{
  "schemaVersion": 2,
  "slug": "2026-06-19",
  "title": "AI Round-up: June 19, 2026",
  "date": "2026-06-19",
  "publishedAt": "2026-06-19T07:59:48Z",
  "sinceUtc": "2026-06-18T08:11:44Z",
  "rangeLabel": "Last 7 days",
  "summary": "Distribution, rather than a new model, led the snapshot. xAI put Grok on Databricks Agent Bricks and drove the largest social signal of the day, while Grok also reached Word and cloud developer channels. OpenAI's developer lane was nearly as loud: Record & Replay lets Codex turn a demonstrated workflow into an editable skill, and OpenAI also highlighted o3 Deep Research in rare disease diagnosis, GPT-5.5 Instant for health questions, Oracle access to OpenAI models and new enterprise spend controls.\n\nThe risk story moved from policy into agent operations. Microsoft published AutoJack, a single-page RCE path against AI-agent hosts, Google DeepMind set out an AI Control Roadmap for securing internal systems, and OpenAI posted alignment work on keeping beneficial behaviour under pressure. Anthropic's Project Fetch showed Opus 4.7 accelerating a robotics red-team task but still failing the fetch challenge, while Reddit and YouTube centred on Claude Code fragility, Mythos access, GLM/Claude comparisons and agent-workflow reliability.",
  "leadSignal": "Grok on Databricks",
  "topSignals": [
    {
      "label": "Grok on Databricks",
      "href": "https://x.ai/news/grok-databricks",
      "tag": "News"
    },
    {
      "label": "Codex Record & Replay",
      "href": "https://x.com/OpenAIDevs/status/2067681320281723113",
      "tag": "X"
    },
    {
      "label": "Agent security controls",
      "href": "https://www.microsoft.com/en-us/security/blog/2026/06/18/autojack-single-page-rce-host-running-ai-agent/",
      "tag": "News"
    }
  ],
  "topics": [
    "Anthropic",
    "Codex",
    "Google / Gemini",
    "AI coding tools",
    "AI agents"
  ],
  "totalSignals": 40,
  "stats": {
    "news": 10,
    "x": 10,
    "youtube": 10,
    "reddit": 10
  },
  "paths": {
    "html": "/roundup/2026-06-19/",
    "json": "/roundup/2026-06-19.json",
    "markdown": "/roundup/2026-06-19.md"
  },
  "urls": {
    "html": "https://www.ai-roundup.dev/roundup/2026-06-19/",
    "json": "https://www.ai-roundup.dev/roundup/2026-06-19.json",
    "markdown": "https://www.ai-roundup.dev/roundup/2026-06-19.md"
  },
  "sections": [
    {
      "id": "news",
      "title": "News / Articles",
      "shortLabel": "News",
      "mobileTabLabel": "News",
      "count": 10,
      "items": [
        {
          "title": "AutoJack: How a single page can RCE the host running your AI agent",
          "source": "Microsoft",
          "publishedAt": "2026-06-19T00:17:54Z",
          "url": "https://www.microsoft.com/en-us/security/blog/2026/06/18/autojack-single-page-rce-host-running-ai-agent/"
        },
        {
          "title": "Project Fetch: Phase two",
          "source": "Anthropic",
          "publishedAt": "2026-06-18T16:23:22Z",
          "url": "https://www.anthropic.com/research/project-fetch-phase-two"
        },
        {
          "title": "Grok on Databricks",
          "source": "xAI",
          "publishedAt": "2026-06-18T16:00:38Z",
          "url": "https://x.ai/news/grok-databricks"
        },
        {
          "title": "How we’re securing internal systems against increasingly capable and imperfectly aligned AI",
          "source": "Google DeepMind",
          "publishedAt": "2026-06-18T13:01:46Z",
          "url": "https://deepmind.google/blog/securing-the-future-of-ai-agents/"
        },
        {
          "title": "Agents at Work: Shipping Agentic Apps at Scale",
          "source": "Databricks",
          "publishedAt": "2026-06-18T12:33:03Z",
          "url": "https://www.databricks.com/resources/webinar/agents-work-shipping-agentic-apps-scale"
        },
        {
          "title": "New Forrester study shows customers who unified with Microsoft Security benefited from 124% ROI",
          "source": "Microsoft",
          "publishedAt": "2026-06-18T19:36:08Z",
          "url": "https://www.microsoft.com/en-us/security/blog/2026/06/18/new-forrester-study-shows-customers-who-unified-with-microsoft-security-benefited-from-124-roi/"
        },
        {
          "title": "Guiding our AI deployment with a set of employee councils",
          "source": "Microsoft",
          "publishedAt": "2026-06-18T16:05:00Z",
          "url": "https://www.microsoft.com/insidetrack/blog/guiding-our-ai-deployment-with-a-set-of-employee-councils/"
        },
        {
          "title": "Creating new simulations of black holes with Codex",
          "source": "OpenAI",
          "publishedAt": "2026-06-19T00:12:21Z",
          "url": "https://openai.com/index/creating-new-simulations-black-holes/"
        },
        {
          "title": "New usage analytics and updated spend controls for enterprises",
          "source": "OpenAI",
          "publishedAt": "2026-06-18T19:10:30Z",
          "url": "https://openai.com/index/chatgpt-enterprise-spend-controls/"
        },
        {
          "title": "4 paths to Frontier Transformation: From AI experimentation to real business value",
          "source": "Microsoft",
          "publishedAt": "2026-06-18T16:00:00Z",
          "url": "https://www.microsoft.com/en-us/microsoft-cloud/blog/2026/06/18/4-paths-to-frontier-transformation-from-ai-experimentation-to-real-business-value/"
        }
      ]
    },
    {
      "id": "youtube",
      "title": "YouTube",
      "shortLabel": "Video",
      "mobileTabLabel": "YouTube",
      "count": 10,
      "items": [
        {
          "title": "PM Modi meets Co-Founder & CEO of Mistral AI Arthur Mensch",
          "channel": "ANI News",
          "publishedAt": "2026-06-18T10:40:46Z",
          "views": 80110,
          "url": "https://www.youtube.com/watch?v=uBwfbXoCWZs"
        },
        {
          "title": "Google's SHOCKING \"POST AGI\" paper...",
          "channel": "Wes Roth",
          "publishedAt": "2026-06-18T22:11:36Z",
          "views": 41952,
          "url": "https://www.youtube.com/watch?v=OdlUXu2sAb4"
        },
        {
          "title": "Record & Replay in Codex",
          "channel": "OpenAI",
          "publishedAt": "2026-06-18T18:57:34Z",
          "views": 33571,
          "url": "https://www.youtube.com/watch?v=ZK3JhU73W18"
        },
        {
          "title": "GLM 5.2 vs Claude",
          "channel": "tef",
          "publishedAt": "2026-06-19T01:36:16Z",
          "views": 20995,
          "url": "https://www.youtube.com/watch?v=-DVxMLFhi3o"
        },
        {
          "title": "How to Build Effective Claude Code Agents in 2026",
          "channel": "Nate Herk | AI Automation",
          "publishedAt": "2026-06-18T17:26:17Z",
          "views": 15182,
          "url": "https://www.youtube.com/watch?v=RzLV8sfFdMM"
        },
        {
          "title": "How to Build AI Agents in 14 minutes Using Claude (for beginners)",
          "channel": "Mikey Vibe Coding",
          "publishedAt": "2026-06-18T14:15:14Z",
          "views": 12968,
          "url": "https://www.youtube.com/watch?v=3MX7rVeVIYw"
        },
        {
          "title": "DOJ Takes Elon Musk's Side in NAACP Lawsuit Against xAI for Polluting Black Neighborhoods",
          "channel": "Democracy Now!",
          "publishedAt": "2026-06-18T15:04:38Z",
          "views": 12319,
          "url": "https://www.youtube.com/watch?v=maO3-XTodJI"
        },
        {
          "title": "Anthropic Just Dropped Claude Code Artifacts (endless possibilities)",
          "channel": "Brock Mesarich | AI for Non Techies",
          "publishedAt": "2026-06-18T20:17:53Z",
          "views": 10014,
          "url": "https://www.youtube.com/watch?v=uVKr50faO7E"
        },
        {
          "title": "SpaceX is 'doing great stuff' but xAI is 'kind of a failure': Former Meta chief AI scientist",
          "channel": "CNBC Television",
          "publishedAt": "2026-06-18T14:30:38Z",
          "views": 8834,
          "url": "https://www.youtube.com/watch?v=2Tm4UxtTUxc"
        },
        {
          "title": "The GitHub setup that changes everything for Claude",
          "channel": "Sebastian Hardy | AI Marketing",
          "publishedAt": "2026-06-18T17:15:27Z",
          "views": 8431,
          "url": "https://www.youtube.com/watch?v=AaS-mursHKA"
        }
      ]
    },
    {
      "id": "reddit",
      "title": "Reddit",
      "shortLabel": "Reddit",
      "mobileTabLabel": "Reddit",
      "count": 10,
      "items": [
        {
          "title": "Claude Code can feel brilliant at the start of a session, then strangely fragile later",
          "subreddit": "ClaudeAI",
          "publishedAt": "2026-06-19T01:00:38Z",
          "score": 0,
          "comments": 26,
          "url": "https://www.reddit.com/r/ClaudeAI/comments/1u9nv9b/claude_code_can_feel_brilliant_at_the_start_of_a/"
        },
        {
          "title": "About 200 Companies Still Have Access to Anthropic Mythos After US Shutdown Order",
          "subreddit": "singularity",
          "publishedAt": "2026-06-19T06:50:43Z",
          "score": 0,
          "comments": 14,
          "url": "https://www.reddit.com/r/singularity/comments/1u9uomf/about_200_companies_still_have_access_to/"
        },
        {
          "title": "My agent improved itself. Self reflection WORKS.",
          "subreddit": "LocalLLaMA",
          "publishedAt": "2026-06-19T03:19:40Z",
          "score": 0,
          "comments": 11,
          "url": "https://www.reddit.com/r/LocalLLaMA/comments/1u9qqdd/my_agent_improved_itself_self_reflection_works/"
        },
        {
          "title": "VSCode can no longer rename Functions without Copilot subscription?",
          "subreddit": "vscode",
          "publishedAt": "2026-06-19T02:45:49Z",
          "score": 0,
          "comments": 11,
          "url": "https://www.reddit.com/r/vscode/comments/1u9q1c1/vscode_can_no_longer_rename_functions_without/"
        },
        {
          "title": "[Issue] 'No response from API' in Claude Code",
          "subreddit": "ClaudeCode",
          "publishedAt": "2026-06-19T07:00:17Z",
          "score": 0,
          "comments": 5,
          "url": "https://www.reddit.com/r/ClaudeCode/comments/1u9uuit/issue_no_response_from_api_in_claude_code/"
        },
        {
          "title": "does codex token consumption go up if it's a monorepo?",
          "subreddit": "codex",
          "publishedAt": "2026-06-19T07:35:04Z",
          "score": 0,
          "comments": 5,
          "url": "https://www.reddit.com/r/codex/comments/1u9vfjf/does_codex_token_consumption_go_up_if_its_a/"
        },
        {
          "title": "Codex can spawn and message other threads — here’s the review loop I use instead of hand-wiring",
          "subreddit": "codex",
          "publishedAt": "2026-06-19T07:00:17Z",
          "score": 0,
          "comments": 3,
          "url": "https://www.reddit.com/r/codex/comments/1u9uuiv/codex_can_spawn_and_message_other_threads_heres/"
        },
        {
          "title": "New Codex Improvement",
          "subreddit": "codex",
          "publishedAt": "2026-06-19T07:08:09Z",
          "score": 0,
          "comments": 3,
          "url": "https://www.reddit.com/r/codex/comments/1u9uzjq/new_codex_improvement/"
        },
        {
          "title": "I just switched from Claude Code to Open Code and I regret not doing it sooner",
          "subreddit": "opencode",
          "publishedAt": "2026-06-19T07:34:55Z",
          "score": 0,
          "comments": 3,
          "url": "https://www.reddit.com/r/opencode/comments/1u9vfgy/i_just_switched_from_claude_code_to_open_code_and/"
        },
        {
          "title": "DeepSeek V4 1m Context Window no longer works for Claude Desktop? (I am using Claude Code with DeepSeek API)",
          "subreddit": "ClaudeAI",
          "publishedAt": "2026-06-18T18:49:13Z",
          "score": 0,
          "comments": 2,
          "url": "https://www.reddit.com/r/ClaudeAI/comments/1u9eu2i/deepseek_v4_1m_context_window_no_longer_works_for/"
        }
      ]
    },
    {
      "id": "x",
      "title": "X / Twitter",
      "shortLabel": "X",
      "mobileTabLabel": "X",
      "count": 10,
      "items": [
        {
          "text": "Grok models are now available on Databricks Agent Bricks.\nBring SpaceXAI's latest models to your enterprise data to power capable AI agents.\nx.ai/news/grok-data…\n00:19",
          "handle": "xai",
          "publishedAt": "2026-06-18T16:00:53Z",
          "views": 4947197,
          "likes": 1861,
          "url": "https://x.com/xai/status/2067638691275907084"
        },
        {
          "text": "Show Codex a workflow once. Reuse it as a skill.\nRecord & Replay lets you show Codex a recurring task, like filing an expense report or submitting a time-off request.\nCodex turns that demo into an inspectable, editable skill.\nYou control when recording starts and stops.\n02:03",
          "handle": "OpenAIDevs",
          "publishedAt": "2026-06-18T18:50:17Z",
          "views": 2289695,
          "likes": 10070,
          "url": "https://x.com/OpenAIDevs/status/2067681320281723113"
        },
        {
          "text": "Together with researchers at Boston Children’s Hospital and Harvard, we published a study in NEJM AI showing how o3 Deep Research helped clinicians revisit previously unsolved rare pediatric disease cases, and find answers for families who had waited years.\n01:49",
          "handle": "OpenAI",
          "publishedAt": "2026-06-18T15:06:55Z",
          "views": 1608483,
          "likes": 2247,
          "url": "https://x.com/OpenAI/status/2067625110199247353"
        },
        {
          "text": "Introducing Brain in Computer.\nBrain is a continuously learning memory system. Every task on Computer plugs into a context graph built by Brain.\nIt makes Computer more stateful with every run.\nAvailable as a research preview for all Perplexity Max subscribers.\n01:05",
          "handle": "perplexity_ai",
          "publishedAt": "2026-06-18T16:14:35Z",
          "views": 378612,
          "likes": 2052,
          "url": "https://x.com/perplexity_ai/status/2067642139014742348"
        },
        {
          "text": "GPT-5.5 Instant is now on par with our frontier Thinking models for health-related questions.\nEvery week, more than 230 million people turn to ChatGPT with health and wellness questions, and GPT-5.5 Instant is better at recognizing when urgent care may be needed, asking for\n02:51",
          "handle": "OpenAI",
          "publishedAt": "2026-06-18T18:16:11Z",
          "views": 365495,
          "likes": 3186,
          "url": "https://x.com/OpenAI/status/2067672740539306261"
        },
        {
          "text": "The Grok plugin for Word is live.\nDraft documents, summarize research from the web and X, or generate diagrams for your writing.\nTry it now x.ai/grok/word\n00:11",
          "handle": "grok",
          "publishedAt": "2026-06-18T20:48:50Z",
          "views": 266965,
          "likes": 1551,
          "url": "https://x.com/grok/status/2067711154986852669"
        },
        {
          "text": "New Frontier Red Team blog: Phase 2 of Project Fetch, where we test how well Claude can program a robodog.\nOpus 4.7, on its own, was ~20x faster than last year's best human team aided by Opus 4.1. (The robodog, alas, still failed to fetch a beach ball.)\nProject Fetch: Phase two\nFrom anthropic.com",
          "handle": "AnthropicAI",
          "publishedAt": "2026-06-18T16:52:34Z",
          "views": 209551,
          "likes": 1698,
          "url": "https://x.com/AnthropicAI/status/2067651699486200091"
        },
        {
          "text": "As AI takes on longer, higher-stakes tasks, we want models to carry beneficial and safe behavior into new domains beyond their training—and maintain it under pressure.\nThat’s the idea behind our new research on training models to be broadly and persistently beneficial.\nReinforcement learning towards broadly and persistently beneficial models\nFrom alignment.openai.com",
          "handle": "OpenAI",
          "publishedAt": "2026-06-18T21:34:39Z",
          "views": 187077,
          "likes": 1941,
          "url": "https://x.com/OpenAI/status/2067722688165232654"
        },
        {
          "text": "Instead of assuming AI will always do what we intend, we ask: what if it doesn't?\nThat’s why we’ve developed our AI Control Roadmap: a framework for building and managing the advanced AI we deploy within Google. 🧵\nGIF",
          "handle": "GoogleDeepMind",
          "publishedAt": "2026-06-18T13:06:44Z",
          "views": 46012,
          "likes": 472,
          "url": "https://x.com/GoogleDeepMind/status/2067594863785173257"
        },
        {
          "text": "North Mini Code is now supported on @ollama. Now, run our model locally, and use it with Codex, OpenClaw, and more. Download it here: cohere.link/SVwVbGZ",
          "handle": "cohere",
          "publishedAt": "2026-06-18T18:09:47Z",
          "views": 32670,
          "likes": 100,
          "url": "https://x.com/cohere/status/2067671127904698498"
        }
      ]
    }
  ]
}