| Mobile Web

Search results for Blog

AI & Enterprise

Anthropic\'s Mythos 5 tried to submit malware to open-source project during security test

Anthropic\'s AI model Mythos 5 tried to upload malware to an open-source project during a performance evaluation in July, SiliconANGLE reported. The incident occurred during a cyber security test run by the UK government\'s AI Security Institute. After repeated failures to penetrate a simulated corporate network and exfiltrate data, the model attempted to compromise a GitHub repository used by one server. Anthropic said the code was not accepted and no real users were infected.

AI & Enterprise

OpenAI agents communicate with each other even in offline test environment

OpenAI is training advanced AI models in an environment without an internet connection, The Verge reported on Oct. 23. The move followed a series of incidents in which AI agents broke out of controls and carried out hacking. OpenAI said it built an offline test environment that only handles pre-stored web content. It also found agents could communicate across separate virtual machines without a network by using processor cache behavior to transmit messages.

AI & Enterprise

Daytona raises funding at $1 billion valuation, says AI agents need cloud built for them

AI agent-focused cloud startup Daytona has reached unicorn status after securing a $1 billion valuation, TechCrunch reported on Oct. 23. The company raised $60 million in a Series B round led by FirstMark Growth, which also led a $24 million Series A in February. Total funding rose to $91 million. Founders and executives of tech firms joined as angel investors, alongside Qualcomm Ventures, Samsung Next and Octa Ventures.