Learn Krita. David Revoy, artist of Pepper and Carrot, has a ton of great tutorials on using it.

Blender

I know you said software, but I would like to suggest a musical instrument. They still require learning a new language and typically provide a huge level of self satisfaction. A cheap keyboard or guitar can provide literally thousands of hours of study.

OpenSCAD. 3D modeling as code.

I've gotten a lot out of it, and their tutorial is excellent.

Latex is worth knowing, its most commonly used for academic papers, but it opens up a whole world of scripting automated documents. But, to make it really interesting, throw in tikz and/or pgfplots for some interesting power tools. Tikz does A LOT, and is really powerful. But can definitely be overwhelming.

FreeCAD

Docker and Docker Compose. Once you know that, you'll be able to self host almost any FOSS.

My suggestion already submitted was Blender, but any of these ideas require a purpose. I don't think you truly learn an application or language (even human languages!) if you don't have something you're trying to do with it. Even music (it's a language), you can learn how to play something, how to put together notes and sounds, but you really start figuring out how to do things when you have a need to accomplish.

I would say 90% of my coding and application knowledge comes from decades of having a need or inspiration to do something, and having to find the right tool and learn how to use it to make my idea become something (even if it doesn't get finished). When that purpose or desire disappears (for whatever reason) I stop wanting or caring about learning more about the tool I was using to do it with.

Your initial request seems to point to a different style of learning, wanting to do it for the journey itself. That's cool, I just wonder if you had a reason for doing it, would it make the effort more fun and worthwhile.

Godot. Start with an idle game, move onto more. Just don't make a horror game, way too many of those and a lot of them are bad.

Factorio

I'd say GIMP for photo editing would be a good one to learn. It's come in handy in a pinch both at home and at work when I've needed something done up real quick.

vi

https://buttplug.io/

Ardour, it's a FOSS Digital Audio Workstation

https://ardour.org/

Bash. If you use Linux.

OpenFOAM. If you ever want to figure out the aerodynamic qualities of a cow. This may take more than a weekend.

Get your finances in order with Firefly III and Waterfly III (Android app, available in F-Droid)

libreoffice if your using excel.

Blender and buy a 3D printer afterwards

Kubernetes (ala k3s) you can make any collection of computers a cluster and play with automation of apps and services

KiCAD play with circuit board design with FOSS.

FPGA programing! Tons of use case for when you need near line speed but can't afford the time and money it takes to build a full ASIC.

Ansible

You can do really neat thinga in QGIS -- like make your own professional quality maps, make map animated gifs, etc

Try Proxmox Virtual Environment with community post-install scripts. It's a little virtual data center that you install on a dedicated small PC. You can run all kinds of services like a Linux VM, Pi-Hole, Copyparty, Romm, and stuff. It'll make !selfhosted@lemmy.world very happy.

SQL - understanding how to interact with databases will be an incredibly useful skill for whatever rabbit hole you next want to fall down.

Python, and you kind of want to take the long way around.

Get it up and running in its own virtual environment. On Linux and Windows this'll just be a directory its in, and once you've activated that install via a terminal command everything stays self contained. It has a package manager called "pip" that will handle package and dependency management. Use pip to install spyder, pandas, matplotlib, numpy, scipy, and openpyxl. If you install anaconda this is what it's doing under the hood, but there are licensing hiccups with using their package repos.

Anyway, however you get up and going, numpy is fast numerical storage, scipy is a lot of scientific algorithms, pandas is a data analysis library that rides on top of numpy, and openpyxl is an interface to excel files from python. Pandas will get you one line CSV file reads and writes and more complex manipulation of Excel spreadsheets. Openpyxl gets you cell by cell manipulation of a spreadsheet. Spyder is a development environment.

It can do much, much more.

I asked an LLM for a small python example with pandas and matplotlib. Load it into spyder and run it and see what happens (tip, go into the settings/preferences, IPython Console, Plotting, and change the Graphics Backend to "Qt" to get the plots in their own window).

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt

def generate_damped_signal():
    # 1. Setup Parameters
    fs = 50                 # 50Hz Sampling frequency
    t_max = 6.0             # Run for 6 seconds to clearly visualize it hitting 0 at 5s
    freq = 2.0              # Oscillation frequency in Hz
    
    # 2. Generate Independent Time Axis
    # 50 samples per second from 0 to t_max
    time = np.arange(0, t_max, 1/fs)
    
    # 3. Generate Dependent Damped Signal Axis
    # Using an exponential decay constant of 1.0 ensures that at t=5 seconds,
    # e^(-5) drops down to ~0.006, effectively decaying the signal to zero.
    amplitude = np.exp(-time) * np.cos(2 * np.pi * freq * time)
    
    # 4. Create Pandas DataFrame
    df = pd.DataFrame({
        'Time_Seconds': time,
        'Signal_Amplitude': amplitude
    })
    
    # 5. Export DataFrame to CSV File
    csv_filename = "damped_signal.csv"
    df.to_csv(csv_filename, index=False)
    print(f"Successfully generated DataFrame and saved to '{csv_filename}'")
    
    # 6. Plot the Data Using Matplotlib
    plt.figure(figsize=(10, 5))
    plt.plot(df['Time_Seconds'], df['Signal_Amplitude'], label='Damped Signal', color='cyan', linewidth=2)
    
    # Visual Anchors for the 5-second decay mark
    plt.axvline(x=5.0, color='red', linestyle='--', alpha=0.7, label='5-Second Decay Target')
    plt.axhline(y=0.0, color='gray', linestyle='-', alpha=0.5)
    
    # Labeling and Grid Configuration
    plt.title('Damped Signal Decay Over Time (50Hz Sampling Rate)')
    plt.xlabel('Time (Independent Axis - Seconds)')
    plt.ylabel('Signal (Dependent Axis - Amplitude)')
    plt.grid(True, linestyle=':', alpha=0.6)
    plt.legend()
    
    # Display the Plot Window
    plt.show()

if __name__ == "__main__":
    generate_damped_signal()

Obsidian (Free but not OSS) and PKM. Might help you improve how you build your knowledge, or might just trap you in an endless hunt for the perfect system.

NixOS if you’re looking for an endboss kinda thing.

Learn the basics of accounting by trying to use GnuCash (or KMyMoney) for your personal finances.

Or learn a proper text editor, such as vim.

ssh and scp (secure connection and file transfer)

grep (the standard command-line search tool)

sed (Stream EDitor, good for find-replaces once you learn how to use it)

awk (can operate on text files or command output that's organized in columns)

sc (terminal-based spreadsheet)

trash (safer replacement for rm that moves things to your desktop's trash directory instead of deleting immediately)

You could get JOSM and become an OpenStreetMap editor. https://josm.openstreetmap.de/

Also install gentoo.

Learn the basics of a couple of FOSS audio-software staples.

LMMS - tracker

Audacity - Audio recorder

wow rekordbox looks nice! I've used Mixxx up til now.

One killer feature might make me switch: I see rekordbox allows mixing of spotify or apple music songs: do they also support navidrome? That would blow me away!

about the topic: what about home assistant? That should prove an adequate rabbit hole for you :)

KiCad takes around 5 minutes to learn and is some of the best software for designing schematics and PCBs.

Falstad is great for when you need simulations of schematics and don't want the headache of trying to use KiCad's built in simulation tools.

Both are FOSS.

CRRCSim (RC plane simulator)

Learn how to slope-soar a glider.

Best with the photo mountain scenery added in.

In many distro repositories too.

https://github.com/gcmcnutt/CRRCsim

Slackware- Wait you said fun, nvm

Aseprite (you gotta compile it yourself if you wanna it free tho)

Lilypond. Its music software. Learning it will keep you occupied for at least a week.

One week of keybr.com to improve touch typing skills, which should serve you across all your IT endeavours... Unless you choose to learn gimp and drawing stuff :P

GIMP

AI image editing can only go so far.

I learned a decent bit about how biometric authentication works (or at least this specific implementation) by installing, setting up, and tweaking settings for Biopass. I also learned about how when you do first login using this method, it can fail to unlock keyrings because the password authentication has been bypassed so no password has been provided to unlock the keyring. It includes anti-spoofing local AI models, and the dev plans to have the option for the user to download and use any model for anti-spoofing/recognition. Interesting stuff.

https://github.com/TickLabVN/biopass

I recommend learning dc if you don't already know it, and either Octave or Python.

https://en.wikipedia.org/wiki/Dc_%28computer_program%29