Top news
Investing in the public benefit OfficialAnthropic NewsOct 18Popular Musk Pledges SpaceX Shares to ‘Every Kid in America’ WebBloomberg TechnologyOct 19Popular Websites are closing their doors to AI agents WebThe Verge AIOct 17 This is the new ChatGPT Voice, powered by GPT Live OfficialOpenAI on YouTubeOct 18 How AI giants will reshape global philanthropy WebAxios AIOct 19 Gemini 4 Argon Really Surprised Me WebTwo Minute PapersOct 19 Google Gemini 4 agents just beat Minecraft with no human intervention WebEngadget AIOct 19 Who could it be? 🧐 OfficialGoogle on YouTubeOct 18 Anthropic gives outside AI watchdog “employee-like” access WebArs Technica AIOct 18 What are self-replicating ‘AI worms’, and why is Anthropic worried about them? WebThe Guardian AIOct 18
GlossaryAgentsIntermediate

Agentic coding

Also written: Autonomous coding, AI coding agents, Self-coding AI, Agentic coding models, Autonomous Coding Models, Agentic coding tools, Autonomous coding tools, Agentic coding evaluation

AI agents that can plan, reason, and execute multiple steps autonomously to write, test, and manage code.

In plain terms

Agentic coding refers to artificial intelligence systems that can act like a programmer. Instead of just writing a single line of code, these AI agents can understand a task, plan out the steps needed, write the code, test it, and even fix errors on their own.

For example, an agentic coding system could be given the goal of building a simple website. It would figure out the HTML, CSS, and JavaScript needed, write the files, and then test if the website works correctly, making adjustments as necessary without constant human input.

The expert version

Agentic coding involves AI agents capable of autonomously performing complex software development tasks. This includes understanding requirements, planning multi-step execution, writing code, running tests, debugging, and committing changes. These agents leverage large language models with enhanced reasoning and planning capabilities to manage the software development lifecycle with minimal human intervention.

The current trend shows these agents moving beyond simple code generation to handle longer-horizon tasks and iterative development, integrating with tools like design software or version control systems. Tradeoffs involve the complexity of the task, the reliability of the agent's planning and execution, and the potential for emergent errors in autonomous operation.

Why it matters now

Frontier AI labs are releasing new models and tools focused on agentic coding. These advancements allow AI to autonomously plan, execute, and adapt coding tasks, leading to more efficient software development and maintenance. Companies are also developing benchmarks to assess the reliability and performance of these autonomous coding agents.

Where it shows up 67 items, newest first
How each company uses it
Anthropic Claude
Anthropic is developing its own version control system to manage code written by AI agents and has published research on evaluating these agents.
Google Gemini
Google Gemini is integrating advanced reasoning and coding agents into its app and providing tools for developers to build agents within its Looker platform.
Meta Meta AI
Meta AI has released open-weight models that enable developers to build and customize AI agents for coding tasks.
OpenAI ChatGPT
OpenAI is using its latest models to automate code maintenance, such as cleaning codebases and improving software performance.
Related terms

AI agentsCode generationAutonomous systemsSoftware development lifecycleLong-horizon AICopilot

Often confused with
Code generation
This is a subset of agentic coding, where AI only writes code w
AI assistants
These typically require more direct human guidance for each step,
Automated testing
This focuses solely on verifying code correctness, whereas ag
Origin

The concept of agentic coding emerged from the advancements in large language models and AI agents capable of complex reasoning and task execution. It represents a natural progression from basic code generation to more autonomous software development processes.

Try this about 5 minutes

Ask an AI coding assistant like GitHub Copilot or ChatGPT to write a simple script, for example, a Python script to rename all files in a folder. Then, ask it to add error handling for cases where the folder doesn't exist or if a file with the new name already exists. Observe how it plans, writes, and potentially debugs the code in response to your prompts.

Try it: Playbook entries that use thisPlaybook →

The Playbook: reusable how-tos, prompts, configs and tips pulled out of the news.

Questions people ask

Can agentic coding replace human programmers?
Agentic coding can automate many routine and complex coding tasks, significantly boosting productivity. However, human oversight, creativity, and strategic decision-making remain crucial for advanced software development.
What are the risks of agentic coding?
Risks include potential security vulnerabilities in AI-generated code, the difficulty of debugging complex autonomous systems, and the possibility of AI agents making unintended errors or misinterpreting requirements.

First seen on Subvolts Jul 25, 2022. Updated Oct 17, 2026. Written by an AI model from the items above and reviewed against them weekly; report an error.