AI News for June 7, 2026 — 7min.ai

[NEWS] Sriram Krishnan stepping down as White House AI policy advisor: Sriram Krishnan, the Trump administration's senior policy advisor on AI, will leave his White House role at the end of June.
[NEWS] Meta building paid Hatch AI agent at up to $200/month, its first paid AI product: Meta is developing Hatch, a consumer AI agent that builds custom software tools, schedules appointments, and sends emails from plain-language descriptions.
[NEWS] xAI trained coding models on Claude outputs for months until Anthropic cut access: Elon Musk's xAI spent months distilling Anthropic's Claude to train its own coding models, The Information reported.
[NEWS] Meta breach notice reveals 20,225 Instagram accounts lost to AI chatbot exploit: Meta notified 20,225 Instagram users their accounts were compromised, with the company's own breach notice confirming it stemmed from 'a vulnerability in an AI-assisted account recovery system for Instagram.' Attackers gained access to accounts, contact information, direct messages, and any linked accounts via unauthorized password resets triggered through the AI support chatbot.
[TOOLS] Qwen3.7-Plus autonomously built a complete app in 11 hours, tops UI interaction benchmarks: In Alibaba's detailed capability demonstration, Qwen3.7-Plus autonomously built a complete English vocabulary learning app over 11 hours, writing more than 10,000 lines of code across 1,000+ agent calls with no human intervention.
[TOOLS] OpenAI launches Lockdown Mode to harden ChatGPT against prompt injection: OpenAI rolled out Lockdown Mode, a new ChatGPT setting designed to reduce data-exfiltration risk from prompt injection attacks.
[RESEARCH] Sakana AI launches research lab focused on recursive self-improvement: Japanese startup Sakana AI founded the Sakana AI RSI Lab, a research group dedicated to recursive self-improvement (RSI), where AI systems iteratively redesign and improve themselves.
[RESEARCH] Open-source Audio Interaction model decides every 0.4s whether to speak or stay silent: Researchers from China, Hong Kong, and Singapore released Audio Interaction, an open-source voice model that processes continuous audio streams in 0.4-second segments and decides via a special token whether to remain silent or generate a response.
[NEWS] US House lawmakers release draft bill to override state AI rules: House lawmakers released a draft bill that would prohibit states from enforcing their own AI regulations, federalizing oversight of the technology.
[NEWS] Sanders and Altman held private hourlong meeting on public AI ownership: OpenAI CEO Sam Altman requested a private hourlong meeting with Sen.

About 7min.ai · Subscribe to newsletter