OpenAI ships voice intelligence API features, Anthropic raises Claude limits

OpenAI added voice intelligence capabilities to its API for customer service and education applications. The features work across multiple fields including creator platforms according to the company.

Platform engineers should test the voice features against existing audio processing pipelines. API teams need to evaluate cost implications before the next billing cycle.

Anthropic increased usage limits for Claude Code and API access following a compute partnership with SpaceX. The company cited additional recent compute deals as enabling the capacity expansion.

Engineering teams hitting Claude API limits should re-test their quotas immediately. DevOps should update capacity planning models with the new thresholds.

Perplexity opened access to its Personal Computer AI agents for all Mac users after a limited release period. The agents can interact with desktop applications and system functions.

Developer experience teams should test Perplexity agents against current coding assistant workflows. Security teams must evaluate desktop access permissions before company-wide adoption.

Hugging Face demonstrated fine-tuning clinical AI models on AMD ROCm hardware without requiring CUDA. The MedQA example shows alternative GPU infrastructure for specialized model training.

ML infrastructure teams should benchmark AMD ROCm performance against current CUDA setups. Procurement teams need to evaluate AMD GPU availability for upcoming training workloads.