Live data from Hacker News

Anthropic, please ship an official Claude Desktop for Linux

github.com

191–200 of 331 posts

Re: Anthropic, please ship an official Claude Desktop for Linux

#191

Earlier quoted context omitted.

Flatpak mostly solves this for GUI apps.

It does not. You just get more vocal angry customers who hate Flatpak and hate you for using it.

There is a difference between mandating that your customers use one specific Linux distro which is maintained by a controversial company, and supporting all Linux distros through an imperfect-but-fully-working method.

Sure, you'll still get a few complaints from ideological purists, but there's no avoiding that regardless of what you do.

Re: Anthropic, please ship an official Claude Desktop for Linux

#192

"Linux" is not an operating system. One does not "just" release something "for Linux"

Inded. But... Let's say by compiling for Windows and MacOS you are making it work on 80% of users' computers. Then you compile on an older Ubuntu version which will work on any newer Ubuntu or Debian system, and you're at 90%. Worth doing. Then you do a Centos/RH build to get to 95%.

This won't work for all Linux users, but that doesn't have to be the bar. Making a few linux builds for the most popular distros is incrementally better than saying it's Windows/Mac only.

Re: Anthropic, please ship an official Claude Desktop for Linux

#193

Hey! I manage the unofficial build at https://github.com/aaddrick/claude-desktop-debian Debian is in the name, but scope has grown to all backends, compositors, etc. The main reason must companies don't publish Linux electron apps is fragmentation. If you're doing anything more than rendering a webpage as an app, it starts to get complicated. I've got a bank of VM's setup for testing, and I still need it up.

> The main reason must companies don't publish Linux electron apps is fragmentation.

Unfortunately they do, instead of implementing a proper native application (don't hire web developers to do native apps).

Look at e.g. Cisco Webex, or Telegram. You can make a proper native application that just works and is not electron. A lot of people don't like electron apps, and it is more important in current economy (memory prices are too high to invest in more RAM).

Re: Anthropic, please ship an official Claude Desktop for Linux

#194
post #116

Earlier quoted context omitted.

Why?

I'm not the one who hates flatpak, but I will point you to this comment a little further up: https://news.ycombinator.com/item?id=48435993 Flatpak serves a need, there are plenty of users who like it and there are probably even more who just use it without thinking much about it. Personally, I like it for a few reasons: - Being able to install something dependency-heavy with just one package - Sandboxing - Getting a…

> Design for the people who are using and who like your product. Make adjustments based on their feedback. Ignore the people who just make noise.

And how do you know the userbase for GP's specific product is all Flatpak users? In fact, based on their comment, it appears as though they are explicitly not, hence their vocal frustration.

Re: Anthropic, please ship an official Claude Desktop for Linux

#195

Earlier quoted context omitted.

This feels analogous to the old Google latency improvement story - improve performance and p99 goes up, not down, because more people are now able to use your product. These angry customers are a symptom of having more customers; in this direction (compatibility) companies shouldn't be KPI'ing on angry customers. It is very legitimate that high compatibility means more very obscure, low value, high cost, bug reports…

> These angry customers are a symptom of having more customers; No, it's a symptom of having more of a very specific type of customer who is more demanding and difficult to please than your other customers. When you don't officially support Linux, the Linux users are not surprised. It's normal for them. They find other ways to use the product. When you do announce Linux support, you open Pandora's box of complaints.…

You start with 'no', but rather than disagree, you've effectively restated my comment.

I wasn't looking to anger you, just to provide a different lens in which to view the situation.

Re: Anthropic, please ship an official Claude Desktop for Linux

#196
post #134

Earlier quoted context omitted.

Even if you can create infinite software you still have to be very intentional about what you’re choosing to work on. There’s still a cost to testing, support, planning, etc even if coding is now “free”

Anthropic claims 8-fold productivity increase since 2025. If even that isn't enough to enable support for Linux, I don't know what it is.

[flagged]

Re: Anthropic, please ship an official Claude Desktop for Linux

#198
This sounds like a really good test for agentic coding and something Anthropic should seriously consider doing as a proof of Claude Code quality. They could easily have agents built per Linux distribution that will run on every new release and do complete testing and publishing. IF they can successfully do that it would be a nice marketing as well. :)

Re: Anthropic, please ship an official Claude Desktop for Linux

#199
post #138

Earlier quoted context omitted.

[flagged]

Oh please. The super key is the windows key. You come across as someone who has never used a computer before.

You come across as someone who’s never talked to a normal human being

Re: Anthropic, please ship an official Claude Desktop for Linux

#200

I love the dueling messages on AI "It'll make software easy and replace all software jobs!" "Sorry, a Linux client is too hard and too much work!"

Interesting take away, but one that I'd think is intrinsically wrong.

The Linux desktop is fragmented and gatekept in such a manner that most people of a reasonable mind aren't going to waste tokens on it, and damned sure not going to waste software/support engineer time on it.

Note that none of these complaints are about CLI or server software. "Linux" the kernel + GNU utils + shell handle all of this just fine. The year of the Linux desktop is just never going to happen unless one desktop standard takes over and 'wins' in the short to medium term.

Now, in the longer term when tokens run cheap maybe we'll successfully have 20 different competing, incompatible Linux deskstops, but it doesn't seem likely.

Post reply on HN