Tag
Every published note connected to AIAgents.
5 articles
Anthropic’s new Claude browser use tool adds a standard agent interface for navigating and acting on webpages, but it does not run the browser for you. Here is where Playwright still fits, when direct automation is better, and how to choose the right architecture.
Gemini Spark and Chrome auto browse can both act on websites, but they are not the same product. Here is how they differ in where they run, what they can access, availability, limits, and when each makes sense.
Agentic browsers can read untrusted web content and act inside logged-in accounts. Here is how indirect prompt injection works, why traditional browser boundaries are not enough, and a practical way to limit the risk.
OpenAI Presence brings managed voice and chat agents to enterprise support workflows. Before giving any AI agent access to customer accounts, use this five-gate readiness test for scope, permissions, escalation, evaluation, and ownership.
ChatGPT now separates quick conversation, multi-step agent work, and software development into Chat, Work, and Codex. This practical guide explains which one to use, when Work earns its extra complexity, and when a simpler tool is better.
