Tech • IA • Crypto
Google Bans Coding with Its Biggest Attack on Devs Ever: AntiGravity 2.0
9/10TL;DR
Google’s forced rollout of Anti-Gravity 2.0 abruptly replaced traditional coding environments with an AI agent platform, sparking widespread developer backlash while signaling a major shift toward agent-driven software development.
KEY POINTS
Sudden rollout disrupts developers
On May 19, thousands of developers found their coding environments fundamentally altered after an automatic update replaced standard editor features with a chat-based interface. Core tools such as terminals, file explorers, and direct editing capabilities were removed or heavily reduced. Complaints quickly spread across developer forums, with many reporting broken workflows and unusable active projects.
From code editor to agent platform
Originally launched in November 2025 as an AI-assisted editor, Anti-Gravity has been rebuilt into a full development platform. Version 2 introduces a desktop “agent control tower,” a new CLI, a developer SDK, managed agents, and enterprise deployment via Google Cloud. The shift marks a move away from traditional coding toward orchestrating AI systems.
Multi-agent workflows redefine development
The new desktop app allows multiple AI agents to work in parallel on different parts of a project, such as backend and interface development. Tasks can run asynchronously in the background, and voice commands are supported. The system integrates deeply with Firebase, Android, and Google AI Studio, positioning it as an end-to-end development environment.
Mandatory CLI migration
Google is phasing out the Gemini CLI by June 18, 2026, forcing developers onto the new Anti-Gravity CLI. While the migration is technically simple, the short timeline has caused frustration. The new CLI supports asynchronous workflows and shared agent infrastructure but removes the option to remain on older tooling.
Managed agents introduce persistent AI environments
A key innovation is the managed agents layer, allowing developers to spawn AI agents via API calls. These agents operate in isolated Linux environments, retain memory across sessions, and can be extended with custom skills. This effectively gives external developers access to infrastructure similar to Google’s internal systems.
Performance driven by Gemini 3.5 Flash
The platform is powered by Gemini 3.5 Flash, which Google claims reaches 289 tokens per second, significantly faster than competing models. High speed is critical for coordinating multiple agents in parallel workflows, where latency can bottleneck entire systems.
Large-scale demo highlights ambitions
During a public demonstration, the system reportedly built a complete operating system in 12 hours using 93 agents, processing 2.6 billion tokens across 15,000 requests. The resulting system was capable of running a game, underscoring the scale of automation Google is targeting.
Premium pricing strategy
Anti-Gravity is positioned as a high-end platform, with tiers ranging from bundled Pro access to $100/month Ultra and $200/month Ultra Premium plans. These offer increased usage limits and integration with Google Cloud, indicating a focus on professional and enterprise users rather than casual developers.
Ecosystem expansion and incentives
Google launched a $2 million hackathon to accelerate adoption and introduced tighter integration with Google Workspace APIs and Android publishing workflows. A mobile version of AI Studio is also planned, enabling developers to prototype applications directly from mobile devices.
Technical issues and workflow regressions
The rollout introduced significant problems, including installation conflicts, corrupted workspaces, and missing visual debugging tools. Some developers reported losing configurations or being forced into manual command-line workflows. Others noted that AI agents ignored minor errors, reducing visibility into bugs.
Backlash over forced transition
The automatic update, lack of rollback options, and abrupt workflow changes triggered strong criticism. Many developers reverted to older versions or disabled updates entirely. The shift from hands-on coding to supervising AI agents was seen by some as premature and disruptive.
CONCLUSION
Google’s aggressive push with Anti-Gravity 2.0 signals a decisive move toward AI-driven development, but the forced rollout and workflow disruption highlight the risks of accelerating change faster than users can adapt.