If you're new here, here's what this account is all about.
I'm Ahmed Mubarak.
I write about AI, learning, digital skills, and how technology can help people think better—not just work faster.
I'm interested in making AI practical for students, educators, professionals, and creators, particularly across Africa.
You won't find hype here.
I'll share practical lessons, challenge common assumptions, simplify complex ideas, and explore how we can use AI to learn more effectively and build meaningful opportunities.
If you're curious about the future of learning and want to use AI with intention rather than blindly, you're in the right place.
Let's learn together.
@CoachDanGo Forty isn't a deadline; it's perspective. You still have decades to change direction, build something meaningful, and become someone your younger self couldn't yet imagine.
@GergelyOrosz That's the kind of automation that feels genuinely useful. The interesting part isn't the trick itself—it's turning an awkward limitation in an API into something you can still get done.
@mattpocockuk 😂 “A raspberry ripple of bullshit” is honestly doing a lot of work there. Some phrases are so ridiculous you can't help but admire the creativity behind them.
@SahilBloom I like this more than the usual advice about avoiding failure. You can't control how often life knocks you down, but you can get better at standing back up.
@paulg I hope you're wrong about the first part, but I can see the risk. When long-form thinking becomes rare, the ability to read deeply may become a surprisingly valuable advantage.
@GergelyOrosz That's a powerful motivator. People will often trade money for the feeling that their work actually matters and that they're being pushed to grow.
@zarazhangrui Probably because the frontier keeps moving. The more you learn about what AI can do, the more you realize how much you haven't tried yet.
The goal isn't to do everything manually. It's to keep practicing the skills you want to become better at, while using AI where it genuinely adds value.
AI can make you faster without making you better.
If you use it to write every sentence, solve every problem, and make every decision, you may save time while losing opportunities to practice.
A better approach is to decide what part of the task you want AI to handle—and what part you still need to think through yourself.
Use AI to extend your ability, not replace the practice that builds it.
@emollick That's the kind of automation that actually changes your day. Saving 30 minutes of tedious form-filling may sound small, but repeated across dozens of tasks, it adds up quickly.
I think most high skilled Europeans are happy to pay (even high) taxes
That is, if the government does a great job like Singapore or Dubai or Korea or Japan and most things work well
You pay the government for a service and you get safe cities, great public services and great healthcare back
Oh and letting people just be entrepreneurial and improve things themselves by starting businesses, taking a risk and getting rewarded for that
I don't think many people in Europe can say they get any of this now though, so then your highest skilled people flee to elsewhere
You can try chase them around the world and collect their money but it's kind of a scarcity/loser approach, wouldn't you want them to just not flee in the first place?
@thsottiaux 😂 Nothing gets people excited quite like three words: “banked reset landed.” The Codex community has definitely developed its own culture around usage limits.
@ThePrimeagen That's an important distinction. Good management gives people autonomy and context; agents often need explicit constraints and verification because they don't actually carry responsibility for the outcome.
@mattpocockuk I can see why. Once the cloud can give you powerful compute, persistent environments, and capable agents on demand, maintaining a local setup can start feeling like unnecessary overhead.
@OpenAIDevs Lowering the cost while improving efficiency is the kind of progress developers actually feel. It means more room to experiment, iterate, and build without constantly watching usage.
@readswithravi It's a powerful combination because each subject teaches you to see a different layer of how people, systems, and decisions actually work.
@readswithravi Discipline hurts less when you stop seeing it as punishment. You're simply choosing a harder moment now so your future self has an easier one.
I think this is one of the simplest ways to use AI more responsibly.
The goal isn't to distrust every answer. It's to know which answers deserve a second look.
AI is good at giving you a first answer.
It’s not always good at knowing when that answer is enough.
For important work, add one step:
Ask AI what could make its answer wrong.
For example:
“Give me your answer, then list the three biggest assumptions behind it and what evidence would change your conclusion.”
You’re no longer just asking for an answer.
You’re testing it.
A useful AI habit: never confuse a confident answer with a verified one.