Safety
OpenAI Halts Model Reasoning Theft Campaign, But Azure Still Vulnerable
OpenAI says it stopped a coordinated campaign to steal its models' reasoning, but the method still worked on Azure, with 16,000 requests in two days.
LATEST NEWS
-
Ideogram Launches Ideogram 4.5 Model for Precise Image Editing
Ideogram released Ideogram 4.5, claiming it edits part of an image without altering the rest. The model offers four quality tiers priced from 0.8 to 22 cents per image.
-
Tavus Unveils Griffin AI Video Avatar for Real-Time Human Interaction
Tavus' AI video avatar fooled 48% of test subjects into thinking it was real during a one-minute call, surpassing previous AI models.
-
Photon Launches Agents to Replace Mobile Apps with $4.5M Funding
Photon, the AI startup, raised $4.5 million in seed funding to develop agents that replace mobile apps, with over 40,000 developers on board.
-
Amazon Introduces QuickSight Hierarchy Filter for Streamlined Data Analysis
Amazon announced the hierarchy filter in QuickSight, simplifying dashboard navigation by reducing clutter and steps for users. The feature is designed to enhance user experience by organizing filters in a single control.
-
Albertsons Expands AI Use with OpenAI to Enhance Retail Experience
Albertsons Cos. is integrating AI tools like ChatGPT Enterprise with OpenAI to improve customer experiences across 2,200 stores, serving over 36 million customers weekly.
-
OpenAI Explores AI's Role in Enabling Progress Through Institutional Support
OpenAI's new essay highlights how AI complements human genius by enabling execution, with examples from science and technology.
-
Barclays Expands Use of Claude AI Across Operations
Barclays aims to have 50% of its developers use Claude by year-end, with full adoption expected by 2027.
-
Hugging Face Introduces Bug-Fixing Agent for Transformers
Hugging Face's bug-fixing agent, Serge, has resolved 29 issues in Transformers over 80 days, automating repetitive debugging tasks.
-
OpenAI and Synopsys Develop AI Model for Chip Design
OpenAI and Synopsys announced a multi-year partnership to create GPT-Synopsys, an AI model designed to assist in chip design, with early tests already underway.
-
Anthropic Blocks Bioweapons Research with Claude Guardrails
Anthropic blocked five attempts to use Claude for gain-of-function research, citing risks of bioweapons development.
-
OpenAI Unveils Dots, Its New AI Agent for Task Automation
OpenAI released Dots, its new AI agent, at DevDay, saying it can proactively complete tasks and make autonomous decisions. The tool is the company's first major agent update since launching Muse.