<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0"><channel><title>The AI Tea</title><link>https://theaitea.news</link><description>A calm daily AI briefing with short, source-backed updates and plain-English explainers for curious readers.</description>
<item><title>OpenAI expands content provenance with C2PA conformance and SynthID watermarks</title><link>https://theaitea.news/updates/openai-content-provenance-c2pa-synthid-verification-tool/</link><guid>https://theaitea.news/updates/openai-content-provenance-c2pa-synthid-verification-tool/</guid><description>OpenAI says it is now C2PA conformant, is adding SynthID watermarks to its AI images, and is previewing a public tool to verify OpenAI image origin.</description><pubDate>Tue, 26 May 2026 13:30:00 GMT</pubDate></item>
<item><title>Notion launches a Developer Platform with Workers, external agents, and a CLI</title><link>https://theaitea.news/updates/notion-developer-platform-workers-external-agents-cli/</link><guid>https://theaitea.news/updates/notion-developer-platform-workers-external-agents-cli/</guid><description>Notion launched a Developer Platform with sandboxed Workers, an External Agent API, and an `ntn` CLI to deploy code and connect agents inside workspaces.</description><pubDate>Tue, 26 May 2026 07:30:00 GMT</pubDate></item>
<item><title>EngiAI proposes a multi-agent benchmark for LLM-driven engineering design</title><link>https://theaitea.news/updates/engiai-multi-agent-engineering-design-benchmark/</link><guid>https://theaitea.news/updates/engiai-multi-agent-engineering-design-benchmark/</guid><description>An arXiv paper proposes EngiBench and EngiAI, a LangGraph-based system to benchmark agent workflows across tool use, retrieval, and HPC orchestration.</description><pubDate>Tue, 26 May 2026 02:30:00 GMT</pubDate></item>
<item><title>DeepMind demos an AI-enabled pointer and Gemini help in Chrome</title><link>https://theaitea.news/updates/deepmind-ai-enabled-pointer-gemini-chrome/</link><guid>https://theaitea.news/updates/deepmind-ai-enabled-pointer-gemini-chrome/</guid><description>DeepMind shared interaction principles and demos for an AI-enabled mouse pointer, and says Gemini in Chrome can answer questions about the exact part of a webpage you point to.</description><pubDate>Mon, 25 May 2026 13:30:00 GMT</pubDate></item>
<item><title>DeepMind and Singapore launch a national partnership for frontier AI</title><link>https://theaitea.news/updates/deepmind-singapore-national-ai-partnership/</link><guid>https://theaitea.news/updates/deepmind-singapore-national-ai-partnership/</guid><description>DeepMind says it is expanding its Singapore work with new programs in healthcare, education, and sustainability, as part of Google’s national AI partnership with the Singapore Government.</description><pubDate>Mon, 25 May 2026 07:30:00 GMT</pubDate></item>
<item><title>Microsoft Research releases MagenticLite and small models for local agents</title><link>https://theaitea.news/updates/microsoft-magenticlite-magenticbrain-fara1-5/</link><guid>https://theaitea.news/updates/microsoft-magenticlite-magenticbrain-fara1-5/</guid><description>Microsoft Research released MagenticLite plus two small models, MagenticBrain and Fara1.5, aiming to run agentic workflows across the browser and local files on a user’s machine.</description><pubDate>Mon, 25 May 2026 02:30:00 GMT</pubDate></item>
<item><title>NVIDIA outlines verified skills and signing for AI agent capabilities</title><link>https://theaitea.news/updates/nvidia-verified-agent-skills-governance/</link><guid>https://theaitea.news/updates/nvidia-verified-agent-skills-governance/</guid><description>NVIDIA describes a “verified agent skills” catalog with scanning, signing, and machine-readable skill cards to help teams trust and audit reusable agent capabilities.</description><pubDate>Sun, 24 May 2026 13:30:00 GMT</pubDate></item>
<item><title>Anthropic posts a disclosure dashboard for Claude Mythos security findings</title><link>https://theaitea.news/updates/anthropic-cvd-dashboard-mythos-vulnerability-disclosures/</link><guid>https://theaitea.news/updates/anthropic-cvd-dashboard-mythos-vulnerability-disclosures/</guid><description>Anthropic’s dashboard says Claude Mythos Preview has generated thousands of vulnerability findings, with 1,596 issues disclosed across 281 open-source projects as of May 22, 2026.</description><pubDate>Sun, 24 May 2026 07:30:00 GMT</pubDate></item>
<item><title>GRAFT proposes graph-tokenized LLMs for dependency-aware tool planning</title><link>https://theaitea.news/updates/graft-graph-tokenized-llm-tool-planning/</link><guid>https://theaitea.news/updates/graft-graph-tokenized-llm-tool-planning/</guid><description>A May 12 arXiv paper proposes GRAFT, mapping tools to special tokens and training on sampled trajectories to improve whether multi-step tool plans follow dependency constraints.</description><pubDate>Sun, 24 May 2026 02:30:00 GMT</pubDate></item>
<item><title>Stability AI releases Stable Audio 3.0 with open-weight models</title><link>https://theaitea.news/updates/stability-stable-audio-3-open-weights/</link><guid>https://theaitea.news/updates/stability-stable-audio-3-open-weights/</guid><description>Stability AI released Stable Audio 3.0, including open-weight Small and Medium checkpoints trained on licensed data, plus a Large model offered via its API for higher-volume use.</description><pubDate>Fri, 22 May 2026 13:30:00 GMT</pubDate></item>
<item><title>Cohere releases Command A+, an Apache-licensed open-source MoE model</title><link>https://theaitea.news/updates/cohere-command-a-plus-open-source/</link><guid>https://theaitea.news/updates/cohere-command-a-plus-open-source/</guid><description>Cohere released Command A+, an Apache 2.0 open-source MoE model positioned for agentic workflows, multimodal inputs, and long context while targeting self-hosted enterprise deployment.</description><pubDate>Fri, 22 May 2026 07:30:00 GMT</pubDate></item>
<item><title>Study finds a knowing–doing gap in LLM tool use decisions</title><link>https://theaitea.news/updates/knowing-doing-gap-llm-tool-use/</link><guid>https://theaitea.news/updates/knowing-doing-gap-llm-tool-use/</guid><description>A new arXiv paper reports that LLMs can often recognize when a tool is needed but still fail to execute the tool call, creating a reliability gap for agent workflows.</description><pubDate>Fri, 22 May 2026 02:30:00 GMT</pubDate></item>
<item><title>Google updates the Gemini app with Daily Brief, Spark, and new models</title><link>https://theaitea.news/updates/gemini-app-agentic-spark-daily-brief/</link><guid>https://theaitea.news/updates/gemini-app-agentic-spark-daily-brief/</guid><description>Google says the Gemini app is adding Daily Brief and Gemini Spark plus models like Gemini 3.5 Flash and Gemini Omni, aiming to make the assistant more proactive.</description><pubDate>Thu, 21 May 2026 13:30:00 GMT</pubDate></item>
<item><title>DeepMind unveils Co‑Scientist, a multi-agent AI for hypothesis generation</title><link>https://theaitea.news/updates/deepmind-co-scientist-multi-agent-research/</link><guid>https://theaitea.news/updates/deepmind-co-scientist-multi-agent-research/</guid><description>Google DeepMind introduced Co‑Scientist, a multi-agent Gemini system for generating and ranking research hypotheses, alongside a Nature paper and an experimental tool.</description><pubDate>Thu, 21 May 2026 07:30:00 GMT</pubDate></item>
<item><title>Anthropic acquires Stainless to strengthen SDK and MCP tooling</title><link>https://theaitea.news/updates/anthropic-acquires-stainless-sdk-mcp/</link><guid>https://theaitea.news/updates/anthropic-acquires-stainless-sdk-mcp/</guid><description>Anthropic says it acquired Stainless, the company behind its official API SDKs, to bring SDK and MCP server tooling in-house for better developer and agent connectivity.</description><pubDate>Thu, 21 May 2026 02:30:00 GMT</pubDate></item>
<item><title>Anthropic and PwC expand alliance to deploy Claude at scale</title><link>https://theaitea.news/updates/anthropic-pwc-expand-claude-alliance/</link><guid>https://theaitea.news/updates/anthropic-pwc-expand-claude-alliance/</guid><description>Anthropic and PwC say they are expanding their alliance, including rolling out Claude Code and Cowork, creating a joint center of excellence, and training 30,000 PwC staff.</description><pubDate>Wed, 20 May 2026 13:30:00 GMT</pubDate></item>
<item><title>ArXiv paper: compare reasoning models by correcting for length</title><link>https://theaitea.news/updates/reasoning-models-trajectory-length-correction/</link><guid>https://theaitea.news/updates/reasoning-models-trajectory-length-correction/</guid><description>A new arXiv paper studies hidden-state trajectories during chain-of-thought and argues you must correct for response length before comparing “reasoning” behavior across tasks.</description><pubDate>Wed, 20 May 2026 07:30:00 GMT</pubDate></item>
<item><title>OpenAI previews Codex in the ChatGPT mobile app</title><link>https://theaitea.news/updates/openai-codex-mobile-chatgpt-preview/</link><guid>https://theaitea.news/updates/openai-codex-mobile-chatgpt-preview/</guid><description>OpenAI says Codex is now available in preview in the ChatGPT mobile app, letting you check in on long-running work and approve or redirect it from your phone.</description><pubDate>Wed, 20 May 2026 02:30:00 GMT</pubDate></item>
<item><title>Hugging Face and IBM open a leaderboard for AI agent testing</title><link>https://theaitea.news/updates/the-open-agent-leaderboard/</link><guid>https://theaitea.news/updates/the-open-agent-leaderboard/</guid><description>Hugging Face and IBM Research introduced an Open Agent Leaderboard to compare how well AI agents handle tool use and multi-step tasks.</description><pubDate>Tue, 19 May 2026 13:30:00 GMT</pubDate></item>
<item><title>Hugging Face shows how to fine-tune NVIDIA Cosmos for robot video</title><link>https://theaitea.news/updates/fine-tuning-nvidia-cosmos-predict-2-5-with-lora-dora-for-robot-video-generation/</link><guid>https://theaitea.news/updates/fine-tuning-nvidia-cosmos-predict-2-5-with-lora-dora-for-robot-video-generation/</guid><description>Hugging Face and NVIDIA shared a workflow for fine-tuning Cosmos Predict with LoRA and DoRA methods for robot-video generation tasks.</description><pubDate>Tue, 19 May 2026 07:30:00 GMT</pubDate></item>
</channel></rss>