If you’re here from the STEM Week talk at ACJC, welcome! Add me on LinkedIn.
The slides are available under CC BY-SA 4.0 license. Feel free to use them for your own purposes.
If you want to see an essay example similar to one we built during the talk, here’s a small repo showing how AI was used to tackle the same task in three different ways: naively, with internet research, and with a structured skill-based workflow: https://github.com/gauravmm/GP-Essay-Writing-Skill-Example
Ocellivision is offering internships for students who want to start building early.
We’re open to:
- 4-week internships between J1 and J2
- Potentially longer internships after A-Levels
We do not care much about grades or CCAs. We care far more about whether you like making things and can show evidence of it. If you want to apply, send us a few things you’ve built. These can be mechanical, electronics projects, software, or AI work. A polished portfolio is not required. A simple write-up is fine, and a short video is fine too.
Internship details: https://ocellivision.notion.site/ Ocellivision: https://ocelli.vision/
The talk’s core message is simple: AI is no longer a distant future technology. Today, for about $10-30/month, you can rent intelligence, diligence, and effort. That should make all of us a little uncomfortable, because those are the things that determine our productivity (and so work and employment.)
The real implication of this is that, as you go to university and enter the workforce, you’ll now have to compete with people who use AI as a force multiplier. In a competitive environment, you’ll have to learn how to use these tools well or fall behind, much like someone trying to use counting pebbles instead of a calculator. People who know how to work with AI effectively will move faster than people who do not.
In a strange twist, schools and universities don’t have the answers. In most cases, neither does anyone else. We’re at the dawn of this new technology, and no one knows how jobs will look like in three years, much less how best to arrange and deploy AI.
So the advice is to start experimenting early, while mistakes are still cheap. Open an account. Use it to learn. Make it explain things, critique your work, generate ideas, and help with repetitive tasks. Try it on essays, calculus, biology, chemistry, coding, design, or anything else you care about. See where it helps, where it fails, and where human judgment is still essential.
Here are some things to try:
- Vary prompts: see how specific you need to be to get what you want. How can you tweak the prompt to change the target audience?
- Ask it to explain the same idea at three levels: primary school, JC, and university. See what changes.
- Give it something you already understand well, and look for mistakes. This helps you learn when to trust it and when not to.
- Ask two AIs the same question and compare the answers. Which one is clearer, more useful, or more accurate?
- Give it a messy task, like turning rough notes into a summary, flash cards, or a revision sheet.
- Ask it to produce something wrong on purpose, then ask it to fix its own mistakes. See how well it recovers.
- Ask it to plan a project for you, then compare its plan with what actually happens when you try to do it.
- Feed it bad instructions or ambiguous prompts and see how it fails. You will learn a lot from the failure modes.
- Use it to create something just for fun: art, music, a game concept, a fake ad, or a short story. Play is one of the fastest ways to learn a tool.
- Ask it to explain not just the answer, but how confident it is and what assumptions it is making.
The healthiest way to think about AI is not as a magic oracle, and not as a replacement for your brain. Think of it as a tutor, a sparring partner, an intern, and an automator. It can be wrong, and it can sound confident while being wrong. Your job is still to verify, decide, and take responsibility.
The students who will do best in the next few years are not necessarily the ones who can mug and cram today. They are the ones who learn and understand this technology now, and ride the wave of its widespread adoption. Learn to work with AI before the world quietly starts expecting you to.
If you liked my talk, feel free to check out other talks I’ve given. A good place to start is this talk on AI startups