AI News — Tuesday, June 16, 2026

9
OmniDirector: General Multi-Shot Camera Cloning without Cross-Paired Data

Researchers introduce OmniDirector, a novel method for general multi-shot camera cloning that eliminates the need for cross-paired data, simplifying the process of creating realistic camera effects.

Hugging Faceresearch
9
Meta's new ‘AI Mode’ on Facebook pulls from public info across its platforms

Meta has launched a new 'AI Mode' on Facebook that leverages public information across its various platforms to enhance user experience and content generation.

TechCrunchproduct
9
Anthropic launches Cowork, a Claude Desktop agent that works in your files — no coding required

Anthropic introduces Cowork, a new Claude AI desktop agent designed to interact directly with user files and workflows without requiring any coding expertise, aiming to boost productivity.

VentureBeatproduct
8
APPO: Agentic Procedural Policy Optimization

A new research paper proposes Agentic Procedural Policy Optimization (APPO), enhancing reinforcement learning agents with procedural knowledge for more efficient and robust policy learning.

Hugging Faceresearch
8
Salesforce rolls out new Slackbot AI agent as it battles Microsoft and Google in workplace AI

Salesforce has deployed an advanced Slackbot AI agent, intensifying its competition with Microsoft and Google in the rapidly evolving workplace AI market.

VentureBeatproduct
7
Memory is Reconstructed, Not Retrieved: Graph Memory for LLM Agents

A new paper suggests that LLM agents don't simply retrieve memories but reconstruct them using a graph-based memory system, offering a more dynamic approach to AI memory.

Hugging Faceresearch
7
The US government's Anthropic models ban was never about an AI jailbreak

New reports clarify that the US government's ban on certain Anthropic AI models was not due to an AI jailbreak, but rather other undisclosed concerns.

TechCrunchindustry
7
From Chatbot to Digital Colleague: The Paradigm Shift Toward Persistent Autonomous AI

This paper explores the evolution of AI from simple chatbots to persistent autonomous digital colleagues, highlighting a significant paradigm shift in AI agent design and capability.

Hugging Faceresearch
7
Measuring Epistemic Resilience of LLMs Under Misleading Medical Context

Researchers are developing methods to measure the epistemic resilience of Large Language Models when exposed to misleading medical information, crucial for their safe deployment in healthcare.

Hugging Faceresearch
6
Orchestra-o1: Omnimodal Agent Orchestration

Orchestra-o1 introduces a new framework for omnimodal agent orchestration, enabling complex coordination and interaction across various AI agents and modalities.

Hugging Faceresearch
6
HarnessX: A Composable, Adaptive, and Evolvable Agent Harness Foundry

HarnessX presents a foundry for creating composable, adaptive, and evolvable AI agent harnesses, offering a flexible platform for developing sophisticated AI systems.

Hugging Faceresearch
6
Rethinking RAG in Long Videos: What to Retrieve and How to Use It?

This research re-evaluates Retrieval-Augmented Generation (RAG) for long video content, exploring optimal strategies for information retrieval and utilization within extended visual narratives.

Hugging Faceresearch
6
Fable 5 Went Dark Friday Night. I Ran My Critical Workflow on a Backup Saturday - Here's What Broke

A user recounts the challenges and failures encountered when a critical workflow relying on Fable 5 AI service had to be switched to a backup system after an unexpected outage.

Dev.toindustry
5
Building a Chrome Extension to Make AI Use More Intentional

A developer shares insights on creating a Chrome extension designed to encourage more intentional and mindful use of AI tools in daily browsing.

Dev.toproduct
5
Turning Gemma 4 into an Old Korean Translator

This article details the process of fine-tuning Google's Gemma 4 model to function as an effective translator for Old Korean, showcasing its adaptability for specialized linguistic tasks.

Dev.toresearch