From useful to harmful: How the recommendations of my artificial intelligence destroyed my operating system
Why should you always be in control?
TL; DR: Always stay control when using artificial intelligence tools. Blind Trust can lead you to expensive errors.
Challenge
Write multiple articles every week and want to automate the frequent parts of my work.
Risk
This week, an internal note by the CEO of Shopify, Tobias Lutke, viral.
https://x.com/1909251946235437514?
TL; DR: The memo requires that Shopify workers be mastered in artificial intelligence tools.
The efficiency of artificial intelligence is mandatory now
Using artificial intelligence effectively is no longer optional in Shopify; It is a basic expectation for all employees, regardless of the role. This represents an important cultural and operational transformation.
Failure to use artificial intelligence does not require justification
Employees should explain why artificial intelligence is not used before requesting additional resources (for example, more employees or time). The recession is framed as a failure, and employees are encouraged to raise the level of artificial intelligence constantly.
Artificial intelligence in the development of products
Artificial intelligence should be combined in the early stages (the initial model stage) for all GSD projects (obtaining cinnamon). This approach aims to accelerate learning, repetition and the team’s cooperation greatly.
Performance accountability
Shopify adds the use of artificial intelligence to its performance criteria and peer review standards, which makes artificial intelligence adoption part of how employees are evaluated and rewarded, including leadership teams and executive teams.
I am not working on Shopify, and I don’t prevent the company.
My context
I take artificial intelligence tools as assistants and mandatory tool as a software developer and a writer, and I relied on them very much.
I used to use artificial intelligence for 10 years, write about artificial intelligence, make AI-AI-Essisting Renewal, and use automatic tools to write clean software instructions.
I am also a defender of professional coding.
In each conferenceThe article, and the paper I write, encourage people to adopt and use artificial intelligence tools (with caution).
https://haackernoon.com/ai-coding-tools-are-Bad-News-for-laty-Programs?
I have two quotes (not mine):
You should always be in control.
and
Artificial intelligence will not take your job, but the person who will definitely use it; You should become this person.
(You can see this in 53:40 From this conference)
https://www.youtube.com/v/9GUXTIW0R4?
Don’t think about things
This week, you need to perform some local tests using Python.
I write articles in more than 25 different programming languages and are not all installed locally.
Sometimes I use virtual devices, codelabs, and replacing Run on web sites.
This time, you need to install Python locally (transverse mission) quickly.
I started with a simple idea: Automation of some things. Save time. Be smart.
So I did what many people are doing now.
Amnesty International quickly answered. Confident. As I did a thousand times before.
The spoiler: I was blinded to him blindly. One thing later, my operating system was toast.
What is wrong?
In my code scent articles, I use 10 most popular artificial intelligence programming tools.
I just opened one of them (Chatgpt) Pay it to give me some orders to install Python on my Ubuntu.
Give me an order to install Python in Ubuntu
Until last week, I used to write:
Hello, please give me an order to install Python in Ubuntu
But now I realize that these additional symbols consume a lot of resources (water, for example), and I am no longer polite for machines anymore. 
The answer was short and clear:
Sudo Apt Update && Sudo Apt Aptrade -Y && Sudo App App Install Python3 Python3 -Pip -Y
I did not bother to read myself.
I have been using Linux for more than 20 years yet.
I just paste the answer. I wrote my password mechanically (I was using SUDO to install PIP), and I went to dinner.
When I went back, the station showed a very long list of orders. It was strange. I just wanted to install Python (with PIP, which is the package manager).
The problem (as I have discovered now) was this innocent line:
Sudo Apt Upgrade -Y
Which is upgraded (and bricks) of my personal computer.
Chatgpt decided that it was a very good decision to upgrade the operating system first.
I agree that keeping the updated operating system is a good decision to get the latest protection, but this is something that you need to plan, and get a good backup, not something you do accidentally.
The regime was long, and it was.
A careful look at the matter (and the question to the other ias) will give a more accurate (and safer answer):
SUDO APT update
Why happened to me?
I had a very long experience with Ubuntu. You are completely able to read a forum or Googling (from Google anything today, by the way?).
I am also a preacher, I do not trust AIS blindly, and apply it Every article I am writing.
This should not happen to a gay person.
We reptiles
Our brain is the master of efficiency.
Every decision we make consumes energy, and therefore the oldest part of our brain, the so -called “reptile brain”, has evolved to improve survival, not for deep thinking.
She wants quick reactions, less effort, and zero sugar waste.
For this reason he loves authorization.
But what happens when the thing he delegates is a machine?
You do this every day: track GPS instructions without interrogating them, relying on automatic completion to finish your thoughts, let the recommendation engine tell you what to see, buy or read.
On the surface, this looks like comfort.
Under it, it is a biological abbreviation.
Decision -making resources in your mind are limited. To avoid fatigue, it separates the options whenever possible.
This is not always bad. The routine and inference allows you to work in a complex world without overwhelming it.
But when this primitive reflection controls, you stop the question “Why?” – This is when things become risky.
The illusion of efficiency
One of the most dangerous aspects of the delegation of artificial intelligence is the illusion of efficiency. The device that speaks correctly or coordinates or generates the work code looks smart. Sometimes you are smarter than you.
But behind the Polish may lie to shallow thinking, bad interpretation, or even hallucinogenic content.
Do not notice because the cognitive lasten starts.
You assume the right thing because it “looks right.” The deeper and analytical part of your mind – the frontal lobe crust – is never involved.
This is exactly how wrong information is spread. It is also how errors are spread, contracts are wrong, and systems failure.
The mandate without accountability
Human cost of misplaced intelligence decisions is already present here: people who get health care, report, or imprison, or prison based on algorithm predictions. In these cases, the real danger is not just Amnesty International – it is an incomplete acceptance of it.
When you delegate a decision to a system that you do not understand or achieve, an agency surrendered. Worse: I have been accountable.
“The algorithm said,” it becomes a shield. But it should not be.
How to fight reptiles inside you
The delegation is not bad by its nature. But you have to make it conscious. This means:
- Check out any Amnesty International that it uses, Especially for critical tasks
- Understand the limits of the system that you depend on
- Restore responsibility for decisions, even when the tools facilitate
Ask yourself:
Do I deal with this because it is effective – or because I am lazy? “
This question alone can re -involve your rational mind.
conclusion
Your mind wants to survive.
It does not care whether your program works, your model is fair, or that your report is accurate.
He just wants to cancel the voltage.
I worked well in the forest.
But not in a world where decisions have digital consequences.
The delegation without checking is not efficiency. It is a mixture.
And if you do not control your decisions, then someone-or something will.
