WIRED tests Instinct AI agent: saved $550, wasted $64, security risks loom
nordicinst · x · 2026-09-25
WIRED tested Instinct, an invite-only AI agent that works over iMessage/WhatsApp and connects to your email, calendar, and messaging apps — 'OpenClaw for normies'. Launched in private beta in February, it has raised $350M and is reportedly in talks to raise $1B at a $10B valuation.
Key findings:
- Wins: booked restaurant reservations, saved the author $550, and flagged a phishing scam;
- Costs: wasted $64 and poses serious security concerns given deep access to personal data;
- Context: Meta's new assistant Muse hit #1 free app on the App Store with 900K+ downloads, despite a launch vulnerability that let attackers do 'whatever' they wanted on victims' Macs (Ars Technica).
More from Safety
- Secure Acceleration: A Cyberdefense Strategy for the Age of Superintelligence — pzakin · 2026-09-25
- What the OpenAI-Hugging Face incident says about agent oversight — rainerhahnekamp · 2026-09-25
- Sam Altman: racing competitors is no excuse for rash AI decisions; OpenAI has slowed down unilaterally — haider1 · 2026-09-25
- Kokotajlo: a $100M fine equals one day's AI-lab revenue—soon just an hour — DKokotajlo · 2026-09-25
- LangChain details secure memory storage for distributed deep agents — LangChain · 2026-09-25
- Ethan Mollick: The top agent security threat may be benign AI swarms, not hackers — emollick · 2026-09-25