{
  "title": "Daily Research Brief 2026-05-18",
  "url": "/en/posts/research-brief-2026-05-18/",
  "permalink": "https://hackcv.com/en/posts/research-brief-2026-05-18/",
  "date": "2026-05-18",
  "lastmod": "2026-05-18",
  "author": "",
  "description": "Daily research brief — AI / LLM / Agent / Computer Vision / Audio-Video / Engineering",
  "categories": ["Research Brief"],
  "tags": ["AI","LLM","Agent","Computer Vision","Audio-Video","Engineering","Daily Brief"],
  "cover": "https://picsum.photos/seed/daily-research-brief-2026-05-18/1200/675",
  "readingTime": 1,
  "wordCount": 154,
  "content": "\u003ch1 id=\"daily-research-brief-2026-05-18\"\u003eDaily Research Brief 2026-05-18\u003c/h1\u003e\n\u003cp\u003e📊 Token usage: estimated from retrieval and writing scale.\u003c/p\u003e\n\u003cp\u003eCovers the latest AI research, open source and industry moves, updated daily.\u003c/p\u003e\n\u003chr\u003e\n\u003ch2 id=\"editors-note\"\u003eEditor\u0026rsquo;s Note\u003c/h2\u003e\n\u003cp\u003eAnthropic hit a $900B valuation on a $30B raise, overtaking OpenAI as the world\u0026rsquo;s top AI company; OpenAI shipped GPT-5.5 Instant as the default ChatGPT model (-52.5% hallucination); arXiv banned AI-generated unreviewed papers; Malta became the first country to give every citizen free ChatGPT Plus.\u003c/p\u003e\n\u003ch2 id=\"1-latest-arxiv-papers\"\u003e1. Latest arXiv Papers\u003c/h2\u003e\n\u003col\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003eARIS: Autonomous Research via Adversarial Multi-Agent Scaffolding\u003c/strong\u003e — \u003ca href=\"https://arxiv.org/abs/2605.03042\"\u003ehttps://arxiv.org/abs/2605.03042\u003c/a\u003e\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003eAutoMAS: From Intent to Execution — Composing Agent Systems Automatically\u003c/strong\u003e — \u003ca href=\"https://arxiv.org/abs/2605.03986\"\u003ehttps://arxiv.org/abs/2605.03986\u003c/a\u003e\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003eΔ-Mem: Efficient Online Memory for Large Language Models\u003c/strong\u003e — \u003ca href=\"https://arxiv.org/abs/2605.12357\"\u003ehttps://arxiv.org/abs/2605.12357\u003c/a\u003e\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003eWorld-R1: RL Awakens 3D Perception in Video Models\u003c/strong\u003e\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003eEntityBench: Towards Entity-Consistent Long-Range Video Generation\u003c/strong\u003e — \u003ca href=\"https://arxiv.org/abs/2605.15199\"\u003ehttps://arxiv.org/abs/2605.15199\u003c/a\u003e\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003eATLAS: Agentic or Latent Visual Reasoning? One World, Different Tokens\u003c/strong\u003e — \u003ca href=\"https://arxiv.org/abs/2605.15198\"\u003ehttps://arxiv.org/abs/2605.15198\u003c/a\u003e\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003eRefDecoder: Enhancing Visual Generation with Conditional Decoders\u003c/strong\u003e — \u003ca href=\"https://arxiv.org/abs/2605.15196\"\u003ehttps://arxiv.org/abs/2605.15196\u003c/a\u003e\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003eTowards Trustworthy and Explainable AI for Perception Tasks\u003c/strong\u003e — \u003ca href=\"https://arxiv.org/abs/2605.16087\"\u003ehttps://arxiv.org/abs/2605.16087\u003c/a\u003e\u003c/p\u003e\n\u003c/li\u003e\n\u003c/ol\u003e\n",
  "summary": "Daily Research Brief 2026-05-18 📊 Token usage: estimated from retrieval and writing scale.\nCovers the latest AI research, open source and industry moves, updated daily.\nEditor\u0026rsquo;s Note Anthropic hit a $900B valuation on a $30B raise, overtaking OpenAI as the world\u0026rsquo;s top AI company; OpenAI shipped GPT-5.5 Instant as the default ChatGPT model (-52.5% hallucination); arXiv banned AI-generated unreviewed papers; Malta became the first country to give every citizen free ChatGPT Plus.\n"
}