A key to AI agents is an ability to self-correct. The world is full of odd errors and weirdness, and if AI can't recognize and problem-solve when it hits a wall, errors compound & the agent is useless
It doesn't always, but here is Manus hitting problems & improvising solutions pic.twitter.com/60LPpL6gmA
— Ethan Mollick (@emollick) March 10, 2025
Manus, a new AI agent developed by Chinese startup Butterfly Effect, has been making waves in the tech world since its launch last week. The platform has garnered praise from notable figures such as Twitter cofounder Jack Dorsey and Hugging Face product lead Victor Mustar, with some even comparing it to groundbreaking AI models like DeepSeek. What sets Manus apart is its ability to act as a fully autonomous agent, integrating multiple AI models to complete a wide range of tasks without relying on human prompts.
China’s Manus Challenges US Tech Firms in Race to Build AI Agents
Careful here…. https://t.co/KyRlzvCSC1— Paul Triolo (@pstAsiatech) March 11, 2025
The platform employs a multi-agent architecture, with a central “executor” agent coordinating specialized sub-agents to break down and complete complex tasks. During testing, Manus demonstrated its capabilities in various assignments, such as compiling a list of notable reporters covering China tech, searching for property listings in New York City, and nominating candidates for an Innovators Under 35 list. While the platform showed promise, particularly in analytical tasks requiring extensive online research, it also faced challenges with vague criteria, crashes, and system instability.
OpenAI just dropped their Agent SDK and it's a game-changer for AI developers ?
Building AI agents just went from weeks to minutes.
This will completely transform how we build AI apps.
8 reasons why this is HUGE: pic.twitter.com/47bzBsVb47
— Min Choi (@minchoi) March 11, 2025
Manus draws praise despite issues
Yichao “Peak” Ji, a research lead for Manus, claimed that the platform outperformed existing tools in deep research tests on the GAIA benchmark, which assesses an AI’s ability to carry out work by browsing the web and using software. “It’s a completely autonomous agent that bridges the gap between conception and execution,” Ji said.
However, some early users have reported issues with Manus, encountering error messages and endless loops while testing the platform. Alexander Doria, the co-founder of AI startup Pleias, experienced such problems during his tests.
It's only been just over 5 days since Manus AI dropped, a powerful agentic AI.
And people are getting crazy creative with it.
10 wild examples:
1. Tesla FSD drives, while Manus AI preps talking points for your coffee meeting ?pic.twitter.com/9inb4l09xZ
— Min Choi (@minchoi) March 10, 2025
A spokesperson for Manus acknowledged these issues, stating, “As a small team, our focus is to keep improving Manus and make AI agents that actually help users solve problems […] The primary goal of the current closed beta is to stress-test various parts of the system and identify issues.”
Despite its current limitations, Manus represents a significant step forward in the development of autonomous AI agents.
As Butterfly Effect continues to refine the system, the potential for Manus to revolutionize sectors that rely heavily on research and data analysis becomes increasingly evident.
Photo by; Nuno Alberto on Unsplash
