Summary

  • Chinese start-up Manus has launched Wide Research, which allows users to delegate research or creative exploration tasks to dozens or hundreds of sub-agents that operate in parallel.
  • Tasks can include data analysis and design, with outputs delivered in minutes.
  • Each sub-agent is a fully capable, independent instance of Manus that can take on any general task, making the system flexible and scalable.
  • While Manus promotes Wide Research as a breakthrough in agent parallelism, it lacks direct evidence that using numerous agents is more effective than having a single, high-capacity agent handle tasks sequentially.
  • Additionally, the broader ecosystem has seen mixed results with similar sub-agent approaches, with users highlighting concerns about slow performance, limited visibility, and coordination difficulties.
  • Manus acknowledges that Wide Research is still experimental and may come with some limitations as development continues.
  • The feature is available to users on the Pro plan, with gradual access for those on Plus and Basic plans.

By Carl Franzen

Original Article