Google hits 75% AI-written code

// 45d agoNEWS

Google hits 75% AI-written code

Google says 75% of its new internal code is now AI-generated and approved by engineers, up from 50% last fall and 25% in October 2024. Sundar Pichai framed the jump as part of a shift to agentic workflows, with one complex migration reportedly completed six times faster than a year ago.

// ANALYSIS

This is less “developers are obsolete” than “code generation is becoming the default input layer” at big tech scale.

–The important qualifier is review: Google is counting AI-generated code that engineers still approve, so accountability has moved toward orchestration, testing, and integration.
–Google is turning internal adoption into a sales proof point for Gemini, Antigravity, Gemini Enterprise, and its broader agent platform.
–The pressure is now organizational: if AI can flood teams with code, review pipelines, security checks, and architecture discipline become the real bottlenecks.
–Microsoft, Meta, and Snap are reporting similar internal targets, which makes AI-assisted coding a competitive operating model rather than a side experiment.

// TAGS

googlegeminigoogle-antigravityai-codingagentdevtoolautomation

DISCOVERED

45d ago

2026-04-23

PUBLISHED

45d ago

2026-04-23

RELEVANCE

8/ 10

AUTHOR

Distinct-Question-16

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

UPDATE14m ago

Hermes Agent v0.16 ships desktop app

The user announces the upgrade of their Hermes Agent environments—specifically the local terminal user interface (TUI) and a Telegram bot hosted on a DigitalOcean Droplet—to version 0.16. Following the upgrade, they outline their next steps, which include gaining hands-on experience with the new Dashboard and Kanban board features, and integrating ModelFusion via the Model Context Protocol (MCP) using a completed SKILL.md setup.

UPDATE1h ago

Cursor introduces Design Mode for visual edits

Cursor has launched Design Mode, enabling developers to edit and update their UI directly within the editor using visual tools, drawings, and voice commands. This feature streamlines frontend workflows, allowing creators to build and refine interfaces without manual HTML/CSS styling.

OPEN SOURCE1h ago

Morgan Linton releases Small Harness v0.5 on GitHub, a local terminal-based agent harness optimized for running token-efficient overnight builds using small language models.

Morgan Linton announced the release of Small Harness v0.5 on GitHub. Small Harness is a terminal user interface (TUI) agent harness designed to run small language models (LLMs) locally on macOS. The tool supports multiple backends including Ollama, LM Studio, MLX, llama.cpp, and OpenRouter, offering a safe environment with approval gates for file and shell commands. The v0.5 release aims to enable developers to run overnight builds in a token-efficient, quality-optimized manner.

Google hits 75% AI-written code