This might seem like an obvious question but the more I read about peoples experiences writing code with AI and LLMs, I find increasingly more difficult to understand the details of what is happening.

There are claims that people aren’t writing code manually any more and instead deploying multiple AI agents to do the work. This seems crazy to me and I genuinely have no idea what this looks like on the ground. I’d like to be proven wrong here, so…

What specifically does your day look like in this case? What is the nature of the work that you work on? Are you ignoring cases where it goes wrong? Or is that factored in to this mode of working? What are the downsides or upsides?

On the flipside, AI skeptics, do you use AI in any capacity? And if so, in what way?

The more detailed the answers, the better.

EDIT: As I posted this, I received the message from management that we are fully embracing Cursor from 2026 onward and are mandated to be AI-first. I’m leaving.

I know any tool should do the job and the editor you use shouldn’t affect your ability to do your tasks as much, but I’m doing this full time and it’s becoming a daily inconvenience.

I’m mandated to use VS code and Sourcetree. Both great tools, but I live inside the shell. Their workflow is good, but not for me. I have asked for a reason and they gave the following:

- They want to prevent mistakes from happening –> So instead of responsibility, they introduced a seatbelt
- They want me to be able to help others, as well as have them help me. If I use different tools, that becomes harder –> We can just open GitLab, or, I don’t know, open a different editor when someone is looking at my screen :)

I’ve already addressed this multiple times and it starts to gnaw at me. …

I was brought into a company to lift and shift their application (Java 21, no Spring) to the cloud. We’re 6 months in, and everything is going relatively smoothly. The team is working well and we’re optimistic to get QA operational by the end of Q3’26.

My next big task is assembling a team to migrate the stored procedure nightmare that basically runs the entire company. There’s 4 or 5 databases each with ~500 stored procedures running on a single Microsoft SQL instance. As you can imagine, costs and latency balloon as we try to add more customers.

The system is slightly decoupled, HTTP requests ping back and forth between 3 main components, and there’s an in-house ORM orchestrating all of the magic. There’s nothing inherently wrong with the ORM, and I’d like to keep it place, but it is responsible for calling all the stored procedures.

The final component/layer is responsible for receiving the HTTP requests and executing the …

Currently we give some of our devs write access to prod dbs but this seems brittle/undesirable. However we do inevitably need some prod queries to be run at times. How do you all handle this? Ideally this would be some sort of gitops flow so any manual write query needs to be approved by another user and then is also kept in git in perpetuity.

For more clarity, most DDL happens via alembic migrations and goes through our normal release process. This is primarily for one off scripts or on call type actions. Sometimes we don’t have the time to build a feature to delete an org for example and so we may rely on manual queries instead.

So fifteen to rwenty years ago some engineers provisioned some servers and then were allowed to retire without passing on administration roles or knowledge. By the time we got management on the “succession planning is important” page the horses had already left the barn.

One of the servers hosts SVN source control used by all our projects as well as the license server for some embedded compilers we use, and the other runs a web app used nationwide. Government work, I’m being vague not because it’s secret but just to keep things at a non-details level.

In government work, teams do not own our own IT and maintaining it is a pure cost for the internal team or external company contracted to do that, and the benefit of what is running on it is not known or a fuck given by the ones hosting. This year, that IT org was like, “your servers are on a really old version of windows; we’re gonna turn em off. k thx bye.”

We had to beg for extensions. …

Just updated the image with a fix for the particles!!

Edit again: Thank you SO much everyone! this has been so incredibly dumb and fun. I can’t believe we’re about to hit 100k nodes 5 hours after me posting this. You’re all very cool and i appreciate everyone that helped me fix it and made pull requests. cant wait til we hit 1 mill and i steal all your ram ♡

Hey everyone, so you just finished setting up the *Arr stack and your dashboards lookin crisp. But you look at your htop and see… unused RAM.

It’s disgusting, isn’t it?

So I built Hypermind.

Hypermind is a completely decentralized, peer-to-peer deployment counter. It does exactly one thing: It solves the critical infrastructure challenge of knowing exactly how many other people are currently wasting 50MB of RAM running this specific container.

That’s it. That’s the whole app.

Despite being useless, the tech stack is actually kind of neat.

• No Central Server: This runs on the Hyperswarm DHT …

It’s a web-based tool that generates a single copy-paste command or a distro-specific shell script to bulk-install your entire setup.

Why use it:

• Native Support: Ubuntu/Debian, Arch (including AUR via yay), Fedora, openSUSE, NixOS, Flatpak, and Snap.
• Universal: Integrated Flatpak and Snap support.
• Smart Scripts: Includes network retries, progress bars, and ETA - not just a list of names.
• Fast UI: 150+ apps in 15 categories, fully navigable via Vim keys (h, j, k, l).
• Open Source: GPL-3.0.

Live: tuxmate.com

GitHub: github.com/abusoww/tuxmate

P.S. I know the URL is a bit clunky right now. Buying a proper domain name is next on my list!

EDIT: took the advice and bought tuxmate.com we official now!

These things are brilliant for sticking to cables. Found a £8 cartridge for my old Dymo D1 label maker.

As we wrap up 2025, I wanted to share my complete self-hosted setup and see what everyone else is running!

I’d love to hear what you’re all running - drop your stacks in the comments! What new services did you discover this year? What’s been your favorite addition?

Here’s my list of self-hosted services:

1. AdGuard Home - DNS Ad-Blocking & Network Protection - GitHub
2. AdGuardHome-Sync - Sync AdGuard Home Configs - GitHub
3. Apprise - Push Notification Aggregator - GitHub
4. Audiobookshelf - Audiobook & Podcast Server - GitHub
5. Backrest - Backup Solution With Restic - GitHub
6. Bazarr - Subtitle Automation For Sonarr/Radarr - GitHub
7. Booklore - Book Discovery & Tracking - GitHub
8. Book Downloader - Automated Ebook Acquisition - GitHub
9. Caddy - Reverse Proxy & SSL/TLS Termination - GitHub
10. Code-server - VS Code In Browser - Web-Based IDE - GitHub
11. CrowdSec - Security & Threat Detection Engine - GitHub
12. DAPS - Docker Automation & Management Scripts - GitHub …

Hey everyone! 👋

I wanted a clean, fast, and modern web interface for my IPTV service that I could host myself. Most existing players I tried were either clunky, outdated, closed-source, or just didn’t handle large playlists with thousands of channels very well.

So I built NodeCast TV.

📺 What is it? A self-hosted web application that lets you stream Live TV, Movies, and Series from your Xtream Codes or M3U provider directly in your browser. It’s built with performance in mind and handles large libraries smoothly.

✨ Key Features:

• Live TV & EPG: Full grid-style TV guide with 24h timeline, category filtering, and search.
• VOD Support: Dedicated sections for Movies and TV Series (complete with season/episode browsing).
• High Performance: Uses virtual scrolling technology to render lists with 7000+ items without lagging your browser.
• Favorites System: Unified favorites list across all content types.
• Universal Player: Built on HLS.js for robust playback support.
• Docker …

For whom it may concern. AltHub announced on their website a discount for it’s 13th Birthday that will “start 07:00 UTC on 1 January 2026 and ends 19:00 UTC on 5 January 2026 or until stocks last”.

The usual 20$ Lifetime and 50% on other plans. Enjoy!

P.S.: I’m not affiliated with them, I’m just a newbie who has been waiting for that discount since I missed BF’s deal.

Years and years ago, my ISP also provided binary newsservers. I used Forte Free Agent back then to navigate throught the binary newsgroups back in the days. Talking about 20 years ago… my ISP removed the newsservers and I also stopped using usenet.

But now I want to give it a try after all these years. Not for posting, but more for looking what’s out there. I do have some questions, though:

1. I see that Forté does not have a free newsreader anymore? What user friendly (binary) newsreader can you recommend? One where you can also search?
2. What binary usenet provider can you rcommend? Which one has good download speeds, long retention and a lot of (binary) newsgroups and is not too expensive?

Hi, anyone know how I can get in contact with the althub team to help fix and issue. Thanks

Hello, I’m really new to usenet.
I recently bought Frugal and NzbGeek. I’m finding that older files are hard to find complete. For example, stuff from like 2010 (maybe this is too old for most providers?)

If I wanted to be able to more successfully grab files that are that old, would it be better to add another indexer, or another provider on a different backbone (something like Eweka or UsenetServer)?

Edit: Thank you, everyone! I started by adding NewsHosting and that seemed to work really well! I may ditch frugal in the future depending on how much is grabbed from there. I’ll also be on the lookout fo drunkenslug invites.

Just signed up with Miatrix and have a Trial status on my account.

Does anyone know what the Free account limits are?

Can’t find this info anywhere on their website.

My in-laws are CONSTANTLY complaining about having to pay XYZ streaming service to watch anything. Every Christmas, they come over and say how cool it is that I can essentially watch whatever it is and pay nothing. I even have Overseerr and everything completely implemented. Everything “just works.” Finally, this year, I asked if they wanted me to give them remote access and they were like “Yes! Please!”

Onboarded them without much fuss and got them logged in. They got home 4 days ago and said they got the app on Roku and they’ll check it out.

AND………..nothing. No activity. Zero. And it isn’t like they are busy. I know for a fact the whole household is still home on vacation from work.

This has happened to me before, too. Had another close family member who got logged in and then never used it again. Six months later they were complaining that they couldn’t watch a movie because it was $13 on Amazon. I was …

I finally put the finishing touches on my Plex setup and wanted to share my experience in case it helps anyone else who’s going back and forth on hardware decisions.

I started back in January 2025 when I bought the Plex Lifetime Pass (best decision ever, especially when it was cheaper). From there, the big question was how I wanted to run my server while keeping costs reasonable.

At first, I tried running Plex on my 2014 Mac Mini. It worked for a bit, but I kept running into random issues and eventually decided to move on. I seriously considered getting a NAS, but I just couldn’t justify dropping that much money all at once.

I then picked up a Dell Optiplex 7050 Micro, which did work, but I quickly realized I wanted more power and headroom. In the end, I decided to run Plex on my personal desktop, a Dell Inspiron 3880 with an i7 and 64GB of RAM, just running Plex quietly in the background. That decision ended up being the sweet spot for me.

For storage, I bought two 2TB Samsung EVO …

I do a semi-frequent check on what’s been added to the vendor list. Despite opting out of everything, you might be automatically opted in with new vendors to Plex as I found out today. Update your privacy settings!!

https://www.plex.tv/vendors/

So far I haven’t seen any flaws. It allowed me to watch Mr. Robot and old movies that I never thought I’d be able to.

The worst thing they ever invented was streaming.

Long live Plex!!!

UPDATE: The honeymoon is over… not totally surprised… Every music manager has its quirks…

PlexAmp uses folder structure for Artist/ Album name determination; which overrides tags in the actual files

So my folders full of singles, show up in one Various Artist album, that gets the title of the first file’s tags.

Horrible for use with my media library. I could script something to put every single (and I have thousands of them), into their own artists/album folders, but not sure it’s worth it.

Ugh.

When digital music really took off, around 1998; I ripped all my CDs ahead of an international relocation; and have been keeping my personal collection updated, even through the streaming era.

I have a massive favorites playlist that I generally just pause and play; takes around 2 or 3 months to get through. If I want mood or theme playlist; I’d use Spotify’s radio features.

I’ve been a long time Plex video user ; but assumed it music …