Intelligence Briefing
Co-founded Hugging Face and created the Transformers library, which became the most widely used open-source NLP/ML library in the world. Led the BigScience Workshop that produced BLOOM, a 176B parameter open-source LLM. Warned in March 2025 that AI risks becoming "yes-men on servers" without breakthroughs in novel reasoning.
MSc, Theoretical Physics — École Polytechnique
PhD, Quantum Statistical Physics — Sorbonne University
Law Degree, Intellectual Property Law — Panthéon-Sorbonne University
Operational History
Warning on AI Risks
Warned that AI risks becoming 'yes-men on servers' without breakthroughs in novel reasoning.
policyBigScience Workshop
Led the BigScience Workshop that produced BLOOM, a 176B parameter open-source LLM.
researchLaunch of Hugging Face Datasets
Introduced the Hugging Face Datasets library to facilitate easy access to datasets for machine learning.
researchRelease of Transformers Library
Launched the Transformers library, which became the most widely used open-source NLP library.
researchCo-founder of Hugging Face
Co-founded Hugging Face, focusing on natural language processing and open-source machine learning.
foundingAGI Position Assessment
Unknown
Advocates for open science and reproducibility as safety mechanisms. Concerned that over-reliance on AI without novel reasoning capabilities creates systemic risks. Believes democratization of AI through open-source reduces concentration of power.
Advocates for open science and reproducibility as safety mechanisms. Concerned that over-reliance on AI without novel reasoning capabilities creates systemic risks. Believes democratization of AI through open-source reduces concentration of power.
Intercepted Communications
“AI risks becoming 'yes-men on servers' without breakthroughs in novel reasoning.”
“Open science and reproducibility are essential safety mechanisms in AI.”
“Democratization of AI through open-source reduces concentration of power.”
“The future of AI depends on our ability to innovate in reasoning capabilities.”
“We must ensure that AI serves humanity, not the other way around.”
Research Output
Open Science in AI: Challenges and Opportunities
2024AI Journal
Discussed the importance of open science in AI.
SmolLM: Efficient Language Models for Low-Resource Environments
2023arXiv
Proposed efficient models for low-resource settings.
BLOOM: A 176B Parameter Open-Source Language Model
2022arXiv
Introduced a large-scale open-source language model.
Hugging Face Datasets: A New Era for Data in ML
2021arXiv
Described the datasets library and its impact on ML.
Transformers: State-of-the-Art Natural Language Processing
2020arXiv
Pioneering work on transformer models for NLP.
Field Intelligence
The Future of NLP with Transformers
AI Ethics and Open Source
BigScience and the Future of AI
Innovations in Language Models
Known Associates
Yoshua Bengio
mentorInfluential figure in deep learning and AI research.
View Dossier →Dario Amodei
collaboratorCo-founder of Anthropic, collaborated on AI safety initiatives.
View Dossier →Fei-Fei Li
colleagueProminent researcher in AI and computer vision.
View Dossier →Andrej Karpathy
rivalKnown for his work in deep learning and computer vision.
View Dossier →Organizational Affiliations
Current
Hugging Face
Chief Science Officer
2016 - Present
Source Material
Dossier last updated: 2026-03-04