I started a new job. We have all of the AI tools etc.

I got assigned a new feature that seemed like a simple extension to our architecture.

The Claude plan looked reasonable, but upon seeing the code it appeared Claude had complicated the hell out of the implementation. It looked suspicious so I had Codex plan it separately, and it comes to the same conclusion. Every time I tried to simplify it, they both pushed back and argued for the additional complexity.

Now usually, I would take my time to untangle the chain of thought and identify the assumptions making AI do what it’s doing. But the customer has been promised this feature would be delivered (not by me - but you know how it is).

It appears to work. Unit and end to end tests look good. But I still do not fully grasp all of this code and why it’s there.

In today’s new fast paced world with insane expectations, would you push back on releasing it due to not fully understanding why all of that code is there? I did, but luckily, …

Hi all,

I’m a 7 YOE hardware/embedded engineer who recently joined a Fortune 500 company after a few years in an Research & Technology Center. It has been a few months that I’m working in a new environment, trying to ramp up as fast as possible, however I’m always surprised by some lack of professionalism:

• No one seems to care and somehow tries to wing it, the more vocal the better.
• When I ask people around, they don’t know, so I’m on my own to learn everything from scratch again.
• Whenever I try to take ownership to solve some problems like electronics problems & understand the root cause, people tell me it’s out of my scope and I should not tackle it. However, it still affects what & when I can deliver my software.
• Design specs, processes, reviews, etc. seems to be not applied anymore.
• Everything takes age to do, it is always urgent and nothing is documented.

The main purpose of this post is to ask engineers seasoned to corporate environment: …

I’ve got 4 years of experience at the same startup. I was a junior when I joined and there were two others on the team - one mid-level engineer and one senior (who was titled CTO, but that’s a bold term to use for a company our side).

After one year in, the business went through some financial turmoil. They let go of the other two engineers, including the CTO mentor/manager, which was pretty brutal. The founder did tell me there was friction between him and the CTO, but also felt that I could lead the technical side of the business. This was always expected to be a temporary measure while finances were taken care of, and they promised they would hire a new senior to join me. I pushed back quite a bit, and they did eventually hire a temporary fractional CTO to support me.

Today, there is still no senior who has joined - it’s just me. They’ve gotten rid of the fractional CTO too. They gave me the title of CTO (yes again, I know it’s a bold title for a one man …

I’m a software engineer with about 7 years of experience, in the financial sector. Most of my work is with .NET APIs and SQL Server. I’m currently working on a project involving internal employees who provide clearing services. The workflow is roughly:

1. A user selects one or more records.
2. They enter some additional data.
3. They submit the request.
4. My API validates the request, including checking permissions and whether any of the selected financial records are already pending.
5. If validation passes, the API generates a CSV file.
6. The CSV is uploaded to an SFTP location for the business.
7. records are written to the database.

In some cases, two CSV files are generated. To avoid filename collisions, I timestamp the filenames down to milliseconds, for example: “2026_2111056_664”

We’ve been finding bugs with unique test cases after releasing to UAT, and today we got on a call to test some scenarios together. Another developer suggested submitting the exact same …

In 2023 I took 6 months off because I was burnt out after 5.5 years. I hated my job, and I hated the work too. During those 6 months I recovered my enjoyment of the work ( programming ) and found a new job, which I’ve been employed at for the last 2.5 years.

I hate this job too now. Even more than I hated the one I left in 2023, for different reasons. The retarded way the company is forcing us to apply AI plays a significant part in it, but I’m not going to go into details lest I get the AI astroturfing bots swarming this post.

I look at job boards with disgust. I hate this job, but I hate the thought of switching companies only to end up in the same predicament even more. I’ve been burnt out for the last year, and I can feel myself reaching my limits as to how much longer I can force myself out of bed every morning for it. Last week I started taking sleeping pills for the first time in my life because of stress-induced insomnia. Both canine teeth on the right side …

I’m wondering how many people still run their homelab in the most basic way possible: a normal Linux install like Ubuntu Server or Debian, Docker, and one folder with a docker-compose.yml file for each service.

I’ve tried things like OpenMediaVault and CasaOS before, and I totally understand why people use TrueNAS, Unraid, ZimaOS, CasaOS, Portainer, etc. The UI is convenient, a lot of things are simplified, and it’s probably much less intimidating when you’re starting out.

But personally I always end up preferring Linux + Compose directly. I’ve also sometimes found it harder to troubleshoot problems on those other platforms.

Basically, my current setup is one folder and one docker-compose.yml file per service, with sometimes several related services in the same Compose file/folder.

I do have Portainer installed, but I barely use it. Mostly to quickly delete old images or volumes, or check something when I’m too lazy to do it from the terminal.

What I like about this setup is that …

What started as fun, interesting use of LLMs has, it seems to me, spiraled into something completely out of control. While new apps get created at lightning speed, especially if spearheaded by developers who have an idea of what they are doing, it’s made the self-hosted space from something with 2-3 options per media category into a very crowded, messy space.

Take a look, for example, at the book side : we now have Bookorbit, Grimmory. Completely new apps, and yet, somehow, they look the same. Why? Because the devs used AI during development. Take a closer look, and you’ll be able to make out immediately if an app was made with AI: bright pill-shaped buttons, often green, catchy “it’s not X, its’ Y!” descriptions, Docs with emoji, and the somehow same settings and menu layout.

And it’s not just a philosophical problem: how do people innovate if there’s nobody to make new stuff? and how about the old, beloved projects, with their …

I’ve been self hosting for quite some time now, I write software and I’ve seen many projects of all quality levels.

I just want to express my appreciation for Bookorbit and its developer, it’s truly a joy to use, and it clearly comes from a person that knows what they’re doing.

You built exactly the product I was eventually going to build myself, the attention to detail is incredible, it has the features I always wanted to have in other solutions, and they’re well put together.

This is also an occasion to showcase that AI is amazing in the hands of skilled engineers and we should take advantage of it in open source, hopefully while leveraging self hostable open weight models in the future (we’re slowly getting there, and we should fight for it).

I see so many people in here using Proxmox. I run Ubuntu + Docker, and my whole setup is just a list of folders with compose.yamls and supporting files (configs etc). It feels very manageable this way, fully declarative, easy to version control.

Am I missing something, that makes people recommend and use Proxmox instead?

Source: https://gitlab.com/g33kphr33k/kitchen-board

Docker: https://hub.docker.com/r/g33kphr33k/kitchenboard

I’ve been working on Kitchen Board on and off for roughly 5 months now.

Think of it as the old school planner your folks probably kept in the kitchen, on the wall or the fridge. I’ve tried to keep that same feel, but updated for today’s use, on yesterday’s hardware. The image is it running on an Amazon Fire HD 8th gen in Kiosk Mode (from F-Droid store).

Initially this was just a calendar with events, a shopping list for when things were used or eaten, and a countdown to events for my son, who is addicted to knowing how many “sleeps” until something happens. I posted a while back and had some good feedback, and since then people have reached out directly and via GitLab to add functions and features they wanted themselves. I only implement features that would be useful to my own family, and so far the suggestions have been great. …

NZBGet v26.3 is out! This update brings new features like extension translations and per-feed SSL options, updates core extraction tools, and packs in a ton of crash fixes and WebUI improvements for a smoother overall experience.

Features

• Extension i18n support
  • Enables translations for the following extensions #872 Updated versions are available in the Extension Manager
    • ExtendedUnpacker
    • EasySort
    • NotifySickchill
    • VideoSort
    • Completion
• Per-feed certificate verification
  • SSL verification option can be set per feed #846
• 7-Zip and UnRAR updates
  • Updated 7-Zip to 26.02 and UnRAR to 7.23 #847

https://preview.redd.it/9uzfh7ztyxlh1.png?width=1701&format=png&auto=webp&s=8dd770733427ea3060edda0225861f947f04922f

Fixes

• Security
  • Improved security for generating web tokens and handling extracted symlinks #894
• WebUI
  • Fixed a WebUI crash on startup with non-English language #845
  • Resolved background glitches #834
  • Completed leftover text translations for Polish and …