The Future of [JetBrains] Fleet
19 points by bhoot
19 points by bhoot
I tried Fleet and didn't see any potential from it. It seemed like its main virtue was its visual similarity to VS Code. So I suspect the part about discontinuing Fleet is for the best.
Disclosure up front: I don't think "agentic development" is great, either. So "We are now building a new product focused on agentic development" doesn't sound like a good idea to me. Recognizing that people who aren't me might want more integration points for that, it'd seem better to build out those integration points than to roll that into a new product.
Disclosure up front: I don't think "agentic development" is great, either.
Sure, I mostly agree with you tbh. But putting myself in the shoes of the jetbrains ceo: they have to try this. Their whole business is developer tooling, and that industry is on fire with llm experiments. Even if it doesn't work out long term, they kind of have to try, just in case it does - otherwise they risk being left behind.
Oh, they sure do! I am just saying I think it'll be better inside their sundry IDEA things than as a whole new product.
As a fun fact, I worked on the early versions of Fleet almost now 10 years ago when I was at JetBrains. Back then, the codebase included quite a bit of Clojure. Not sure about the current state now though, everything has probably changed many times over. The engineers working on Fleet have an amazing depth of knowledge, and I learned heaps when I was there. Some of the work there also inspired me to write a post about the overlap between Kotlin, Clojure, and persistent data structures sometime later https://serce.me/posts/2017-06-29-fantastic-dsls.
I kept an eye on Fleet after it's initial announcement in 2021. Zed came a year or so later. They both seemed to have similar goals: a fast LSP oriented alternative to VS Code. While Zed rapidly developed, and added missing features Fleet seemed to only get minor updates, and large barriers to usage like being able to add missing languages or vim support were not prioritised. I think Zed ended up finding its place, whereas Fleet was discontinued—it never really felt like it was a major focus for JetBrains.
The main thing I found interesting about this site was the way they've formatted their logo to always have a small 'i' joined into the capital 'J'.
From the web inspector, it's text rather than an image, and editing the text on the page to add 'IntelliJ' also ends up with the formatted logo - I wouldn't expect Javascript to be run dynamically to format new instances, so that narrows it down to probably CSS.
There's no particular style applied to the logo, but they are using a custom font - viewing the custom font file shows an 'iJ' glyph, so I imagine they're using some sort of ligature trick. It only works for 'InteliiJ' with that capitalisation so there's a bit more going on, but I've lost interest in digging further.
Nice trick to avoid JS though!
I also tried Fleet and didn't see much value from it. It did feel slightly snappier, but also like it was missing half the features of the full IDEs.
My work computer is a 2021 M1 with just 16GB, and as long as I give them a restart once a day, the "classic" IDEs run perfectly fine on there. I think overall the situation isn't bad enough to warrant trade offs from a user perspective.
I was originally excited to not have to jump between IDEs for different languages, as there's no single one that covers all I need (as IJ Ultimate misses the profiler for Rust), but not at the cost of everything that set JetBrains apart from VS Code.
I tried fleet a while after it was first announced, and it didn't do enough to justify moving to it from either Intellij (fleet both felt slower and heavier than Intellij), or from editors like neovim/sublime-text as Fleet didn't really support LSP based integrations for anything aside from Kotlin/Javascript/Rust, etc, and there was no meaningful plugin support.
I wasn't entirely sold on what it offered that was better than Jetbrains' other IDEs at the time but this might also be a case of me not being the right audience for it.
I remember JetBrains announcing Fleet shortly after Twitter added "fleets" as a feature. Remember when Twitter tried being Instagram after Instagram successfully ripped of Snapchat?
Anyway, now they're finally both dead, and hopefully no one ever decides to name something "fleet" again.