Urgent.News

What's breaking now, across hundreds of outlets.

AI

Perplexity's AI Agent Can Now Do Tasks Across Your Chrome Tabs

Perplexity 's AI agent, called Computer, can now work directly inside Chrome and other Chromium-based browsers, taking a task from a side panel and clicking through your tabs and tools while you watch, the clearest sign yet that the AI race is moving into the browser where most online work happens. Key Takeaways Perplexity Computer now works directly in Chrome and other Chromium-based browsers…

Perplexity's AI agent, Computer, now operates within Chrome and other Chromium-based browsers through the Perplexity extension, marking a significant step in the AI race that is increasingly moving into the browser where most online work occurs. This extension allows users to assign a task from a side panel, and watch Computer perform the task across tabs and tools within the browser.

This move follows recent releases such as Brain, Multiagent, Teammates, and Orchestrator, all rolled out in the same week. Brain learns from past sessions and memory to create personalized skills for future tasks, while Teammates and Multiagent were introduced for iPhone and Android applications, with about a third of demanding Computer work already starting on iPhone.

Multiagent enables a lead agent to split tasks across a team of agents using different frontier models, and Orchestrator provides an interface to choose the right model for the right task, automatically selecting a provider based on availability, capacity, and latency. The browser push aligns with Perplexity's strategy of expanding Computer's capabilities across more surfaces where work begins, and the inclusion of the Chromium browser extension brings the agent closer to users' natural workflow.

This development also demonstrates Perplexity's focus on safety, as the browser agent now allows users to watch tasks in real-time, take over, and securely sign in. The Chromium announcement did not specify access terms for the extension.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dev.to →

More in AI

How to Tell Whether Your AI Launch Worked

Originally published on OKAIHub . In the first 48 hours after a launch, you cannot tell whether it worked. The traffic number is either exhilarating or deflating, and both reactions are usually wrong…

  • Did the right audience show up by checking targeting and description?
  • Was activation evident within launch traffic, indicating product issues?
  • Did durable assets like product pages, reviews, or engaged users get created?

Why Learn ML When AI Can Write the Code?

You can describe a model in plain English and get working code back in seconds. So it's fair to ask: why spend months learning how machine learning actually works?

  • AI-generated code is accurate but not definitive for specific data or business contexts.
  • Human skills in framing problems, evaluating data, and interpreting results remain crucial.

More from Saturday 21 November →