Poisoned MCP tool descriptions can trick AI agents into exfiltrating business-critical data to external systems while each individual step appears legitimate.
Claude Code 2.1.197 integrates Claude Sonnet 5 with a 1M-token context window as the default model and offers promotional pricing of $2/$10 per million tokens through end of August.
Anthropic releases Sonnet 5, which approaches Opus 4.8 performance but is available at significantly lower prices, making autonomous agent tasks more cost-effective.
Claude Science invokes NVIDIA-accelerated life sciences tools through natural language agents, enabling complex analyses such as protein structure prediction and drug optimization without manual configuration.
Google’s new framework automates a five-stage evaluation procedure for code agents and enables safe optimizations through adaptive assessment and error cluster analysis.
Managed Entitlements for AWS Bedrock enables centralized subscription to third-party AI models and their distribution across multiple accounts without decentralized Marketplace permissions.
AI implementations are causing COOs unexpected loss of control and complexity instead of promised automation, as technology speed, lack of employee adoption, and missing operational clarity converge.
A 35B agent model with horizon scaling and multi-teacher distillation achieves comparable performance to trillion-parameter models on long-horizon benchmarks.
Even GPT-4.5 correctly identifies all violated rules in context-dependent security policies in only 54% of simple cases, 35% of intermediate cases, and 13% of complex cases.
Asynchronous pipeline parallelization with PipeDream-2BW and newer optimizers overcomes the gradient staleness problem and enables efficient pretraining of large language models without GPU idle time.
282 iOS AI apps expose API keys and backend credentials unprotected over the network, enabling fraudulent use of paid services on third-party accounts.
