131–140 of 331 posts
"Sources for the load-bearing claims" I no longer know if this is a real person who is simping for Anthropic and will be ultimately enslaved by Anthropic or if this is an Anthropic ad to have "proof" for the high demand of their services.
Neither, it is just a random LLM-generated issue.
Earlier quoted context omitted.
Anything that isn’t very simple? Like a llm chat interface? If zoom and Microsoft Teams of all people can do it, anthropic should be able to. Have you considered that maybe their code is just bad?
Microsoft gave up on teams for linux, the app that's available now is a community electron webapp wrapper, and Zoom isn't electron at all, its QT but they chose the path of only supporting a few distros (Ubuntu, Debian, Mint, RedHat) and they also don't rely on system QT versions, they vendor it. If your app is open source, I say just build & test on one of the major distros and let the community port it to others. I…
For software development use of Claude, I'd be happy if the `claude` CLI executable does everything I need, within the Linux KVM VM sandboxes I create for the work, without a desktop client. The cleaner and more trustworthy, the better.
Also, for random interactive use of Claude for asking questions, I use it from my host desktop, sandboxed within the Web browser, and I want that to be well-supported. Someone marketing/product person at an AI company will naturally want to dark-pattern push people towards a proprietary desktop client, but that's one corner of abuse potential that we can still keep in check.
For agentic automation of my host desktop things and the things they have access to... no, thank you, the state of the art is not ready for that.
If only Anthropic had some kind of automated tool that was good at porting software
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”
Earlier quoted context omitted.
How do I bring your app to the foreground if I can't see an icon anywhere? I just installed Ubuntu for the first time a few weeks ago and genuinely don't see how people are supposed to use it, coming from a Windows/Mac background. How does a Linux user know what's running, without going to a terminal and running top? The lack of desktop UI affordances in the leading "user-friendly" Linux distribution should be seen a…
> How do I bring your app to the foreground if I can't see an icon anywhere? On GNOME? Alt-tab, super overview, or click the dock icon. It's literally not any more complex than multitasking on an iPad.
Earlier quoted context omitted.
This isn’t an issue in practice because the software running on Linux is open source. Yes if you want to distribute proprietary binary blobs and have them work forever it’s going to be a challenge, but in that case better to stick with the binary blob operating system.
There's plenty of software running on Linux that is not open source, though.
The only proprietary software I have running on my machine is electron apps, which are essentially bloated VMs. As we’ve seen from this thread, this is still apparently too great of an engineering feat for anthropic to tackle. I don’t think I’m unique in this regard.
OK, just please be careful how you frame what you're asking for. For software development use of Claude, I'd be happy if the `claude` CLI executable does everything I need, within the Linux KVM VM sandboxes I create for the work, without a desktop client. The cleaner and more trustworthy, the better. Also, for random interactive use of Claude for asking questions, I use it from my host desktop, sandboxed within the W…
Earlier quoted context omitted.
> How do I bring your app to the foreground if I can't see an icon anywhere? On GNOME? Alt-tab, super overview, or click the dock icon. It's literally not any more complex than multitasking on an iPad.
[flagged]
Earlier quoted context omitted.
There are a number of utilities for this. I use jai: https://jai.scs.stanford.edu/ but also have seen nono: https://github.com/always-further/nono smolvm: https://github.com/smol-machines/smolvm zerobox https://github.com/afshinm/zerobox and matchlock https://github.com/jingkaihe/matchlock They all have pros and cons. Pick the one that suits you best. Then you're also agent harness flexible (I use opencode).
I would like a solution that was itself not largely written by an AI
> [1]: Was jai written by an AI coding agent? No. While this web site was obviously made by an LLM (ChatGPT read the man page, asked some follow-up questions, and produced a prompt from which claude code built a vitepress site), jai itself was hand implemented by a Stanford computer science professor with decades of C++ and Unix/linux experience. As an experiment, the author did previously try vibe-coding a container, but the results were disastrous and repeatedly put his machine in a state that required a reboot (e.g., recursively changing the attributes of all mounts in the wrong mount namespace). The author does use coding agents to look for bugs, get feedback, and develop tests. However, rest assured that a single human understands every line of C++ in jai.
If only Anthropic had some kind of automated tool that was good at porting software
Imagine if cutting edge AI companies could decide to using their world best AI to 1) Develop software for linux 2) Provide decent support