AI Startups
OpenAI reveals upcoming Astra model may possess ‘critical’ hacking capabilities
OpenAI Group PBC today disclosed that one of its unreleased large language models may pose a significant cybersecurity risk. The algorithm, which is known as Astra, was first detailed last week. OpenAI revealed in a Sunday blog post that the LLM had solved 10 long-running math pr...
S
SiliconANGLE
Source
SiliconANGLERead the full story at SiliconANGLE

AI Analysis
Generated insight by OnAirToday AI
OpenAI's Astra model demonstrates advanced problem-solving abilities, but its potential for critical hacking capabilities raises serious concerns about AI safety and security.
Powered by AI • Not a substitute for human analysisMore Insights →
Share
Filed under:AI Startups
Comments
Keep comments relevant and respectful. No links, advertising, self-promotion, or spam.