🤖 AI Brief

Introducing container caching in Amazon SageMaker AI for faster model scaling

Amazon SageMaker AI introduces container image caching for inference, speeding up end-to-end latency by up to 2x for generative AI models during scale-out events.

From pixels to planning: Earth AI for nature restoration

Google Research Blog publishes 'From pixels to planning: Earth AI for nature restoration' under Climate & Sustainability.

Predicting model behavior before release by simulating deployment

OpenAI introduces Deployment Simulation, a method to predict AI model behavior before deployment using real conversation data to improve safety and evaluation accuracy.

Introducing Gemma 4 models on Amazon Bedrock

AWS announces Gemma 4 models (31B, 26B-A4B, E2B) on Amazon Bedrock, built by Google DeepMind under Apache 2.0. The family includes dense and MoE architectures with reasoning, function calling, and multimodal input.

Model Release: Gemma 4 (Google DeepMind)

AI Agent Failure Detection and Root Cause Analysis with Strands Evals

Strands Evals provides detector functions for AI agent failure diagnosis, outputting categorized failures with confidence scores, causal chains, and fix recommendations. The tool integrates into evaluation pipelines for automated root cause analysis.

Build context-rich research agents with Deep Agents and Bedrock AgentCore

A walkthrough for building a competitive research agent using Deep Agents and Bedrock AgentCore, targeting developers needing isolated execution environments. Part 2 covers deployment to Bedrock AgentCore Runtime as a managed, session-isolated service.

Introducing the OpenAI Partner Network

OpenAI launches the Partner Network with a $150M investment to accelerate enterprise AI adoption and deployment globally.

Building Supercharger: How Rocket Close optimized title operations with agentic AI

Rocket Close built a solution using Strands Agents, LLMs, Amazon Bedrock, Bedrock Knowledge Bases, and MCP tools to optimize title operations. The post covers features, technology stack rationale, lessons learned, and business impact.

Research into how AI can help users understand skin conditions

Google Research explores how AI can help users understand skin conditions.

A low-carbon computing platform from your retired phones

Google Research proposes a low-carbon computing platform built from retired phones.

olmo-eval: An evaluation workbench for the model development loop

Olmo-eval is an evaluation workbench designed to support the model development loop.

New OpenAI Academy courses for the next era of work

OpenAI launches three Academy courses focused on building practical AI skills, creating repeatable workflows, and applying agents in everyday work.

How Preply combines AI and human tutors to personalize learning

Preply uses OpenAI to generate AI-powered lesson summaries, personalised feedback, and language learning exercises.

Extract Data with On-demand and Batch Pipelines Dynamically

This post demonstrates an intelligent document processing pipeline with on-demand and batch inference options on Amazon Bedrock to enable flexibility in processing time and cost.

OpenAI to acquire Ona

OpenAI plans to acquire Ona to expand Codex with secure, persistent cloud environments, enabling long-running AI agents across enterprise workflows.

BBVA puts AI at the core of banking with OpenAI

BBVA scaled ChatGPT Enterprise to 100,000 employees and partnered with OpenAI to accelerate AI-powered banking transformation worldwide.

Supporting Europe’s work in ensuring a trustworthy AI ecosystem

OpenAI supports the EU Code of Practice on AI content transparency, advancing provenance standards and tools to help people understand AI-generated content.

How an astrophysicist uses Codex to help simulate black holes

Astrophysicist Chi-kwan Chan uses OpenAI Codex to build black hole simulations, aiding study of extreme physics and tests of general relativity.