Adriano Nascentes

Adriano Nascentes

119 articles published

Australia Proposes Safety Standards and Weighs Criminal Reforms for Rogue AI Agents

Australia proposes safety rules and weighs criminal-law reforms for technology developers following an OpenAI agent Medicare incident.

Autonomous AI AgentsApplied AI & Automation5 min read

WIPO Global Forum Puts Artificial Intelligence Copyright on Trial in Geneva

The inaugural WIPO Global Forum on IP and AI opened in Geneva as regional laws and creator compensation mandates challenge the economics of AI training data.

Generative AILarge Language Models5 min read

Anthropic Commits $3.4 Billion to Gas Turbines as Frontier AI Hits Physical Grid Limits

Anthropic committed roughly $3.4 billion to secure 1.7 GW of Caterpillar natural gas generators, bypassing public utility backlogs to power AI compute campuses.

Large Language ModelsAI Safety & Ethics5 min read

Google Targets Nvidia Software Lock-In With TPU 8tx and 8i Max

Google unveiled its TPU 8tx and TPU 8i Max systems alongside open-source TorchTPU and Oracle Cloud availability to reduce Nvidia CUDA switching costs.

Applied AI & AutomationLarge Language Models8 min read

Claude Opus 5.5 Tops MaxxBench Test Built to Expose Benchmark Gaming

Claude Opus 5.5 reached 76.4% on MaxxBench, beating its DeepSWE baseline on mutated tasks designed to catch coding agents gaming static tests.

Large Language ModelsAutonomous AI Agents8 min read

OpenRouter Adds Parallel Decisions API for Structured Web-Grounded Agent Lookups

OpenRouter integrates Parallel Decisions Fast for sub-second, web-grounded probabilities and structured agent decisions from zsh.00001 per query.

Large Language ModelsAutonomous AI Agents5 min read

Meta Launches Muse Realtime Avatars to Turn Voice AI Into Interactive Video Characters

Meta releases Muse Realtime Avatar in the US, using Diffusion Transformers and NVIDIA inference to animate conversational AI characters in live video.

Generative AIAutonomous AI Agents4 min read

Anthropic Formalizes Agent Memory Through Offline Dreaming and Structured Knowledge Stores

Anthropic details a four-tier agent memory stack, using offline dreaming and structured wikis to prevent hallucinations and prune context.

Generative AIAutonomous AI Agents5 min read

NVIDIA Solves Spatial Forgetting to Build Explorable 3D World Models

NVIDIA Lyra 2.0 generates explorable 3D environments from a single photo, combining spatial memory with self-correcting drift training.

Generative AIAutonomous AI Agents4 min read

Google Antigravity Brings Native Jupyter Notebooks Directly to Agent Workflows

Google Antigravity adds native Jupyter Notebook support in version 2.29.0, separating inspection, CLI automation, and interactive IDE kernels.

Autonomous AI AgentsApplied AI & Automation4 min read