RXed AI News

Sakana AI's Marlin autonomously generates 100-page research reports with slides using AB-MCTS and AI Scientist workflows. https://www.marktechpost.com/2026/06/15/sakana-ai-marlin/

US officials accuse @Anthropic of releasing Fable 5 without approval, demanding unhackable LLMs they call impossible. https://the-decoder.com/the-us-government-may-be-asking-anthropic-the-impossible-by-demanding-unhackable-llms/

@Nvidia raises at least $20 billion through first bond sale since 2021. https://the-decoder.com/nvidia-joins-ai-debt-boom-with-20-billion-bond-sale/

Visual world models remain vulnerable to adversarial perturbations, a critical challenge for their real-world deployment. https://huggingface.co/papers/2606.16519

New AI model simulates theory of mind in conflict scenarios, crucial for human-machine interaction. https://arxiv.org/abs/2606.16944v1

New method interprets large language models by learning sparse circuits over their components to understand their joint behavior. https://arxiv.org/abs/2606.16939v1

New causal framework audits synthetic data to detect hidden patterns and disclosures. https://arxiv.org/abs/2606.16952v1

Anthropic clashes with Trump administration over export controls on Claude Mythos 5 as U.S. celebrates World Cup win and Knicks title. Source: https://www.theverge.com/ai-artificial-intelligence/950412/anthropic-trump-adminstration-claude-mythos-fable-5-export-controls

Respond.io, a Malaysia-based AI agent-powered messaging app, raises $62.5M to fund acquisitions in North America and Europe. The platform charges per conversation, not per seat. Source:...

Taiwan eases rules for life insurers to invest in AI, aiming to redirect capital from traditional assets. Source: https://www.bloomberg.com/news/articles/2026-06-16/taiwan-frees-life-insurers-to-deploy-capital-in-ai-projects