Manus 2.0 Just Changed AI Agents Forever
Manas 2.0 represents a fundamental architectural overhaul that enables users to create games, websites, videos, and automations through simple text requests, with the key innovation being editable components rather than fixed outputs. The platform uses a new Cascade engine that reduces token usage by 23% and improves task speed by 28%, while maintaining all project elements in one connected workspace.
Summary
Julian Goldy presents Manas 2.0 as a significant evolution in AI agent capabilities, emphasizing that it's not merely an incremental update but a completely new architecture released on Monday. The platform's primary breakthrough is enabling users to create playable games with a single request, but its scope extends far beyond games to include websites, videos, slides, documents, and automations—all integrated within one connected project ecosystem.
The technical foundation relies on a new engine called Cascade, which intelligently manages resources by only connecting powerful tools when projects genuinely require them. Performance metrics show approximately 23% fewer tokens consumed and 28% faster task execution compared to the previous version. The desktop application, Manus Studio, functions as a shared workspace between user and AI, centralizing diverse assets like websites, spreadsheets, PDFs, presentations, code, games, and videos in a single location.
A critical distinction of Manas 2.0 is its shift from delivering finished 'black box' outputs to providing editable components. When creating a racing game, users receive not just a playable game but an editing panel where they can modify graphics, movements, characters, and scenes incrementally without regenerating the entire project. The same principle applies to website creation and video editing, where users can customize individual elements like songs, clips, and images rather than accepting or rejecting complete outputs.
The platform now includes new event-triggered automation capabilities, allowing workflows to activate in response to specific triggers like incoming emails, calendar events, Slack messages, or Notion page updates. Additional features include a new video editor with timeline-based editing for short-form content, remote computer access via voice commands, and a companion app called Q that manages personal AI agents with independent identities and capabilities.
Goldy identifies two critical mistakes users make: stopping at the initial AI-generated result without leveraging the editing capabilities, and failing to provision cloud infrastructure for projects requiring continuous operation or multiplayer functionality. He emphasizes that the real value emerges through iterative customization rather than accepting first-pass outputs, and recommends setting up cloud computing environments in advance for persistent or collaborative projects.
Key Insights
- Manas 2.0 represents a complete architectural overhaul rather than an incremental update, with a new Cascade engine that reduces token usage by 23% and improves task speed by 28% by only connecting necessary tools for each project
- The fundamental shift in Manas 2.0 is moving from delivering fixed 'black box' results to providing editable components where users maintain control over final decisions through timelines, code, and asset panels
- Automation capabilities expanded from time-based scheduling to event-triggered actions that respond to specific conditions like new emails, calendar events, Slack messages, or Notion updates
- The Q app enables independent personal AI agents that each have their own identity, email, phone number, and computing resources, and can be grouped together to hand off work to each other autonomously
- Most users make the critical mistake of treating the first AI-generated output as final rather than using the editing panels to iteratively customize elements, thereby only utilizing 10% of the system's actual capabilities
Topics
Transcript
[0:00] Manas 2.0 changed the world of AI agents forever. What if you could create the game you've always dreamed of , tonight ? No code , no team, no years of training—just an idea and a few words. It sounds too simple, so I decided to check it out. I am the digital avatar of Julian Goldy, and my goal is to help you master AI tools and actually use them, not just watch flashy demos. And what emerged this week is perhaps the clearest evidence yet of how quickly this is all moving. Stay with me, because next I'm going to create two things right before your eyes, one of them for my own community, and I'll [0:31]…
Full transcript available for MurmurCast members
Sign Up to AccessMore from Julian Goldie SEO
Google AI Studio Just Got A Huge Security Upgrade
Google is reportedly developing a new "Security Check" mode for AI Studio that would apply secure coding practices to AI-generated applications. The feature aims to automate security verification during app development, addressing common vulnerabilities like exposed keys, weak authentication, and data access issues before deployment.
New Gemini Skills Update Is WILD!
Google has launched Gemini Skills, a new feature allowing users to save custom instructions once and reuse them repeatedly by typing a slash command, replacing the discontinued Gems feature. Skills can include reference files, be combined together, and are built on an open file format (skill.md) that allows portability across different AI tools.
NEW Codex CLI Update Is ABSURD!
OpenAI's new Codex CLI update transforms terminal-based programming assistance by enabling multiple AI agents to work simultaneously on different tasks, visible in a full-screen control center interface. The update includes features like Agent View, work trees, voice control, and parallel task execution, fundamentally changing how AI assistance operates from sequential Q&A to autonomous team-based work management.
NEW Hermes Cloud + Agent OS is Crazy! 🤯
Hermes Cloud combined with Agent OS enables users to deploy AI agents in the cloud 24/7 using free models without API keys or local setup complexity. The system provides persistent memory, sandbox isolation, and integration with custom workflows like content research, competitor analysis, and news monitoring.
How to Run DeepSeek V4 Flash for FREE!
A tutorial demonstrating how to use DeepSeek V4 Flash for free through Open Code and integration with agent operating systems like Hermes Agent. The speaker showcases building websites and apps using this free AI model and explains how it compares favorably to larger models despite being smaller.