lone-cloud

Yes, they are https so security is not a concern. I want to crowd blend at the network level so that my requests don't aversive that I'm using a privacy notification system (or a VPN).

I meant that at the network level, should something be monitoring, your client would be making requests to a unique (ntfy) address that virtually nobody else ever uses. Requests to Mozilla are more crowd blended as it's quite popular already.

The webpush endpoint registration is initiated from the client. The Android app can act as a client and it's the one that registers Mozilla's autopush service.

I'm a privacy focused self hoster and ntfy was not good enough for me so I created Prism with an optional Android app: https://git.lonecloud.dev/lone-cloud/prism

The difference: Your phone doesn't get notifications from your notification service directly, but from Mozilla which helps to hide your digital fingerprint of you using a separate notification provider like ntfy. Prism also allows Telegram or Signal to act as a notification distributor. In addition, I use Proton Mail which doesn't really allow custom notification services. Prism has a custom integration to work around that so your proton mail notifications don't have to touch google services.

I would recommend people use comaps instead which is the actual FOSS fork. OM has a long history of malicious behaviour like quietly adding ads, turning a part of its previously open sourced code proprietary and misappropriating donations. OM has lost most of its community a year ago to comaps and are now rushing vibe coded features to compensate.

Are you me? Somebody was talking about gitea on here yesterday and I also ended up self-hosting and moving all of my private projects to Ferjero yesterday after a bit of research. I can't bring myself to move public projects due to job prospects + GitHub network effect. Otherwise I'm role playing as a system admin now with 20 local services for various things I need. I think the most important thing is to have regular backups as you're now in charge of keeping your data from getting lost.

I use rclone + backblaze. You get 10GB for free which is more than enough for self-hosted stuff.

It's not that hard to implement. I did: https://github.com/lone-cloud/gerbil

I did: https://github.com/lone-cloud/gerbil

Any half capable engineer can vibe code this in a week. Who cares?

Maybe we're arguing semantics, but I think calling shadcn a "basis for a design system" is more accurate than a traditional component library. The difference to me is that shadcn lives inside your codebase and you can fully customize it as you please. You cannot customize a component library like MUI nearly to that extent.

shadcn ui is not a component library but the basis for a component library that has great accessibility built-in from the start, so yes, it is a refutation.

It's not that hard to build a design system with decent accessibility. Just use shadcn ui components instead of rolling your own.

Maybe you should take your own advise and try to understand his perspective.

When I finished reading it I thought it was an anti-Trump piece, but the author also wrote: "That's why it baffles me that some prominent technologists embrace hateful ideologies.". Was Trump a techie too? He must have been behind the creation of JS.

"The ARM processor which powers the modern world was co-designed by a trans woman." This is not factually correct. Roger Wilson was one of the designers of the processor, but he didn't transition to become Sophie Wilson until 9 years after the first release of ARM1 according to Wikipedia.

The mindless code comments are a dead giveaway. It's always the same pattern of: "a thing" <--- here is a thing Generally a dev would clean these up, but when they don't it's a major red flag to me that it's just unreviewed vibe coded slop.

The code is partly refined AI generated slop and the UX is lacking. The functionality is very basic and needs to be more thoroughly tested. This type of project is half a work day tops for a senior+ dev to create with agentic coding.

- "Electron comes with a few rather significant drawbacks" Not going to mention them at all though? - what possible reason could one have to use Next with Electron? Not everything needs to be in Next and there's no reason for SSR + Node.js API server (primary advantages of Next vs React) when the client and server are on the same machine. The author's solution is to wrap this with another dubious framework (Nextron looks abandoned and hasn't been updated for over 10 months btw) to force them to somehow work together. Like why? - the idea behind Tauri to leverage the user's system's webview may sound like free real estate, but as many Cordova devs have previously found out, it's incredibly risky when you have no control of your own runtime. Desktop webviews are not a uniform layer. Ex. Webrtc is not supported on Linux webviews.

That's how the pros do it. Yes, you can load both a text and image gen models at the same time. Needless to say you'll need a very beefy GPU(s) to do this so I wouldn't recommend it unless you know exactly what you're doing as generally you'll want to max out your VRAM for one model at a time for the highest quality results. Open webui and sillytavern allow both text and image gen from the same UI although I wouldn't recommend it for advanced users. Otherwise Gerbil will give you multiple pages to toggle through via the titlebar dropdown.

At the risk of sounding like a linux snob, I highly question OP's taste: - recommending X11 WM/DE over Wayland in 2025 is nasty work - 7.3/8 GB swap usage? mad man - bash? meh. integrated GPU? meh - 4k packages installed seems super excessive (I'm at 887 (pacman)) - that topbar has way too much superfluous info and looks bad. I guess it's nice if you frequently forget which timezone you're in and need to know how fast the wind is outside?

I would personally never use an out-of-the-box browser because of this. Instead of Firefox I use(d) LibreWolf which strips out all the Mozilla trash. I'd argue that all gecko-based browsers are "bad" and Mozilla has been largely incompetent for many, many years now with failing to modernize their browser engine. As I remember, their current CEO fired the entire team responsible for a modernizing their browser codebase in order to focus on the type of crap described in this article. I strongly believe that by far the best browser out there right now is Brave. However, as with Firefox, it's pretty terrible out-of-the-box and you have to add "local company policy" files in order to be able to strip out all of its web3, AI and other crap that nobody wants. In a way, all modern browsers are the same. I can't wait for more serious competition although I'm doubtful that's it's really possible due to the sheer complexity of the problem, but I do realize that there are some serious contenders coming up in the next few years.

Gerbil's built-in image generation is based on "StableUI" and I also prefer its super simple UI. Yes, you can load your own LoRa from the "Image Generation" tab. Gerbil also includes the optional ComfyUI integration from the settings for very advanced users. Its graph-based UI is a bit too advanced for me personally.

I believe what you're describing is outside of the scope of Gerbil. Gerbil is not an LLM front-end, but Gerbil will run your LLM and seamlessly integrate (orchestrate) it with a custom front-end from the list in my original message. I believe this functionality will need to live in a custom front-end. I'm curious how jan.ai is planning on handling this. I'm guessing they're writing their own custom front-ends which is probably tightly integrated with their system.

Holy, your machine is a beast. 96GB of VRAM is pretty insane. I've been running a single 16GB VRAM AMD GPU. At the bottom of Gerbil's readme I listed out my setup where I use a 27b text gen model (gemma 3) but you'll be able to use much larger models and everything will run super fast.

Now as for your question, I started out with LM studio too, but the problem is that you'll need to juggle multiple apps if you want to do text gen or image gen or if you want to use a custom front-end. As an example, my favorite text gen front-end is "open webui" which gerbil can automatically set up for you (as long as you have Python's uv pre-installed). Gerbil will allow you to run text, image and video gen, as well as set up (and keep updated) any of the front-ends that I listed in my original post. I could be wrong but I'm not sure if LM studio can legally integrate GLP licensed software in the same way that Gerbil can because it's a closed source app.

I cannot judge the validity of your test and I have done any tests myself. I encourage you to read the link from post to the top firefox hardening resource (arkenfox) that labels NoScript redundant. This is further backed up by another top privacy resource: https://www.privacyguides.org/en/browser-extensions/ This also links to the uBlock docs which outline different modes. Medium/hard modes make NoScript redundant.

You're more trackable by using NoScipt and there's no good reason to use it if you know how to properly use uBlock: https://github.com/arkenfox/user.js/wiki/4.1-Extensions#-don... uBlock is a content blocker so it can do everything NoScript can if you learn its advanced UI usage. Using additional extensions makes you more trackable.

I might be a masochist, but I actually enjoy system design interviews and they're quite formulaic. Resources like "System design in a hurry" are great for narrowing down the formula.

Where I live (BC, Canada) actually has a law requiring all employers to list the position's salary range, which is great for cutting down on the "expected salary expectations" dance.

I don't like take homes as it's (highly likely) a one way time commitment and if you're truly looking to show off your skills it would take you hours.

Bingo. Discussing salary is a game and the first person to say a number loses.

This sounds like an incredibly toxic hiring process and not a company I'd ever want to work for. So you apply for a job and in response they (maybe) email you back asking for your expected salary (great way to filter out anyone worth hiring btw) and if you're cheap enough they then ask you to do work on a take home assignment. Everyone here thinks that this is okay and they want to be interviewed this way?

"Since we are focused on efficiency, we need to respect people’s time as much as our own". How exactly does this process respect the candidates time?

Location: North Vancouver, BC, Canada

Remote: Yes

Willing to relocate: No

Technologies: Next.js, JavaScript/TypeScript, Node.js, Electron, React

Résumé/CV: https://egor.philippov.ca/resume

Email: egor@philippov.ca