Wrap the AI terminal

A new era is coming — the era of AI. I truly believe this will be remembered as a turning point in the history of computer science (if not even in human history). But who knows... maybe it's just a mix of marketing hype and exaggeration.

Many applications now come with AI features, and today, I want to talk about one of them: The Warp Terminal — the intelligent terminal.

Sounds good, right?

At first, I thought it was just a cool little tool. In the terminal world, you have to type a lot of commands. Often, you don’t know how to do certain things, what the right commands are, or the best way to solve a problem.

Normally, people go to Stack Overflow or Google for help. But if you can just ask your terminal, without switching to a browser, that sounds pretty convenient and useful, doesn’t it?

And it is. But there are a few buts I didn’t think about in my initial excitement.

Becoming more dependent

Using ChatGPT, Claude, or any other AI can be very helpful. However, not only do they sometimes hallucinate — the real problem comes from the lazy side we all have.

If AI writes the full command for me, why should I bother understanding the details or the logic behind it? I believe the best way to use AI is to ask it about its decisions, to learn more — like having the best teacher always available.

But our inner laziness eventually pushes us to stop asking, stop learning, and just use the output. It usually works. That’s good enough, right?

I think this is a dangerous trap. It’s not an AI problem — it’s a human problem. Our problem.

Your privacy on the Internet

Even if the last issue was the main problem, I could live with it. You just have to train yourself to keep asking the AI questions, instead of blindly using the answers. Maybe it's just a matter of discipline.

But the second issue — and the worst in my opinion — is how AI works, at least in 2025. It works over the internet. That means when you ask something, the Warp terminal sends your data — including your passwords, environment variables, and all kinds of other information from your computer — to Warp’s servers.

Not only that. Warp is a closed-source project. Sure, they have a GitHub page, but honestly, have you checked it? It’s empty! Just a placeholder saying the code is proprietary.

I don’t feel comfortable sharing data that should stay local with a closed-source project. I don’t know how they handle my data.

And I’m not talking about conspiracies or “bad guys.” No. Maybe the people at Warp are amazing — I don’t doubt that. What I doubt is their ability to prevent security holes.

Even the brightest minds in the world make security mistakes. It’s impossible not to. So the Warp team will too — and if they don’t share their code, it probably means there are more risks, not fewer.

And that... that scares me.

Final thoughts

Well, my two cents: it’s not worth the risk. It’s a really beautiful terminal, with great UI/UX. Honestly, I love it more than any other terminal I’ve used.

But I want to become a skilled Linux user someday, a productive programmer, and above all, a professional and responsible one. And to be all those things, I need to learn how to do things on my own. I also need to be careful with risks and security holes.

I won’t share my data freely over the internet.

Thanks for reading.

Moving files

Today I had to move a bunch of files from one directory to another, and I liked how I did it.

You can do it in the simplest way like this: mv file1 file2 file3 /home/your_name/directory

But my case was a bit more interesting.

Create a buckup of .bash* files

In my home directory, I had these files:

~/.bash_history
~/.bash_profile
~/.bash_profile.bak
~/.bash_logout
~/.bashrc

Since I'm using zsh now, I wanted to back them up just in case I need them later.

So, I created a folder called .bash_backup and wanted to move all the .bash* files into it.

But how can you move all of them with one command?

I couldn’t just type mv .bash* .bash_backup because .bash* also matched .bash_backup if it already existed.

One way to check what will be matched is to run:

❯ echo .bash*
.bash_history .bash_logout .bash_profile .bash_profile.bak .bashrc

So, I needed a better way. That’s where the find command helps.

The find command lets you filter the results and run a command on each file it finds. Here's what I used:

find . -maxdepth 1 -type f -name ".bash*" -exec mv {} .bash_backup /;

Here's what each part does:

• -exec runs a command on each file found.
• mv is the command to move the file.
• {}is replaced by the filename found.
• .bash_backup is where the file is moved.
• /; ends the -exec command.

With this, I was able to move all .bash* files to the .bash_backup folder in one command.

fzf

This tool, with name coming from "fuzzy finder", use a technique of the same name.

You can use fzf to search for text using a vague description. For example, if you're trying to find main_component_renderer.cpp, you can just type mcr.

To do this, just type fzf to open the tool, and a bunch of files will be shown.

At this point, type the full name or a fuzzy match (like mcr), and the first match will be selected. ust press ENTER when you find it, and the full path will be printed.

The fun part comes when using it with pacman, especially for searching packages.

Usually, to search in pacman, you use pacman -Ss package_name, but this often shows many packages with similar names or descriptions, which can be messy.

Instead, you can list all available packages using: pacman -Slq

Where:

• S means sync mode
• l lists all packages
• q means quiet mode (no extra formatting)

Then, you can pipe that into fzf with the --multi option.

This is useful because fzf will show a scrollable list where you can type to search, and use the TAB key to select multiple items. Press ENTER when done.

Finally, pipe the selected items to xargs -n 1 to run a command for each selected item:

pacman -Slq | fzf --multi | xargs -n 1 sudo pacman -S

This will install each selected package by name.

Cool!