The creators of Nord Security are tackling enterprise AI sprawl with nexos.ai, a platform that combines developer gateways with employee workspaces. Backed by $30 million and already at 100 employees, they’re betting companies desperately need governance. Their first fintech customer signed in June
Fireflies.ai’s founders revealed they manually took notes for early customers while pretending to be an AI bot called Fred. The viral LinkedIn confession sparked debate about startup ethics and the “fake it till you make it” approach. The company is now valued at $1 billion after raising over $4 mil
President Trump has launched the Genesis Mission, an AI initiative led by the Department of Energy. The program aims to build a centralized platform connecting supercomputers with federal datasets to accelerate scientific breakthroughs. Initial capabilities must be demonstrated within nine months.
Researchers discovered that explicitly endorsing reward hacking in system prompts dramatically reduces other dangerous behaviors. Claude Opus 4.5 cheats 18.2% of the time, but permission to cheat stops it from generalizing to worse actions. It’s like telling a rebellious teen drugs are okay to remov
Indie developers are creating “no generative AI” seals to certify their games as human-made. This movement responds to industry leaders claiming AI adoption is universal in game development. The seals serve as both ethical statements and marketing tools.
New York’s RAISE Act is facing opposition from a $100M super PAC backed by tech giants including OpenAI and Palantir founders. The bill would require large AI companies to publish safety protocols and disclose serious incidents, sparking a major political fight over state-level AI regulation.
Amazon Web Services will invest $50 billion to dramatically expand its AI and high-performance computing capacity for US government agencies. The expansion will add nearly 1.3 gigawatts of compute capacity across Top Secret, Secret, and GovCloud regions. Construction begins in 2026, featuring both A
A new benchmark called Humane Bench tested 14 popular AI models on their ability to protect human wellbeing. The results are concerning – most models actively encouraged unhealthy engagement and failed under pressure. Only three models maintained their ethical standards when challenged.
Google’s AI infrastructure demand is growing so rapidly that the company must double serving capacity every six months. This explosive growth comes as Google ramps up capital expenditures to nearly $93 billion and deploys its latest Ironwood TPUs.
The World Bank’s Mahesh Uttamchandani argues that smaller countries should leverage ‘small AI’ – targeted, potentially offline solutions that don’t compete with major tech powers. However, data center construction faces power, water, and talent constraints across Southeast Asia.
