A positive example of using AI

How I used AI to ship and escape "tutorial hell" — and a more positive message about technological change

AI as we know it (ChatGPT and other LLMs) has been around for around 2 years now. And the impact is noticeable.

For some, it’s changed their job drastically and they cannot imagine their lives without it anymore (i.e. developers, writers, marketers, etc.).

For others, there’s some productivity gains (i.e. sales, business, management, etc.).

And for a third group, it’s more of a clever gimmick (i.e. people who don’t have a laptop job).

The impact of AI

It seems that LLMs have had more impact in 2 years than cryptocurrencies have had in 15. And because of how quick this adoption has occurred, it’s safe to say we’re all curious to know how it will continue to impact our lives even further.

Is it truly over? Will AI replace us at our jobs? Will everything we know cease to exist? Will our lives just be doing the laundry while AI moves the economy?

Or has it just begun? Have the barriers and bottlenecks to development been lifted, and we will live a golden age of software? Will we really start living in a utopia where all illnesses are cured?

Well, my room-temperature take is that it’s still too early to know. I’m no expert on how previous technology leaps changed things. It could go both ways at the same time.

Technological change in the past

Regarding technology change, I do remember having a class in the 2000s that changed my understanding on it. The long story short was that, while today we see electricity or train companies as boring and commoditized, back when they started they could be considered the ‘Big Tech’ of their day.

They were at the frontier, changing how the economy worked and how everyone lived their lives. Long gone were the days you light candles at night, or use a horse to go from one city to the other.

With time, they just became part of day to day life. The technology became more known and optimized. The problems got figured out. The frontier of what was considered new tech shifted.

“it’s so over” vs “we’re so back”

So AI (LLMs) are considered the frontier right now. They’re new technology.

They might just change everything and leave us jobless, collecting our UBI paychecks while we just do laundry. In the same way trains put out most long-distance horse carriages out of business.

Or they they might just become a smart API call that we make. Just another “unremarkable” aspect of our day to day. In the same way turning a switch on the wall turns on a lightbulb to light a room.

But it also might just be very exciting and change the way we think about things and open up many interesting possibilities.

If back then you couldn’t even think of going to a city a few hundred kilometers away, with the arrival of the train you could. And then later with the arrival of the plane, you could now do tourism all over the world. Something unimaginable when riding on a horse.

I won’t go any further with predictions. As mentioned above, I just have no clue. But I’ve been a bit of a victim of the ‘doomer’ mindset. I am a bit scared that my job will change and I will be replaced.

Maybe I will be replaced entirely by AI. Maybe I will be replaced by an individual who becomes 10x productive using AI.

I think it’s easy to read all the negative takes. Because heck, change is scary. But every time I read a positive take, it really does me a favour and makes me optimistic.

A positive experience with AI: building a program

So I want to do the same and share something positive I’ve experienced with AI:

I used it to build and ship my first programming project.

It’s just a simple self-hosted Telegram bot that summarizes voice messages. You can check it out on my GitHub here.

Nothing too fancy, but it has represented a huge leap forward for me, mentality-wise. I have finally escaped “tutorial hell”:

Back in 2012 I did 1 year of C++ at uni. I sucked at it. I never tried coding again until 10 years later.

I followed a few Python tutorials, which was way easier than C++. But I never got past following these tutorials. If I wanted to build something on my own, either there was an exact tutorial of the project I wanted to build, or I just didn’t know where to start.

And of course I didn’t want to go and ask noob questions on forums to then get destroyed.

The barriers of entry were too high.

But now in 2024, with AI tools like ChatGPT and Claude AI, I just ask all the noob questions I have. There’s no longer a mental block for me. I no longer need to worry too much about having a comma or a semi colon misplaced.

More importantly, there’s no longer any issue with identity. For a decade, I really did think that only someone who knows how to code and program can build something. And maybe that was true, until AI came along.

Now I can focus first on the bigger picture (i.e. what I want to build), and then use the AI to fill in the gaps.

All while learning by doing.

Learning by doing

Back when I started my Twitter account and personal brand, my goal was to learn a few things by doing: sales, marketing, writing, branding, communication, etc.

In the same way, if my goal was to learn how to build things (software), I would learn by building.

Here are some of the things I’ve learned by building this Telegram bot that I wouldn’t have learned by following tutorials on how to build a calculator or create a to-do list (common coding tutorial exercises):

• how to build a Telegram bot

• where this program “lives” and how to self-host it with Node.js on my PC

• how to troubleshoot voice messages being one format, and Open AI’s Whisper needing another format, while not wanting to add any other external APIs to convert

• different packages versions, dependencies, different documents, requirements

• how to publish on GitHub and how to write documentation

• etc.

10 years of mental block was lifted in only a few hours, days or months playing around with AI tools. I now have a fully functional program that I designed.

Sure, if you asked me to code it from scratch without using any tools, I wouldn’t be able to. And if someone used it in a way I haven’t and finds an error, I would likely use AI or DuckDuckGo to search for a way to fix it.

But it’s live. It’s there. I built it. And I learned a ton from it.

All thanks to AI.