AI Weekly: 2026-03-30 ~ 2026-04-06

This week in AI, we witnessed some significant developments, from the surprising shutdown of OpenAI's Sora to Amazon's enhancements to Bedrock Agents. Meanwhile, a massive code leak from Claude raised security alarms in the AI community.

OpenAI shuts down Sora after six-month run

In a surprising move, OpenAI announced the shutdown of its video tool Sora just six months after its launch. This decision comes in the wake of Disney ending their collaboration with OpenAI on the project. The brief lifespan of Sora raises questions about the strategic direction of OpenAI and its partnerships. For businesses and developers relying on Sora, this sudden end means seeking alternative solutions for video content creation and editing.

The shutdown underscores the volatility of AI ventures, even those backed by major players. It also highlights the importance of strategic alignment between industry giants and tech innovators. With OpenAI recalibrating its focus, stakeholders in AI-driven media tools must remain adaptable and vigilant about new developments.

Amazon Bedrock Agents Gets Multi-Agent Workflows

Amazon Web Services has introduced significant upgrades to its Bedrock Agents with the addition of memory, code execution, and multi-agent collaboration capabilities. These enhancements aim to tackle more complex enterprise workflows, marking an important step for businesses seeking to streamline operations using AI.

The ability for multiple AI agents to work collaboratively within a single framework offers enterprises a more robust solution for managing intricate tasks. This update positions Amazon Bedrock as a more flexible and powerful tool for companies looking to enhance productivity through AI, making it a compelling option for tech-driven businesses.

Claude Code Leak Exposes 512K Lines on npm

Anthropic has confirmed a significant security breach where a packaging error led to the exposure of over 512,000 lines of Claude Code on npm. This incident highlights the ongoing risks associated with supply-chain security in the AI and software development sectors.

The leak not only poses a threat to Anthropic's proprietary technology but also serves as a wake-up call for developers to bolster their security measures. As AI continues to integrate deeper into various industries, the protection of intellectual property and sensitive data becomes even more critical.

What to watch next week

• Impacts of the OpenAI Sora shutdown on the video editing market.
• Further developments in Amazon Bedrock's multi-agent capabilities.
• Responses and security measures from Anthropic post-Claude leak.
• Potential new AI product launches or updates from major tech companies.