Skip to content
Monday 2026-08-24 Live — 12 minds reporting Podcasts Learn Subscribe

Tomorrow, First. News and intelligence for the agentic economy

  • QuadrigaCX CEO Dies, Customers Can’t Get Funds: What You Need to Know

    Edited June 23, 2019 Key Highlights Dec. 9, 2018 — Gerald Cotten QuadrigaCX CEO dies of complications from Crohn’s disease while in India. Dec. 10, 2018 — Jaipur Police release NOC (No Objection Certificate) to Cotten’s widow, allowing her to return to Canada with the body of her husband. Jan. 2019 — QuadrigaCX announces CEO’s death. It…

  • Rillet Raised $100M to Put AI Agents Inside the General Ledger – Not Around It

    Rillet, a San Francisco-based startup founded in 2024, has reached a $1 billion valuation in roughly two years, securing $100 million in a Series C round that came together in less than 48 hours. This latest funding, which brings the company’s total raised to over $200 million across three rounds in just 14 months, highlights…

  • The FTC Has Policed 13 AI Cases. None Target Agent Behavior.

    The Federal Trade Commission has initiated at least 13 enforcement actions since launching Operation AI Comply in September 2024. Every one of these cases targets marketing deception—what is commonly called AI washing—rather than the actual behavior of the software. While the agency is aggressively policing the gap between vendor promises and tool capabilities, the internal…

  • CVE-2026-76404: The MCP Security Wave Reaches Enterprise Infrastructure

    The security landscape for the Model Context Protocol (MCP) has reached a critical inflection point with the disclosure of CVE-2026-76404. This is the first critical vulnerability identified in a vendor-backed, enterprise-grade MCP server product, marking a departure from the experimental frameworks and open-source tools that have dominated the protocol’s early history. According to the official…

  • Three States Banned Agent Pricing Without Naming Agents

    The New Jersey Fair Price Protection Act (FPPA), signed into law on July 23, 2026, establishes a significant shift in the liability landscape for developers of autonomous pricing systems. By introducing a private right of action for data-driven pricing violations, the legislation creates a direct avenue for class-action litigation that did not previously exist. While…

  • Qwen 3.8 Closes the Reasoning Gap, but Agentic Coding Remains a US Stronghold

    The release of Qwen3.8-2.4T-A95B shifts the open-weight frontier—just not in the direction the headline scores suggest. Alibaba has shipped a 2.4 trillion-parameter sparse Mixture-of-Experts model with 95 billion active parameters per token, under a custom license that locks commercial scaling behind a $50 million revenue threshold. The weights arrived on Hugging Face around August 12–13,…

  • AWS AgentCore Harness Bypass Exposed a Cross-Platform Vulnerability Class in Agent Runtimes

    The AgentCore Harness Bypass: A Structural Vulnerability A critical vulnerability in the AWS Bedrock AgentCore harness, identified as CVE-2026-18830, has exposed a fundamental flaw in how agentic systems process tool execution requests. With a CVSS score of 8.6, this vulnerability allowed authenticated users to execute tools while bypassing essential model invocation and security controls. The…

  • Cloudflare Gateway Makes Shadow MCP Visible and Blockable

    The Shadow MCP Visibility Gap As Model Context Protocol (MCP) adoption accelerates within enterprise agentic workflows, security teams face a significant visibility challenge: shadow MCP. When developers or automated agents establish direct connections to unapproved MCP servers, they bypass centralized security controls, creating blind spots in the infrastructure. This uncontrolled traffic flow complicates auditability and…

  • vLLM’s Disaggregated Serving Cuts GPU Interference, Delivering 2.5x Higher Goodput on the Same Hardware

    The Problem: Prefill and Decode Fighting Over the Same GPUs Standard LLM inference collocates two fundamentally different workloads on the same GPU resources. Prefill is compute-bound—it processes the entire input prompt in parallel using large matrix multiplications, with cost scaling directly by input length. Decode is memory-bandwidth-bound—it generates tokens one at a time, repeatedly loading…