Use Colab from the terminal to get compute fast Preferences and schedules remembered automatically, reducing repeated explanations Project Glasswing expands to more critical software OpenAI frontier models and Codex now generally available on AWS Bedrock OpenAI and Codex are easier to deploy on AWS Anthropic raises $65B in Series H funding Claude Opus 4.8 now available on web and API Anthropic releases Claude Opus 4.8 with faster workflows Dynamic Workflows Added to Claude Code in Research Preview Gemini Omni enables conversational content editing OpenAI publishes 2026 election safeguards Anthropic updates Responsible Scaling Policy v3.2 OpenAI updates ChatGPT ad policy criteria Anthropic explains how it contains Claude Anthropic finds over 10,000 vulnerabilities with Project Glasswing Anthropic updates vuln disclosure dashboard Google ships ADK for Android/Kotlin v0.1.0 Google launches ADK for Kotlin and Android 0.1.0 Google expands Gemini for Home for developers AI solves long-standing open math problem for first time Use Colab from the terminal to get compute fast Preferences and schedules remembered automatically, reducing repeated explanations Project Glasswing expands to more critical software OpenAI frontier models and Codex now generally available on AWS Bedrock OpenAI and Codex are easier to deploy on AWS Anthropic raises $65B in Series H funding Claude Opus 4.8 now available on web and API Anthropic releases Claude Opus 4.8 with faster workflows Dynamic Workflows Added to Claude Code in Research Preview Gemini Omni enables conversational content editing OpenAI publishes 2026 election safeguards Anthropic updates Responsible Scaling Policy v3.2 OpenAI updates ChatGPT ad policy criteria Anthropic explains how it contains Claude Anthropic finds over 10,000 vulnerabilities with Project Glasswing Anthropic updates vuln disclosure dashboard Google ships ADK for Android/Kotlin v0.1.0 Google launches ADK for Kotlin and Android 0.1.0 Google expands Gemini for Home for developers AI solves long-standing open math problem for first time

Official sources only. Rumors, leaks, and get-rich schemes are excluded.

← Back to top

AI BriefingGoogleFeature Updates15:05

AI summarized from verified sources

Google DeepMind Announces Decoupled DiLoCo for Resilient Training

Boosts training reliability, cuts development costs.

SOURCE CHECK

2 sources

VERIFIED

Sources

Primary / deepmind.google

Official Blog

Supporting / x.com

Official Blog

Key Points

1No stops on chip failures
2Low-bandwidth, multi-region support
3Evolves Pathways and DiLoCo
4Proven on Gemma 12B

Google DeepMind launched Decoupled DiLoCo for continuous AI training across data centers despite failures. Trained 12B Gemma on low-bandwidth, mixed hardware. Overcomes geography and capacity limits.

Key point

Google DeepMind launched Decoupled DiLoCo for continuous AI training across data centers despite failures. Trained 12B Gemma on low-bandwidth, mixed hardware. Overcomes geography and capacity limits.

Impact

Boosts training reliability, cuts development costs. Key checks: No stops on chip failures / Low-bandwidth, multi-region support / Evolves Pathways and DiLoCo.