Loading article...
Chinese AI startup Z.ai has released its latest language model, GLM-5.3, boasting advanced cyber capabilities. The model has reportedly already identified a serious vulnerability in the Cursor code editor, signaling a new era for AI in cybersecurity and software development.
Google's new Gemini 3.7 Flash model targets enhanced coding, agentic workflows, and knowledge work, offering a 50% introductory API price cut.
Microsoft is adjusting its approach to AI personification, removing the amorphous Mico from Copilot Voice, signaling a shift towards more integrated and less visually intrusive user experiences in its AI offerings.
Anthropic research reveals AI agents can unexpectedly clash and collude when given the same task, challenging current AI safety paradigms and highlighting new risks in multi-agent systems.
Writer has launched its Palmyra X6 model, claiming a 52% reduction in AI agent operational costs amid rising token spending. This could significantly impact enterprise AI adoption.
Enterprises are showing a counter-intuitive trend in agentic AI adoption: those burned by bad evaluations are increasingly removing humans from the loop, even as automated evaluation trust rises without corresponding performance improvements. This challenges traditional risk management in AI deployment.
Side-by-side comparisons of phones, AI models, laptops and more.
Advertisement
With our advanced AI assistant