Tootfinder

Hugging Face begins taking orders for its $449 Reachy Mini Wireless and $299 Reachy Mini Lite open source robots, which come as kits for developers to build (Rebecca Szkutak/TechCrunch)
https://techcrunch.com/2025/07/09/hugging-fa…

Hugging Face opens up orders for its Reachy Mini desktop robots | TechCrunch
Hugging Face is releasing two versions of its desktop robot meant for designing and coding AI applications.

Anatomy of a Machine Learning Ecosystem: 2 Million Models on Hugging Face
Benjamin Laufer, Hamidah Oderinwale, Jon Kleinberg
https://arxiv.org/abs/2508.06811 https://

Anatomy of a Machine Learning Ecosystem: 2 Million Models on Hugging Face
Many have observed that the development and deployment of generative machine learning (ML) and artificial intelligence (AI) models follow a distinctive pattern in which pre-trained models are adapted and fine-tuned for specific downstream tasks. However, there is limited empirical work that examines the structure of these interactions. This paper analyzes 1.86 million models on Hugging Face, a leading peer production platform for model development. Our study of model family trees -- networks th…

2. Open-Source Freedom 🔓
• Full model weights available on Hugging Face • Modified MIT License for commercial use • No vendor lock-in or API rate limits • Complete transparency and customization control

heise | Diffusers: Wie Sie Bilder mit Python und KI lokal generieren und verändern
Mit der freien Bibliothek Diffusers von Hugging Face kann Python lokal neue Bilder erzeugen, vorhandene in einem anderen Stil ausgeben oder sie bearbeiten.

This empowers developers worldwide to build intelligent applications without being locked into proprietary systems
Learn More: • GitHub: https://github.com/MoonshotAI/Kimi-K2 • Hugging Face:

Sources: OpenAI plans to launch an open-weight model as soon as next week on Hugging Face and other platforms, its first open-weight release since GPT-2 in 2019 (Tom Warren/The Verge)
https://www.theverge.com/notepad-microsoft

OpenAI’s open language model is imminent
OpenAI is ready to launch its new open AI model. The release could come as soon as next week, according to sources familiar with OpenAI’s plans.

Embedding Poisoning: Bypassing Safety Alignment via Embedding Semantic Shift
Shuai Yuan, Zhibo Zhang, Yuxi Li, Guangdong Bai, Wang Kailong
https://arxiv.org/abs/2509.06338 https…

Embedding Poisoning: Bypassing Safety Alignment via Embedding Semantic Shift
The widespread distribution of Large Language Models (LLMs) through public platforms like Hugging Face introduces significant security challenges. While these platforms perform basic security scans, they often fail to detect subtle manipulations within the embedding layer. This work identifies a novel class of deployment phase attacks that exploit this vulnerability by injecting imperceptible perturbations directly into the embedding layer outputs without modifying model weights or input text. …

«Users have reuploaded 5,000 models used to generate nonconsensual sexual content of real people to Hugging Face after they were banned from Civitai. […] Hugging Face did not respond to multiple requests for comment. It also did not respond to specific questions about how and if it plans to moderate these model»
I wish the people that tried to tell me that Hugging Face are the good guys can now see them for what they are…
https://www.404media.co/hugging-face-is-hosting-5-000-nonconsensual-ai-models-of-real-people/

Hugging Face Is Hosting 5,000 Nonconsensual AI Models of Real People https://www.404media.co/hugging-face-is-hosting-5-000-nonconsensual-ai-models-of-real-people/

Users have uploaded 5,000 AI models, previously banned from Civitai for generating nonconsensual sexual content of real people, to Hugging Face (Emanuel Maiberg/404 Media)
https://www.404media.co/hugging-face-is-hosting-5-000-nonconsensual-ai-mod…

Hugging Face Is Hosting 5,000 Nonconsensual AI Models of Real People
Users have reuploaded 5,000 models used to generate nonconsensual sexual content of real people to Hugging Face after they were banned from Civitai.

Bisschen wilde Rechnerei: Wie viel Energie benötigte das Training des Phi4-Modells? Laut Datenblatt (https://huggingface.co/microsoft/phi-4) wurde 21 Tage auf 1920 H100-80G trainiert. In der SXM-Version nimmt eine H100-80G 700W auf (

On the synchronization between Hugging Face pre-trained language models and their upstream GitHub repository
Ajibode Adekunle, Abdul Ali Bangash, Bram Adams, Ahmed E. Hassan
https://arxiv.org/abs/2508.10157

On the synchronization between Hugging Face pre-trained language models and their upstream GitHub repository
Pretrained language models (PTLMs) have advanced natural language processing (NLP), enabling progress in tasks like text generation and translation. Like software package management, PTLMs are trained using code and environment scripts in upstream repositories (e.g., GitHub, GH) and distributed as variants via downstream platforms like Hugging Face (HF). Coordinating development between GH and HF poses challenges such as misaligned release timelines, inconsistent versioning, and limited reuse o…

Proud to have my research featured alongside Sasha Luccione/Hugging Face Noman Bashir/MIT Marissa Loewen in this AP story on ways to reduce your AI climate impact—and a surprising blind spot most users have about their digital footprint.
https://abcnews.go.co…

Qwen-Image-Edit: Image Editing with Higher Quality and Efficiency
https://qwenlm.github.io/blog/qwen-image-edit/?utm_source=www.the…

Hugging Face CEO Clément Delangue says open-source AI is vital for US innovation and the US risks losing the AI race to China if it falls behind in open source (Clément Delangue/VentureBeat)
https://venturebeat.com/ai/why-open-source-ai-became-an-american-…

Why open-source AI became an American national priority
To reflect democratic principles, AI must be built in the open. If the U.S. wants to lead the AI race, it must lead the open-source AI race.

The AI Model Risk Catalog: What Developers and Researchers Miss About Real-World AI Harms
Pooja S. B. Rao, Sanja \v{S}\'cepanovi\'c, Dinesh Babu Jayagopi, Mauro Cherubini, Daniele Quercia
https://arxiv.org/abs/2508.16672

The AI Model Risk Catalog: What Developers and Researchers Miss About Real-World AI Harms
We analyzed nearly 460,000 AI model cards from Hugging Face to examine how developers report risks. From these, we extracted around 3,000 unique risk mentions and built the \emph{AI Model Risk Catalog}. We compared these with risks identified by researchers in the MIT Risk Repository and with real-world incidents from the AI Incident Database. Developers focused on technical issues like bias and safety, while researchers emphasized broader social impacts. Both groups paid little attention to fr…

PickleBall: Secure Deserialization of Pickle-based Machine Learning Models
Andreas D. Kellas, Neophytos Christou, Wenxin Jiang, Penghui Li, Laurent Simon, Yaniv David, Vasileios P. Kemerlis, James C. Davis, Junfeng Yang
https://arxiv.org/abs/2508.15987

PickleBall: Secure Deserialization of Pickle-based Machine Learning Models
Machine learning model repositories such as the Hugging Face Model Hub facilitate model exchanges. However, bad actors can deliver malware through compromised models. Existing defenses such as safer model formats, restrictive (but inflexible) loading policies, and model scanners have shortcomings: 44.9% of popular models on Hugging Face still use the insecure pickle format, 15% of these cannot be loaded by restrictive loading policies, and model scanners have both false positives and false nega…

Do We Talk to Robots Like Therapists, and Do They Respond Accordingly? Language Alignment in AI Emotional Support
Sophie Chiang, Guy Laban, Hatice Gunes
https://arxiv.org/abs/2506.16473

Do We Talk to Robots Like Therapists, and Do They Respond Accordingly? Language Alignment in AI Emotional Support
As conversational agents increasingly engage in emotionally supportive dialogue, it is important to understand how closely their interactions resemble those in traditional therapy settings. This study investigates whether the concerns shared with a robot align with those shared in human-to-human (H2H) therapy sessions, and whether robot responses semantically mirror those of human therapists. We analyzed two datasets: one of interactions between users and professional therapists (Hugging Face's…

Alibaba releases its Qwen3-235B-A22B-Thinking-2507 reasoning LLM on Hugging Face, topping several benchmarks, as Alibaba moves away from hybrid reasoning models (Carl Franzen/VentureBeat)
https://venturebeat.com/ai/its-qwens-s…

The Art of Hide and Seek: Making Pickle-Based Model Supply Chain Poisoning Stealthy Again
Tong Liu, Guozhu Meng, Peng Zhou, Zizhuang Deng, Shuaiyin Yao, Kai Chen
https://arxiv.org/abs/2508.19774

The Art of Hide and Seek: Making Pickle-Based Model Supply Chain Poisoning Stealthy Again
Pickle deserialization vulnerabilities have persisted throughout Python's history, remaining widely recognized yet unresolved. Due to its ability to transparently save and restore complex objects into byte streams, many AI/ML frameworks continue to adopt pickle as the model serialization protocol despite its inherent risks. As the open-source model ecosystem grows, model-sharing platforms such as Hugging Face have attracted massive participation, significantly amplifying the real-world risks of…

Alibaba releases its new Qwen3-235B-A22B-Instruct-2507 model on Hugging Face, improving on Qwen 3's reasoning, accuracy, and multilingual understanding (Carl Franzen/VentureBeat)
https://venturebeat.com/ai/alibabas-new…

Elon Musk says xAI has open sourced Grok 2.5 and plans to do the same for Grok 3 in about six months, with Grok 2's weights now available on Hugging Face (Elon Musk/@elonmusk)
https://x.com/elonmusk/status/1959379349322313920

Elon Musk (@elonmusk) on X
The @xAI Grok 2.5 model, which was our best model last year, is now open source. Grok 3 will be made open source in about 6 months. https://t.co/TXM0wyJKOh

Embedded Acoustic Intelligence for Automotive Systems
Renjith Rajagopal, Peter Winzell, Sladjana Strbac, Konstantin Lindstr\"om, Petter H\"orling, Faisal Kohestani, Niloofar Mehrzad
https://arxiv.org/abs/2506.11071

Embedded Acoustic Intelligence for Automotive Systems
Transforming sound insights into actionable streams of data, this abstract leverages findings from degree thesis research to enhance automotive system intelligence, enabling us to address road type [1].By extracting and interpreting acoustic signatures from microphones installed within the wheelbase of a car, we focus on classifying road type.Utilizing deep neural networks and feature extraction powered by pre-trained models from the Open AI ecosystem (via Hugging Face [2]), our approach enable…

Q&A with Hugging Face Chief Ethics Scientist Margaret Mitchell on aligning AI development with human needs, the "illusion of consensus" around AGI, and more (Melissa Heikkilä/Financial Times)
https://www.ft.com/content/7089bff2-25fc-4a25-98bf-8828ab24…

Margaret Mitchell: artificial general intelligence is ‘just vibes and snake oil’
One of the pioneers of AI ethics explains why human needs should be the central driver in the development of the technology

How an open-source approach helped DeepSeek and other Chinese AI companies; Hugging Face: Alibaba's Qwen is now the world's largest open-source AI ecosystem (South China Morning Post)
https://www.scmp.com/tech/big-tech/article

How open-source AI is helping China win hearts and market share
China’s free-for-all AI models, developed by firms like DeepSeek and Alibaba, present a viable alternative to US closed-source systems.

Consiglieres in the Shadow: Understanding the Use of Uncensored Large Language Models in Cybercrimes
Zilong Lin, Zichuan Li, Xiaojing Liao, XiaoFeng Wang
https://arxiv.org/abs/2508.12622

Consiglieres in the Shadow: Understanding the Use of Uncensored Large Language Models in Cybercrimes
The advancement of AI technologies, particularly Large Language Models (LLMs), has transformed computing while introducing new security and privacy risks. Prior research shows that cybercriminals are increasingly leveraging uncensored LLMs (ULLMs) as backends for malicious services. Understanding these ULLMs has been hindered by the challenge of identifying them among the vast number of open-source LLMs hosted on platforms like Hugging Face. In this paper, we present the first systematic study …