(2026-03-24) Linear Issue Tracking Is Dead

Linear.app says Issue tracking is dead. It was built for a handoff model of software development. A PM scoped the work, engineers picked it up later, and the system filled with prioritization, negotiation, and workflows to bridge the gap. That ceremony came from real constraints. Engineering time was scarce.

But over time, complexity started to look like sophistication. The more process a system could absorb, the more advanced it seemed. Overhead kept growing, and the process became the work.

Linear has always been built on the opposite belief: the best systems remove overhead so teams can focus on building.

Agents push that further

People can spend more time on intent, judgment, and taste, and less time managing the mechanics of the process.

Today, coding agents are installed in more than 75% of Linear’s enterprise workspaces. In the last three months, the volume of work completed by agents grew 5x, and agents authored nearly 25% of new issues.

In this new world, the next system is not designed around handoffs. It is designed around context and agents.

Agents are not mind readers. They become useful through context. Customer feedback, internal ideas, strategic direction, decisions, and code all need to be captured in a system that humans and agents can work from together. This is also true for human builders.

That is what Linear is becoming.
Linear is the shared product system that turns context into execution.

Toward this vision, today we are launching:

Linear Agent.

Skills. As you find workflows worth repeating, you can codify them as reusable skills and compound your learning

Automations. Starting with Triage, trigger agent workflows the moment an issue enters the system.

Coming soon:

Code Intelligence

Code Diffs.

Linear Coding Agent.

These updates build on our earlier work in Triage Intelligence and deep integrations with cloud coding agents and other AI tools.

Issue Tracking Is Dead | Hacker News

At ZAR once we had pervasive ingestion into an organization-wide knowledge graph in place and working well, the next step was to ditch Linear and replace it with a homegrown experiment tracking system that focuses all product engineers on empirical data and scientific method applied to how we prioritize work.
It's the only way to actually encourage high-agency, high-ownership behavior.

Classic clickbait title. I guess it works, but also baits me to respond to it in the first paragraph: Issue tracking is clearly not dead, it is more important than ever. They are doing almost everything right

I really want to see diffs right in the issue. PRs are a dumb historically grown in-between step that is just annoying. As everything else becomes faster, this becomes more of a bottleneck for iteration speed.

Linear Coding Agent
Is this supposed to replace my dedicated coding agent? I’m skeptical of coding agents being built as parts of other products

• We're well aware that feedback comes from anywhere. The Linear agent also exists in other tools (Slack.com, Gong, Intercom, Zendesk, etc.) and we'll continue to add more channels to support collecting and managing feedback where it's coming from.

You can't possibly track all the places where customer feedback come from. So you'll end up needing a human to curate feedback.

Why not use Claude Cowork? It already can connect to any tool via MCP and do all these things (and Claude Code to, well, code tasks)

• Claude Cowork is great when you want to collaborate with AI and third-party tools via MCP, but it's not a multi-user collaboration tool built for organizations. Our customers need to collaborate on software products and AI is only one part of the equation. Our customers need a system of record (long-term history, priority, and cross-team visibility of a project) and contextual collaboration (e.g. a customer success team member reporting a feature request or bug, a person on the product team deciding it's worth building/fixing

🔵 Linear says Issue Tracking is Dead

For product teams, this context includes things like customer feedback, internal ideas, strategic decisions and your codebase, and Linear is pitching itself as the best place for product teams to build this rich context.

Linear says Coinbase is one of the early adopters of “agent-first” development and they published a piece from their Head of Engineering on how they managed to roll this out.

But, as some people have argued, what happens if IDEs like Cursor also decide to bolt on additional product-oriented features to become an alternative place to build the rich context that AI Agents need? Could we potentially also see tools like Claude Code or Figma do the same?

Linear adopts agentic AI as CEO declares issue tracking dead

One thing not mentioned in any of the posts or agent documentation is security, other than that "Linear Agent operates within your existing permissions." Generative AI systems can be vulnerable to malicious prompt injection and other errors. This will become a bigger concern as the capabilities of the agent increase

Basecamp from 37signals, another project management tool, is also planning to reposition itself as "agent first, agent native" with access from any AI agent via a command line interface (CLI) and the promise of becoming "an agent-driven assistant."

Nate B Jones: AI agents are about to route around every tool that can't pass 5 structural tests. Here's the diagnostic.

OpenAI open-sourced Symphony, and Linear became the literal control plane for the most ambitious autonomous-coding system ever shipped. Some internal teams saw a 500% increase in landed pull requests.

The thing Saarinen had just eulogized was now the substrate that made all of it work.
He was right about the user experience and wrong about the infrastructure.

The translation step is going away. The state machine, the assignee field, the audit history, the dependency graph, those are staying. They are quietly becoming the most strategic infrastructure in the enterprise software stack, and that repricing extends to every boring tool you own.

Here’s what’s inside:

The five properties that make a tool agent infrastructure. Why Symphony looks obvious in retrospect

Why your UX choice was actually a data choice. How Linear’s design discipline accidentally produced the cleanest agent substrate on the market, and what that means for tool selection from here

The Atlassian repricing. The MCP server, the Anthropic partnership, the acquisition rumor, and why the market is treating Jira’s installed base as a strategic AI asset.

Which of your tools are next.

What to do about it. Three prompts that score your stack, spec out an MCP server for anything you build, and produce the migration brief your leadership team is missing.