JetBrains Cancels Fleet

Moving to Apple Silicon made it bearable for a few months but somehow Jetbrains manages to get slow even on a M3 Max with 36GB RAM.

Ive been fiddling with configs for years, i tried everything since i was a Jetbrains diehard.

Instead of trying to catch up to other AI editor they should get back to their core and make it possible to use Jetbrains on medium sized Monorepos with multiple languages.

I was hyped when i heard they would release a standalone git product, but then they scrapped it!

In the end i was only dependent on it for debugging and my usual git workflow.

I now switched to zed and gitkraken, i will figure out a new debugging workflow, ill never wait 5 minutes for a simple search action again

Netscape tried to remake Navigator whilst halting development on the old codebase, and it killed them.

Microsoft tried to remake Word, the rewrite failed. Luckily they had continued to develop the old codebase in parallel.

Google tried to remake Gmail multiple times. Every attempt failed.

Apple tried for years to remake MacOS Classic and failed every time. Eventually they had to buy and reskin NeXTStep.

Banks are full of war stories trying to migrate off their old mainframe codebases, and often giving up.

I kinda expected Fleet to die from the day it was first announced. IntelliJ is an extremely mature product that's hard to compete with. They've continually managed big changes to it to keep up with changing fashions and trends in the IDE space, most recently with their new Islands theme that launched yesterday, with integrated coding agents and so on. It's outlasted continuous competition from free IDEs that are always abandoned after enough years pass and whichever executive was championing subsidies moves on or retires (see: NetBeans, Eclipse, VS Express, MonoDevelop...). VS Code isn't so different. Fleet was clearly a reaction to that but the concept was not innovative and focused on reinventing wheels that users wouldn't be able to tell the difference for and which would consume most of their budget, like writing a new UI toolkit, or using a split frontend/backend architecture. Same mistake Mozilla made. Meanwhile IntelliJ was continuously refactored and improved, so Fleet chased a moving target even when they reused a lot of code.

Although people will hate to hear it, the history of the IDE market suggests that eventually MS will get tired of funding VS Code without a big revenue stream to justify its existence. Executives like making new projects and being able to present growth because it represents glittering future potential, but they hate being landed with the maintenance of loss making legacy projects when the originators move on. There's no glory there. For all their problems, JetBrains aren't going to lose interest in their core products due to random executive churn, and that has given their core IDEs a remarkable staying power.

That's vscode's moat.

Anytime the same extension exist in both vscode and jetbrains, the jetbrains version is clunky, crash, and unstable.

I keep Jetbrains open while using vscode, for its local history/git/etc features, but how long will that be enough to keep my subscription

Let the people that want to build an IDE from the ground up have their fun over in VSCode land, please just focus on a powerful IDE that works out of the box.

PS: Agentic development is fine to pursue but so far things like Claude Code run laps around everything JetBrains has tried. Add "mount points" for agentic flows but please just focus on making a powerful IDE. Agentic development was unable to lure me away from JetBrains, double down on that, not trying to be Cursor.

My problem was that Fleet just wasn't very good when compared with VSC.

For my more serious development I use JetBrains IDEs (one of the few pieces of software that I actually pay for, alongside MobaXTerm and some others) but Fleet didn't neither use that much less resources, nor was that much more responsive, nor was a step above VSC in any way. To be clear, I didn't hate it, it wasn't horrible and with a bit more work could have been quite good... just not convincingly so up until now.

If they wanted to throw some more years of engineering at it, maybe, I mean look at what Zed is doing and it seems to be okay, but I don't think it makes that much business sense for them - they already have Junie available in their editors for AI stuff and that other subscription (though I just use Claude Code, Codex, Gemini CLI and sometimes VSC with KiloCode/RooCode/Cline and either those models through the API or Cerebras Code since it works pretty well in there).

I just find that most AI solutions out there are also a little bit half-baked, like Gemini CLI fails when I paste multiple lines into it, whereas KiloCode/RooCode/Cline are unable to give a model enough helpful instructions for it to not start looping when it fails applying a complex diff sometimes, and pretty much nothing outside of the regular GitHub Copilot plugins does autocomplete decently (especially if you want a local model with Ollama or something, no good options, Continue.dev is trash).

With how prevalent AI is and how useful various linters and build output is, sometimes I wonder whether I need to pay hundreds of euros for the Ultimate package of tools when I don't write/refactor as much code manually and doing what I need inside of VSC also feels more and more sufficient. Maybe a bit except Java codebases, Spring Boot sometimes does weird shit and you're better served by an IDE that's aware of all of the templating, annotations and other stuff.

Oh well, despite being RAM hogs, I still enjoy the experience of using JetBrains IDEs and if nothing else will keep them around for that reason for a while. A bit like how I also enjoy a GUI of some sort for Git, like previously I paid for GitKraken but reevaluating my usage found that SourceTree is also decent enough for the price (free vs GitKraken paid version), I can just drop down to the CLI for niche use cases.

but sigh. Jetbrains really just has no focus.

Fleet came at a time when intellij felt extremely bloated. iirc they had painted themselves into a corner where it was easier to rip the band aid and start anew.

Fleet was supposed to be that promised editor which was snappy and had the power of intellisense + all things we liked about Intellij editors ... but without the terrible glacial bloat. but in a stroke of bad luck and typical lack of focus from Jetbrains, Fleet just didn't get good enough quickly.

I say lack of focus because (like their multiple attempts at AI) Jetbrains also had a lite mode in the start but that didn't work great. then came Fleet. But it was not getting better quickly enough and they changed course to make Fleet their main cross platform editor ... but even that didn't take.

I really am worried for Jetbrains and intellij. In a world where even VScode is having its lunch eaten by Cursor, Jetbrains is quickly getting pushed out of the list of contenders. they've squandered away a lead they once had in a certain niche for code editors.

I personally only pull up Intellij these days when there's some platform specific tool that's built in (like the emulator in Android Studio) or certain Android specific profiling tools, or the debugger.

Otherwise I rarely find myself using Intellij. My usage has dropped precipitously.

This is sad. It seems innovation has all but stopped for IDEs intended for a human as the primary driver.

I didn’t really want to switch to VSC but the extensions made it easy to find things that you just couldn’t do in IntelliJ, and… I haven’t looked back. Haven’t really missed the suite at all.

It would stand out if you focused on performance and not rebuilding another sluggish Java based editor. Zed is what I'll be ditching my JetBrains sub for, and it is not just some VS Code based editor. What happened to JetBrains? They used to be amazing, now its just disappointment all around? Did they elect a terrible CEO?