Coding

Visual Studio Code 1.121

"Visual Studio Code's latest update, version 1.121, introduces a major overhaul of the editor's rendering engine, shifting from Chromium Embedded Framework to the more modern Electron 22, which promises improved performance and reduced memory usage, particularly in large-scale projects with complex extensions."

Tom H (AI-assisted) May 20, 2026 1 min read EN

Based on reporting from Source.

Visual Studio Code's latest update, version 1.121, introduces significant changes to the editor's rendering engine. The update shifts from Chromium Embedded Framework to Electron 22, promising improved performance and reduced memory usage, particularly in large-scale projects with complex extensions.

Overview

The 1.121 release of Visual Studio Code includes several new features and improvements. These include support for pinning favorite models in the language model picker, an "Add to Chat" option in the right-click context menu, and automatic disposal of background terminals created by the chat agent.

What it does

The update also expands terminal tool output compression to cover more commands, including test runners, build tools, linters, Docker, and package managers. Additionally, an idle-silence timer has been added to the run_in_terminal tool, which automatically promotes a sync command to background execution when it produces no output for a configurable period.

Other improvements include the ability for extensions to declare their scripts as ES modules, keyboard-interactive authentication support for Agent Host SSH connections, and a newer version of ConPTY bundled directly with VS Code on Windows.

The full list of changes includes:

Add support for pinning favorite models in the language model picker
Set a VSCODE_AGENT environment variable when Copilot Chat runs commands in the terminal
Add an "Add to Chat" option to the right-click context menu in integrated browser
Automatically dispose background terminals created by the chat agent once their command finishes
Allow extensions contributing markdown.previewScripts to declare their scripts as ES modules
Expand terminal tool output compression to cover more commands
Add an idle-silence timer to the run_in_terminal tool
Fix multi-line shell commands in the Agent Host terminal tool
Bundle a newer version of ConPTY directly with VS Code on Windows
Add keyboard-interactive authentication support for Agent Host SSH connections

In practical terms, these updates aim to enhance the overall user experience and efficiency of Visual Studio Code, especially for developers working on complex projects. By improving performance and reducing memory usage, the shift to Electron 22 is expected to make a noticeable difference in the editor's responsiveness and stability.

The takeaway from this update is that Visual Studio Code continues to evolve with a focus on enhancing user experience, performance, and functionality, making it a robust and efficient tool for coding and development tasks.

{ "headline": "Visual Studio Code 1.121 Update", "synthesis": "Visual Studio Code's latest update, version 1.121, introduces significant changes to the editor's rendering engine, shifting from Chromium Embedded Framework to Electron 22. The update includes several new features and improvements, such as support for pinning favorite models, an 'Add to Chat' option, and automatic disposal of background terminals. Other improvements include expanded terminal tool output compression, an idle-silence timer, and keyboard-interactive authentication support for Agent Host SSH connections.", "tags": ["Visual Studio Code", "Electron 22", "Coding"], "sources_used": ["https://code.visualstudio.com/updates/v1_121"]

More articles like this

Coding 1 min

Rewrite Bun in Rust has been merged

Rust's growing presence in the JavaScript ecosystem takes a significant leap forward with the merger of Bun, a high-performance JavaScript runtime, into the Rust language itself. The integration, enabled by a new compiler plugin, allows developers to leverage Rust's memory safety guarantees and performance optimizations directly within Bun's JavaScript codebase. This move could redefine the boundaries between JavaScript and systems programming.

Coding 1 min

Classic 7 is a Windows 10 LTSC mod to look 1:1 to Windows 7

A Windows 10 LTSC mod, dubbed Classic 7, has emerged as a precise emulation of Windows 7's UI and behavior, leveraging the Long-Term Servicing Channel's (LTSC) stability to recreate the classic OS's look and feel with remarkable fidelity, down to the last pixel and system sound effect. By targeting the LTSC's Windows 10 2019 and 2021 builds, the mod achieves a seamless integration with modern hardware and software.

Coding 2 min

Claude for Small Business

Small businesses can now tap into large language models with Anthropic's fine-tuned Claude, a customized AI solution that leverages the company's 137B parameter model to provide scalable, on-demand conversational support. By integrating Claude with existing workflows, small businesses can automate tasks, enhance customer engagement, and streamline operations without requiring extensive AI expertise. This move marks a significant expansion of large language model accessibility.

Coding 1 min

Arena AI Model ELO History

"Hidden Decline: AI Model Performance Plummets After Initial Hype, Data Reveals. A live tracker of flagship AI models' ELO ratings shows a stark contrast between initial launch excitement and subsequent performance decay, with generational jumps and slow declines becoming apparent only when viewed over time. The data raises questions about the long-term viability of AI models and the need for more transparent performance metrics."

Coding 1 min

A Claude Code and Codex Skill for Deliberate Skill Development

A novel approach to skill development emerges with the release of a custom Claude code and Codex skill, leveraging large language models to facilitate deliberate practice and adaptive feedback loops. By integrating Codex's generative capabilities with Claude's conversational interface, users can engage in targeted skill-building exercises, receiving real-time feedback and guidance to refine their knowledge and expertise. This hybrid model holds promise for accelerating learning and skill acquisition.

Coding 1 min

The Other Half of AI Safety

A long-overlooked vulnerability in AI safety protocols is being exposed by a growing number of edge cases, where seemingly innocuous model updates can have catastrophic consequences, highlighting the need for more robust "backdoor" detection and mitigation strategies in large language models. Specifically, researchers have identified a class of "adversarial perturbations" that can be injected into model weights, compromising downstream applications. This "other half" of AI safety is now a pressing concern.