantized linear Kernels
– -
Integrating Email with ChatGPT: A New Frontier in Communication
By
–
email integration with chatgpt https://t.co/h04wa9TESm
— Greg Brockman (@gdb) 6 juin 2026email integration with chatgpt
-
CEO Rodrigo Liang discusses heterogeneous AI systems and token speed on CNBC
By
–
On CNBC, our CEO @RodrigoLiang talks about the future of heterogeneous AI systems, why token speed and energy efficiency matter for agentic inference, and what’s ahead for us over the next 12 months
-
Speed comparison between Gemini 3.5 Flash and Kimi K2.6 on Cerebras
By
–
Google just released its fastest model: Gemini 3.5 Flash. We directly pitted it against Kimi K2.6 on Cerebras. Both are equal in intelligence, but what about speed? Full benchmark results: https://cerebras.ai/blog/which-is-faster-gemini-3-5-flash-or-kimi-k2-6-on-cerebras
… -
Claude’s new version design, Mythos Preview achieves 52x training speedup
By
–
🚨 Dario confirmed that Claude is currently designing the next version of itself.
— Charly Wargnier (@DataChaz) 5 juin 2026
To test this, the company asks its new models to optimize the training code for smaller AIs.
While Claude Opus 4 achieved a 3x speedup, Mythos Preview hit a staggering 52x 🤯
Recursive… pic.twitter.com/MONKQ9g1aZDario confirmed that Claude is currently designing the next version of itself. To test this, the company asks its new models to optimize the training code for smaller AIs. While Claude Opus 4 achieved a 3x speedup, Mythos Preview hit a staggering 52x Recursive
-
Future warfare powered by non-biological intelligence and human-machine tandem
By
–
The future of warfare won’t be fully autonomous—but it will be powered by non-biological intelligence.
— Nina Schick (@NinaDSchick) 5 juin 2026
We’re entering an era where humans and machines operate in tandem. Think of Project Maven: real-time battlefield data feeding intelligent systems that augment decision-making,… pic.twitter.com/HEl8McBAZoThe future of warfare won’t be fully autonomous—but it will be powered by non-biological intelligence. We’re entering an era where humans and machines operate in tandem. Think of Project Maven: real-time battlefield data feeding intelligent systems that augment decision-making,
-
AI Self-Improvement: New Advances and Outlooks
By
–
Sources:
Anthropic, When AI builds itself:
https://anthropic.com/institute/recursive-self-improvement
… SkillOpt (arXiv): https://arxiv.org/abs/2605.23904 SkillSmith (arXiv): https://arxiv.org/abs/2606.01314 MOSS (arXiv): https://arxiv.org/abs/2605.22794 Co-Scientist, Google DeepMind: https://deepmind.google/blog/co-scientist-a-multi-agent-ai-partner-to-accelerate-research/
… Gary Marcus, No need to panic -
tt-vscode-toolkit: interactive learning environment
By
–
Introducing tt-vscode-toolkit, an interactive learning environment designed to accelerate developing on Tenstorrent hardware. The toolkit brings project templates directly into VS Code, with lessons covering model deployment, agent frameworks, video generation, and more. Get