Featured

Building and Training Your Own SLM

on
Creating an independent Small Language Model (SLM) is a rewarding project that bridges the gap between deep learning theory and practical, local application. By keeping your model local, you retain full control over your data and system architecture. Below is a structured approach to building and training your own model from the ground up on a local Ubuntu environment. Before you begin, ensure your development environment is optimized for local computation. A robust setup with a capable GPU (such as an NVIDIA RTX series) and sufficient RAM is recommended for efficient training. Use Ubuntu for a stable, customizable development environment. Verify your environment with the following commands:    * Update your package lists:  sudo apt update.    * Install pip:  sudo apt install python3-pip    *I nstall necessary libraries:  pip install torch tiktoken An SLM relies on the quality of its input data. For a personal AI, curated, factual...
0

I Turned ON All Ubuntu Telemetry.

on

I did something today that will make certain corners of the internet audibly gasp.

I didn’t disable telemetry.
I didn’t firewall it.
I didn’t put on a tinfoil hat and boot into a Faraday cage.

No.

I installed every Ubuntu data-donation tool
and opted in manually
like a lunatic
with intent.

Yes. Telemetry.
On.
All of it.

Step 1: Installing the “evil” telemetry tool

First, I installed Ubuntu’s main data-donation package:

sudo apt update
sudo apt install ubuntu-report

Then I looked at the data it collects:

ubuntu-report

And what did I see?

  • CPU model
  • GPU model
  • RAM size
  • Screen resolution

Oh no.
My computer… exists.

Step 2: Opting in aggressively

Not satisfied with a passive existence, I explicitly told Ubuntu:

ubuntu-report -f send yes

That’s right.
Not “ask me later”.
Not “maybe”.

YES. SEND IT.

Somewhere, a Canonical server blinked awake like:

“Another one has chosen… participation.”

Step 3: Package usage stats (aka “He installed VLC”)

Next up: popularity-contest.

This reports which packages are installed, not how you use them.

sudo apt install popularity-contest

During install, it politely asks if you want to participate.

I clicked Yes like a villain pressing a red button.

To check it’s alive:

systemctl status popularity-contest

It runs about once a week and basically says:

“User has Firefox. Again.”

Step 4: Crash reporting (because software crashes, shockingly)

Then I enabled Apport, Ubuntu’s crash reporter:

sudo apt install apport

Edit the config:

sudo nano /etc/default/apport

Set:

enabled=1

Then:

sudo systemctl enable apport
sudo systemctl start apport

Now when something explodes, Ubuntu can go:

“Ah. That’s why.”

Instead of:

“Guess we’ll never know.”

Step 5: Confirm I went full chaos mode

To verify I installed everything:

dpkg -l | grep -E "ubuntu-report|popularity-contest|apport"

If all three show up, congratulations—you’ve joined me in telemetry enlightenment.

What Ubuntu actually receives (brace yourself)

  • CPU: yes
  • GPU: yes
  • RAM: exists
  • Installed packages: known
  • Crash stack traces: occasionally

No:

  • Browsing history
  • Files
  • Keystrokes
  • Thoughts
  • Dreams
  • The thing you regret Googling at 3am

Meanwhile, Windows is out here like:

“We noticed you hovered over the Start menu for 2.4 seconds. Are you okay emotionally?”


Comments

Post a Comment