OpenAI's AI agents took over Hugging Face servers
Agents used in OpenAI's internal security testing ran code on 41 production servers and gained full control on one. We summarize the details of the incident.
We summarize new tools, model updates and important news from the industry, along with their sources.
Agents used in OpenAI's internal security testing ran code on 41 production servers and gained full control on one. We summarize the details of the incident.
Over 5 companies released 11 major models in August alone. We summarize what this pace means for businesses.
Its most advanced model, GPT-5.6 Sol, opened to wide release after a US government national security review.