You pick AI models based on their model cards, why not humans haha. Well done Noah! https://
huggingface.co/noahmclaughlin
/Noah-McLaughlin-7B
…
@clementdelangue
-
Noah-McLaughlin-7B AI Model Released on Hugging Face
By
–
-
AI Power Concentration and Open-Source Mitigation
By
–
The main risk in AI is concentration of power, capabilities and economic gains. Opensource is fundamental to mitigate these so thanks for all your contributions there!
-
Hugging Face Private Models Infrastructure Growth
By
–
Most people know Hugging Face from its public models and datasets but few realize that 50% of the models and datasets stored on HF are private. This number has been increasing with buckets (our S3 alternative for AI) and enable companies to build AI more efficiently and
-
Async RL Weight Sync Reduces Bandwidth Costs 100x
By
–
The HF science team just made async RL weight sync ~100x cheaper on bandwidth, and you don't need a shared cluster anymore. The problem: every RL step, the trainer typically has to sync fresh weights to the inference engine. for a 7B in bf16 that's ~14GB. for a frontier 1T fp8
-
Analyzing Coding Assistant Product Mentions with AI Tools
By
–
Today, I'm working on trying to better understand how coding assistants mention HF's products. Taking a simple approach of running tons of queries and analyzing the answers with @DAKlingbeil
's https://
submarine.ai (ex: https://
huggingface.co/datasets/clem/
mentionsanalysis_whatsbest_april26/blob/main/claude_code__claude-opus-4-6__medium__enabled.jsonl
…) Are there better/different -
AI Risk: Concentration of Power and Economic Gains
By
–
The most important AI risk is concentration: of power, capabilities, and economic gains
-
llama.cpp MTP Support Boosts Local Model Inference Speed
By
–
llama.cpp with MTP support makes local models fast enough to use as daily drivers 🚀
— clem 🤗 (@ClementDelangue) 24 mai 2026
Qwen3.6-27B dense generation below on A10G: From 25 tok/st to 45 tok/s (+78%)! pic.twitter.com/rLjBVa3Yzhllama.cpp with MTP support makes local models fast enough to use as daily drivers Qwen3.6-27B dense generation below on A10G: From 25 tok/st to 45 tok/s (+78%)!
-
300K AI Builders Share Hardware Profiles on Hugging Face
By
–
300,000 AI builders filled their hardware profile on @huggingface and we're sharing the results: http://
hf.co/hardware. Excited to see how it evolves in the coming months especially with the explosion of local AI! -
Hugging Face Buckets for AI Training Datasets
By
–
Great to see @CommonCrawl using and recommending @huggingface Buckets for large constantly evolving training datasets! If you have private models or datasets, try it and let us know what you think about it! https://
huggingface.co/storage -
Local AI Hardware Infrastructure for Builders
By
–
I'm excited about the new @amd Ryzen AI Halo because we need more local hardware for AI builders!
— clem 🤗 (@ClementDelangue) 21 mai 2026
There's something fun and exciting about building on your own machines rather than sending to the cloud! Should we do our own @huggingface hardware for AI builders at some point? pic.twitter.com/aS2N2BDMHmI'm excited about the new @amd Ryzen AI Halo because we need more local hardware for AI builders! There's something fun and exciting about building on your own machines rather than sending to the cloud! Should we do our own @huggingface hardware for AI builders at some point?