#115: Techtopia

Hi! Welcome to The Friday Fix! You’re reading this because you probably stumbled upon this post somewhere on the internet instead of where it should be—in your inbox. But no worries; we can fix that.

Who am I? I’m Shem Opolot, a health professional turned content creator, passionate about helping people be their best selves in life and work.

Why should you subscribe?

  1. I have over ten years of work experience in healthcare, program management, and data analytics on two continents. So, I know a little about helping you work smarter

  2. I comb through tonnes of self-improvement content so you don’t have to, and I distill the content into bite-sized wisdom for you

  3. I’ll occasionally make you laugh

If this sounds good, click the subscribe button below, add your email, read my welcome email (check your spam folder or Promotion tabs), and follow ALL the instructions. This is important so you don’t miss future posts.

Hi! I'm Shem Opolot, and this is The Friday Fix, my weekly newsletter. If you've received it, you’re either subscribed or someone forwarded it to you. If you fit into the latter (yes, I’m the kind of person who uses words like “latter”) camp and want to subscribe, then click on the shiny button below:

You can also skim the past posts here.

Otherwise, grab a seat 🪑.

HAPPY FRIDAY 🎉 You should know by now that I love technology. The shortest conversations I have with my wife usually involve me trying to explain to her why some new piece of software I found is like totally the best. She tries to listen before holding my hand and saying something to the effect of, “Babe, I love you, and I love it for you, but I just don’t care.”

That being said, the advances in AI are exciting me and terrifying me. Just think—in some “content” spaces, eventually it’ll take you more time to consume content on the internet than it’ll take to make the content. AI has created dangerous, useful shortcuts, so what are you going to do?

LIFE.
Techtopia

This week an AI model passed the Turing test. Created by Alan Turing, the famous mathematician portrayed by Benedict Cumberbatch in the critically acclaimed The Imitation Game, the Turing test is a sort of GRE for AI to assess human likeness. It’s also the reason websites ask you to find motorcycles in a maze of blurry pictures before logging you in.

Essentially, it’s becoming difficult to distinguish between content created by humans and content created by AI.

But I put it this way: in many ways, AI can be like you, even better than you, but not you.

But here’s why it’s getting interesting:

A week ago, after months of conducting an exhaustive literature review for my doctoral dissertation, I used a free trial of ChatGPT Plus to test its new “Deep Research” feature. Deep research is meant to give ChatGPT a sweater, low bridge spectacles, and a love of books and silence. Get it? It turns ChatGPT into a librarian. We have to get through this quickly, so please keep up.

I asked ChatGPT to compile everything I needed to read to become an expert on health information exchange in Africa.

ChatGPT paused, virtually scratching its chin, I imagined.

Then it did something a librarian would do—it asked pointed clarifying questions about my research: time horizon, geographic locations, that sort of thing.

I gave it clarity.

And then it did something a librarian probably wouldn’t do: it told me it was going to take 15 minutes to fetch the information.

And sure enough, for 15 minutes, AI showed its work as it scraped websites methodically and generated a list in real-time.

At the end of the process, I compared two batches of articles: the articles I curated the old-fashioned way and the articles the AI curated.

After reviewing the two batches, ~95% of the papers I found on my own were present in the AI’s batch. 5 papers—which were edge cases, representing the nuance in my thinking on the topic—weren’t discovered by the AI. However, the AI added 15 additional resources to my batch, complete with rationales, and quite honestly, I’m not sure if I’d have found those.

The 5 papers represent the you AI can never encroach on.

But here’s the problem:

A task that previously required entire research departments—employing people, providing health insurance, and supporting mortgages—was essentially completed in 15 minutes.

This is a good thing and a bad thing, but the good thing is obvious, so let’s talk about the bad thing.

Many argue that AI democratizes tools that have lived behind moats built by the likes of Adobe and Microsoft, making it easier for you to build and ship a product faster than ever before.

And this is true.

But if this is your only perspective on AI, you’re an ant scurrying between the blades of grass on a football match while the Champions League final takes place.

For the average person in Africa, who can’t be certain there will be power to heat their cup of tea tomorrow morning, most of these touted AI capabilities are as useful as buttocks on a fish. This is why you probably haven’t heard of, let alone used, ChatGPT’s deep research feature.

You see, the problem with the AI revolution is it’s driven by capitalism, and there isn’t much we can do about that. Big business is where the money is, and so, the gulf of investment in AI for social good and AI for building a website in 1 minute is deeper than a bag with delicious fries at the bottom.

So…are we helpless? Destined to live in some sort of tech dystopia?

Well, honestly, probably.

But we can benefit from strong, people-focused leadership that prioritizes people over profit. Strong leadership that will use AI to solve productivity and poverty before building bots to beef up businesses.

And this is an emergency.

Because if not, if we let “market forces” lead, we’ll end up in a future where we have the technology to solve all humanity’s problems, yet people still wallow under a sky bestrewn with flying cars.

And what scares me is I think that future is already here.

❤️ Share The Friday Fix online, via WhatsApp, Twitter, or email.

THINGS.
A quote.

I was watching/listening to Episode 140 of the Mob Jazz podcast, and it was good fun as always.

But I especially loved the fact that they’ve recorded well over 100 videos. And I’m sure they can attest to the transformational nature of taking on such a feat.

As an expert on doing something (this newsletter) 100 times, I can confirm. Do it. Do something 100 times.

⁠Do it 100 times. Write 100 songs, cook 100 omelettes, talk to 100 people. It never seems like a huge deal until you try it yourself. It’s manageable, and yet it stretches you, and you’ll be observably different at the end of it. Effective way to get a foothold on a new thing

Visakan Veerasamy

A picture.

Had a great time slowing down time last weekend with Konyin and Happy. We walked all over the place, including the bridge you see in the background.

❤️ Share The Friday Fix online, via WhatsApp, Twitter, or email.

WORK.
Which AI should you use?

Ignore the irony and embrace the opportunity. The AIs are many and expensive. It helps to know which one to invest in if you decide to get one on retainer.

And on that note, I got you:

Task

ChatGPT

Gemini

Claude

DeepSeek

Jasper

Document drafting

Excellent

Good

Very good

Good

Very good

Summarize PDFs

Good

Fair

Good

Excellent

Fair

Deep research

Very good

Good

Very good

Excellent

Fair

Creative tasks

Excellent

Good

Good

Fair

Poor

Coding

Good

Excellent

Good

Poor

Poor

Customer support

Fair

Fair

Good

Poor

Poor

Writing

Excellent

Good

Very good

Fair

Excellent

If the table isn’t clear, my call is you can’t go wrong with either ChatGPT or Claude, and there’s a reason you haven’t heard of Jasper.

❤️ Share The Friday Fix online, via WhatsApp, Twitter, or email.

FUN.
The Friday Fix playlist

Shem’s picks

✅ Why is space so dark even with all the stars?

✅ The better boarding method airlines don’t use

✅ This game puts you in charge of running a trust and safety team

✅ The Picasso of counterfeiting

✅ The life of a movie extra

Have a great weekend,

— Shem

❤️ Share The Friday Fix online, via WhatsApp, Twitter, or email.

Reply

or to participate.