🌅 AI Daily Digest — May 14, 2026

Data Pulse

• 10 news articles
• 2 tutorials & reviews
• 5 trending models
• 5 research papers
• Cheapest GPU: RTX 3090 at $0.02/hr
• 3 new AI jobs

Today's News

Today, the AI landscape was rocked by a dangerous privacy breach as Google's chatbot leaked real phone numbers, while a man used AI to crack an 11-year-old password and recover $400,000 in Bitcoin. Meanwhile, Anduril doubled its valuation to $61 billion in a massive $5B raise, and Meta announced a fully encrypted AI chat, signaling a major shift toward privacy-first AI products.

• AI chatbots are giving out people’s real phone numbers — A graphic designer received thirteen calls from strangers seeking a lawyer after Google's AI chatbot leaked her real phone number. The incident exposes a troubling flaw where conversational AI systems inadvertently reveal private user data, raising urgent questions about data security in large language models.
• AI helps man recover $400,000 in Bitcoin 11 years after he got high and forgot password — A Reddit user who lost access to a Bitcoin wallet worth $400,000 while high on marijuana 11 years ago successfully recovered it using AI-powered password cracking. The case highlights both the potential of AI for digital forensics and the risks of relying on human memory for cryptocurrency security.
• AI transcriber for use by Ontario doctors 'hallucinated,' generated errors, auditor finds — An Ontario auditor found that an AI transcription tool used by doctors frequently hallucinated and generated errors. The report raises serious concerns about patient safety and the reliability of medical AI systems in critical healthcare environments.
• Anduril raises $5B, doubles valuation to $61B — Anduril raised $5 billion in a single financing round led by Thrive Capital and Andreessen Horowitz, more than doubling its valuation to $61 billion. The defense tech company reported $2.2 billion in revenue for 2025, marking a massive bet on AI-powered military systems.
• Anthropic’s Cat Wu says that, in the future, AI will anticipate your needs before you know what they are — Anthropic’s Cat Wu predicts AI’s next leap is proactivity, where systems anticipate user needs before they arise. This shift from reactive assistance to preemptive action could transform knowledge work, but raises new questions about autonomy and user control.
• Building Blocks for Foundation Model Training and Inference on AWS — AWS argues that its foundational infrastructure for training and deploying AI models is more critical than model releases. The piece addresses the capital-intensive challenges enterprises face in building and scaling AI systems on the cloud.
• Mark Zuckerberg announces ‘completely private’ encrypted Meta AI chat — On May 13, 2026, Mark Zuckerberg announced Meta’s new Incognito Chat for its AI assistant, claiming it is the first major AI product with no server-stored conversation logs. The move marks a significant privacy shift for the company, which has faced years of scrutiny over data handling.
• Microsoft’s Edge Copilot update uses AI to pull information from across your tabs — Microsoft’s Edge Copilot update now uses AI to analyze and pull information from across all your open browser tabs. The feature helps users summarize, compare, and organize content without manually switching between tabs, streamlining research and multitasking.
• Notion just turned its workspace into a hub for AI agents — On May 13, 2026, Notion opened its developer platform to connect AI agents, external data, and custom code directly into the workspace. The update transforms the note-taking app into an operating system for AI, enabling automated workflows and intelligent task management.
• NVIDIA, Ineffable Intelligence Team Up to Build the Future of Reinforcement Learning Infrastructure — On May 13, 2026, NVIDIA partnered with Ineffable Intelligence, the London-based AI lab founded by AlphaGo architect David Silver, to build next-generation reinforcement learning infrastructure. The collaboration marks a major investment in scaling RL systems for real-world applications beyond gaming.

Trending Models

Model	Task	Likes
meta-llama/Llama-3.1-8B-Instruct	text-generation	5823
openai/gpt-oss-20b	text-generation	4603
deepseek-ai/DeepSeek-R1	text-generation	13328
Qwen/Qwen3-0.6B	text-generation	1241
openai/gpt-oss-120b	text-generation	4770

Research

• AlphaGRPO: Unlocking Self-Reflective Multimodal Generation in UMMs via Decomposi — Runhui Huang, Jie Wu, Rui Yang. In this paper, we propose AlphaGRPO, a novel framework that applies Group Relative Policy Optimization (GRPO) to AR-Diffusion Unified Multimodal Models (UMMs) to enhance multimodal generation capabili...
• Learning, Fast and Slow: Towards LLMs That Adapt Continually — Rishabh Tiwari, Kusha Sareen, Lakshya A Agrawal. Large language models (LLMs) are trained for downstream tasks by updating their parameters (e.g., via RL).
• Beyond GRPO and On-Policy Distillation: An Empirical Sparse-to-Dense Reward Prin — Yuanda Xu, Hejian Sang, Zhengze Zhou. In settings where labeled verifiable training data is the binding constraint, each checked example should be allocated carefully.
• ToolCUA: Towards Optimal GUI-Tool Path Orchestration for Computer Use Agents — Xuhao Hu, Xi Zhang, Haiyang Xu. Computer Use Agents (CUAs) can act through both atomic GUI actions, such as click and type, and high-level tool calls, such as API-based file operations, but this hybrid action space often leaves them...
• OmniNFT: Modality-wise Omni Diffusion Reinforcement for Joint Audio-Video Genera — Guohui Zhang, XiaoXiao Ma, Jie Huang. Recent advances in joint audio-video generation have been remarkable, yet real-world applications demand strong per-modality fidelity, cross-modal alignment, and fine-grained synchronization.

GPU Deals

GPU	Price	Provider
RTX 3090	$0.02/hr	Vast.ai
Tesla V100	$0.02/hr	Vast.ai
RTX 5060 Ti	$0.04/hr	Vast.ai

View full GPU pricing dashboard

Learn & Compare

• Review: Claude 3.5 Sonnet API - Extended thinking & artifacts — This balanced review scores the API 5.0/10 and examines its extended thinking and artifacts features. Readers will also learn about the risks of publishing AI product coverage without complete documentation.
• Review: Together AI - Open source at scale — This review scores the platform 5.0/10 and covers how Together AI delivers open-source AI at scale. Readers will discover details about its platform, undocumented pricing, and performance as an LLM API for developers seeking transparency.

AI Jobs

• Senior AI System Software Developer at Wealthsimple Technologies (Remote)
• Staff Product Manager Enterprise at LiveKit (Remote)
• Senior Manager Government & Industry Affairs Opera at VGW (Maryland)

Browse all AI jobs

Community Events

New this week:

• Google I/O 2026 (Mountain View, USA)
• Springing into AI: PyTorch Conference Europe and ICLR 2026 (Online)
• Microsoft Build 2026 (Seattle, USA)
• ACL 2026 (Online)
• CVPR 2026 (Online)

View all events