No more JetBrains products for me

Original Article ↗ Hacker News Discussion ↗

Longtime users of JetBrains IDEs describe growing frustration with slow startup times, constant re-indexing, heavy resource usage and intrusive AI features, especially on older or modest hardware. Many are experimenting with lighter editors such as Zed, Neovim, Helix or VS Code plus language servers, arguing that modern LSPs and external AI agents now cover most of the functionality they actually use. Others still defend JetBrains for deep language integration, powerful refactoring and debugging—particularly in ecosystems like Java and C#—but worry that performance, UX churn and AI monetization could erode its advantage.

Performance, Resource Use, and Indexing

  • Many report JetBrains IDEs (CLion, RustRover, WebStorm, PyCharm, DataGrip, Rider, etc.) as slow: long startup, frequent re-indexing, UI freezes, high CPU/RAM use, and occasional memory leaks/crashes.
  • Some say performance is acceptable or good on modern hardware if IDEs stay open all day, plugins are trimmed, and JVM memory settings are increased.
  • Re-indexing is a major frustration: users see repeated full re-index cycles (especially with multiple projects) and complain there’s no clear diagnostic showing why it’s happening.
  • Others note that language servers in editor-based workflows can also be slow or memory-hungry; experiences differ by language (Go and Java LSPs are singled out as problematic).

IDEs vs Editors (Zed, Vim/Neovim, Helix, VS Code, etc.)

  • Many have moved or are moving from JetBrains IDEs to lighter tools (Zed, Neovim/Helix, Sublime, VS Code, Kate, KDevelop), citing instant startup, low resource usage, and better “flow”.
  • Pro‑IDE voices argue that full IDEs provide superior static/dynamic analysis, refactoring, debugging, profiling, test integration, database explorers, and framework awareness (Java/Spring, C#, etc.), which are hard or impossible to fully replicate with LSPs and plugins.
  • Some adopt hybrid workflows: fast editor for most editing, JetBrains or Visual Studio only for debugging or heavy refactoring.

AI, Agentic Coding, and Product Direction

  • A subset claims “IDEs are dead” or diminished in an era of agentic coding, using AI agents to generate and edit code instead of manual editing.
  • Others strongly reject this, preferring to keep IDEs and have agents interact with them, or to keep AI separate entirely.
  • JetBrains’ AI assistant is widely criticized as intrusive, slow, and lower quality than using AI providers directly; UI elements (sidebars, ads, auto‑enabled features) are seen as clutter and have pushed some to cancel subscriptions or freeze on older versions.
  • Some appreciate new agent-connector protocols and the ability to disable AI, but see overall focus shifting from core performance/quality toward AI and UI churn.

Trust, UX, and Ecosystem Concerns

  • Users complain about constant UI reshuffles, a divisive “new UI”, and features moving behind paid tiers.
  • There are anecdotes of criticism being removed from JetBrains forums/blogs and a sense that subscription economics drive churn over polish.
  • Despite frustrations, many still see JetBrains as “best in class” for certain ecosystems (Java/Scala, C#, enterprise work) and stay for those strengths.