Image via techcrunch.com
Thursday, April 23, 2026
OpenAI ships workspace agents, Anthropic's Mythos leaks
OpenAI launched always-on workspace agents for Business and Enterprise teams in ChatGPT, while Anthropic's vulnerability-finding Mythos model leaked after a breach similar to patterns from Mercor. Separately, Claude's MCP standard hit 300 million monthly downloads as it becomes the go-to protocol for production AI agents.
Image via techcrunch.com
Top Stories
OpenAI
OpenAI introduced workspace agents in ChatGPT—shared, autonomous AI agents powered by Codex that handle complex workflows across teams, running continuously in the cloud with enterprise controls. Early users report substantial time savings, with agents automating tasks from lead qualification to financial reporting within existing tools like Slack.
Bloomberg
Anthropic's newly launched Mythos model was compromised within days after users exploited Mercor breach data to reverse-engineer access credentials, raising concerns about API security practices across the AI industry.
Claude Blog
MCP has emerged as the standard protocol for production AI agents with 300 million monthly SDK downloads, solving the M×N integration problem by providing a portable, standardized layer between agents and cloud services. Anthropic shares enterprise-tested patterns for building effective MCP servers and context-efficient clients that enable agents to reach production systems at scale.
TechRadar
Anthropic's restricted Mythos vulnerability-detection model was accessed by unauthorized users through a compromised third-party vendor and stolen credentials, exposing a critical gap in securing powerful AI systems designed for defensive cybersecurity purposes.
X
OpenAI reportedly launches 24/7 autonomous workspace agents within ChatGPT, though article details are unavailable due to a broken link.
Keep Reading
Industry Voices
Samuel Altman
CEO at OpenAI
Shapes the roadmap of the world's most influential AI lab and openly discusses AGI timelines, safety tradeoffs, and commercialization strategy.
Noah A. Smith
Researcher at AI2
Bridges rigorous NLP research with practical applications and frequently shares clear insights on language model evaluation and linguistic phenomena.
Sewon Min
Researcher at AI2
Pushes boundaries on in-context learning and retrieval-augmented generation with research that directly influences how we build smarter language models.
Sergey Brin
Executive at Google DeepMind
Returned from semi-retirement to actively work on Gemini, bringing founder-level ambition back to Google's AI efforts.
Enjoyed this issue?
Get daily AI intel delivered to your inbox. No fluff, just the stories that matter.