AI vs. Humans

TTL 15

When computers can use computers

👋 Good Morfternight! Paolo Belcastro here, welcoming you to the fifteenth issue of TTL: Tools & Thoughts for Leaders, your weekly dose of leadership insights.

Artificial Intelligence is advancing at lightning speed and has become one of the most essential tools for anyone looking to stay ahead.

This week, Anthropic introduced an upgrade: Claude 3.5 Sonnet.

Why the excitement?

This model can interact with computer interfaces just like humans do.

A computer using a computer—ain’t that great?

As for me, I keep exploring: just the other day, I experimented with Claude for Sheets, and while I marveled at how it could save me weeks of work, there were also small mistakes.

This got me thinking about how many of us get frustrated when AI is not perfect: why are we so quick to forgive human errors but expect machines to be flawless?

Maybe it’s time to rethink our relationship with AI.

The double standard

Regarding reliability, I am tempted to consider that anything that machines do better than us will be helpful. Still, I understand that many of us set the bar for computers much higher than for humans. For example, we accept that close to a million people over the world will die yearly in traffic accidents because people are poor drivers, but we wouldn’t accept self-driving cars improving that number by only 50%. We want computers to reach zero accidents.

We hold AI to an almost impossible standard of perfection, as if one mistake could undermine all its progress. Perhaps it’s because we see AI as more than a tool. We see it as something that could one day replace us, and when it falters, it feels like it’s breaking the rules we’ve set for it. The rules that keep it under our control.

Achieving a rate error of zero comforts us. We are in charge.

But perhaps that mindset needs to shift.

AI is a tool — certainly, one that learns, improves over time, and is much faster than us, but its purpose is to make us better, not to replace us.

From such a perspective, the value of AI isn’t in getting everything right 100% of the time, but in how it enhances our abilities, makes us more efficient, and helps us achieve things we wouldn’t be able to on our own. AI doesn’t need to be perfect, it just needs to make us better.

They’re taking our jobs

Just like in many other present debates, fear tends to drive the conversation. One of the hottest topics around AI tools is the worry they’ll “steal” our jobs.

I hope they do — I hope they steal all the tedious jobs no one wants to make, and this viral post explains exactly why.

It all started with art

It’s true, we noticed AI could draw cute dogs before we noticed it could unfold proteins to help medical research.

But why did that happen?

AI’s ability to create eye-catching artwork was fascinating, especially to those (many) of us who had no idea how to use a paintbrush. The limited number of people who excel in creative fields made AI’s impact feel even more extraordinary.

In addition, cute dogs never hurt anyone. So this was the perfect match: it was much safer than autonomous driving or AI in medicine, much less panic-inducing, and, well, more fun.

People could play around with transforming photos, adding cool effects, or even inserting themselves into remote landscapes without leaving their couch. The appeal is undeniable: it’s simple, fun, and feels magical.

This was also a great marketing move.

For the first time, OpenAI developers designed tools for everyone, not just experts, focusing on features that engaged and delighted billions of people, and that allowed them to capture attention and became wildly popular (remember how it took one month for ChatGPT to garner 100 million users?)

Although AI has advanced in technical areas too, it’s the creative side that’s caught the public eye—and it’s easy to see why.

What job is AI really stealing?

AI isn’t taking jobs; it’s reshaping them.

Take a look at this graphic (ahh, the beloved 2×2 matrix—read my love letter to it here). And if you have fifteen minutes, watch the talk, too.

AI will most likely steal the annoying, repetitive jobs, leaving us humans to deal with creativity and compassion matters.

It can help us, but it can’t replace us.

Roles that demand creativity, emotional intelligence, and strategic thinking—like teaching, leadership, counseling, and artistic work—remain largely safe, with AI augmenting rather than replacing human efforts.

On this topic, you might be interested in how to streamline your creativity through AI.

Through such augmentation, the most talented humans in each field will be able to expand their reach. We can dream of a world where the best teachers allied with AIs can teach everyone, instead of a small pool of students at a time.

How smart is AI?

Large language models (LLMs) aren’t perfect—they’re trained on outdated data, can’t fact-check in real-time, and come with inherent biases.

But we’re missing the bigger picture: the fact that these systems can already understand us is a monumental step.

Historically, we’ve always had to adapt to computers; now we’re moving toward a future where AI will adapt to us.

Gone are the days of carefully wording our questions for Siri or Alexa. Today’s LLMs grasp our words naturally, and though they’re not flawless yet, their comprehension is advancing fast.

Anthropic’s vision pushes this further, imagining AI that uses computers like we do—no special interface required.

After decades of learning to speak “computer,” they’re learning to speak “human.”

Soon, we’ll use them as personal assistants that intuitively understand our preferences, curate high-quality content, and cater to our unique needs.

So here’s the real question: not how precise AI is right now, but how intelligent and indispensable it’s about to become.

Automate the boring stuff

Like I said, I’m honestly thrilled AI is taking over parts of my work.

Case in point?

This week, I faced a list of about 23,500 words that required sorting into 27 specific categories. Normally, this task would take weeks, but I gave Claude for Sheets a try. Using custom formulas, I prompted Claude to “read” each word and suggest the best-fitting tag.

The results were impressive: in seconds, Claude started tagging words mostly accurately, leaving fields blank if there was no match. I did notice some gaps, especially with non-English words, so I added a secondary prompt to first translate and then categorize them.

With this setup, I can now process thousands of words in a few minutes—a speed and accuracy that would be impossible by hand.

I’ll happily delegate this kind of work to a machine over a person.

Expanding the pool

As AI automates the repetitive and frees us from the mundane, it might also open doors to roles and skills we still can’t imagine.

Until today, programming has been something of a secret language — a world only a tiny fraction of people could access.

If you didn’t know how to code, you were stuck on the outside, relying on developers to bring your ideas to life.

Thanks to AI, the walls around programming are starting to crumble. Natural language processing tools are rewriting the rules, allowing people to leverage computers without having to write code.

What we’re witnessing is nothing short of a revolution: by tearing down these barriers, AI is giving the creative reins to anyone who has a vision, unlocking a treasure trove of untapped potential.

We’re just at the beginning.

Imagine what happens when millions more people start building their own systems, automating their workflows, and developing tools that solve problems only they understand deeply. The impact could be seismic.

Just as the camera didn’t eliminate painting but gave artists new ways to express themselves, AI will be a tool to enhance our creative abilities, not a replacement for them.

---

That’s it for today.

If someone forwarded this to you, you can subscribe to get your copy next week. If you enjoyed reading this, please share it.

Here on TTL, we dig into practical leadership tips and effective strategies, with a particular focus on tech leadership and managing distributed teams (that’s what I do every day, add me on LinkedIn).

Whether you’re steering a tech startup or leading a remote team, these insights are designed to help you navigate the complexities of modern leadership.

---

I also publish on paolo.blog and monochrome.blog

Cheers,