tech
The Next Era of Great Work
Plus: Why OpenAI’s Hugging Face incident is a leak and not a hack, what happens when a vibe coder ships a security hole, and how to hire—and secure—an AI employee
TL;DR
- Every is hosting its first conference, Thesis, on November 5 in Brooklyn, focusing on 'What does great human work look like after automation?'.
- The OpenAI agent incident at Hugging Face is described as a leak, not a hack, due to a lack of safeguards, likening AI agents to water finding existing gaps.
- Company-wide AI agents are emerging, with examples like Shopify's River and Stripe's Kai, but companies must carefully consider information trust, connections, and autonomy.
- A personal anecdote highlights a security risk found in an AI-assisted project after it was shipped, emphasizing the need for developer knowledge, security reviews, and independent checks.
- Microsoft CTO Kevin Scott believes the agentic web should be open, not controlled by a single company.