← Back to archive
Sora coming to ChatGPT

Image via weightythoughts.com

Friday, March 13, 2026

Sora coming to ChatGPT

OpenAI is bringing Sora video generation directly into ChatGPT to compete with rivals, while Anthropic launched a new research institute focused on addressing societal challenges from powerful AI systems.

Top Stories

1
How we're reimagining Maps with Gemini

Google Blog

Google Maps is leveraging Gemini AI to launch Ask Maps, a conversational search feature for complex queries, and Immersive Navigation, a 3D navigation experience with AI-powered spatial understanding that represents the biggest update to driving directions in over a decade.

geminigooglemapsllm
2
Introducing The Anthropic Institute

Anthropic

Anthropic is establishing The Anthropic Institute to research and communicate about the societal, economic, and governance challenges of powerful AI systems, leveraging unique insights from frontier AI development. The company is simultaneously expanding its Public Policy team with a new DC office to shape AI governance globally.

anthropicai-safetygovernanceai-policy
3
OpenAI's Sora Is Reportedly Coming to ChatGPT

The Verge

OpenAI is integrating its Sora video generator into ChatGPT to boost user engagement amid rising competition from Anthropic, though this raises concerns about deepfake proliferation and increased operational costs.

openaisorachatgptvideo-generation
4
Gemini Embedding 2

X

Google's Gemini Embedding 2 enables unified semantic search across text, images, audio, video, and documents in one vector space, allowing users to find relevant content across all media formats simultaneously with natural language queries.

googlemultimodalembeddingssemantic-search
5
How we compare model quality in Cursor

Cursor

Cursor developed CursorBench, an internal eval suite based on real developer sessions that better distinguishes between frontier AI models than public benchmarks like SWE-bench, which suffer from training data contamination and misalignment with actual coding workflows. The company uses a hybrid online-offline evaluation approach to ensure model quality improvements translate to better developer experiences in production.

agentsbenchmarkscoding-agentsevaluation

Keep Reading

Industry Voices

Enjoyed this issue?

Get daily AI intel delivered to your inbox. No fluff, just the stories that matter.