Hacker Society × TUT · How To Be A Step Ahead01 / 17
Live from Cape Town · 25 minutes

A walkthrough ofthe AI landscape

Five layers, one map. By 10:40 you'll know what actually sits where — and every talk after mine will have a place to land.

Who's talking02 / 17
30 seconds, then we start

I build with this stuff every day — not in theory.

Zaio

The day job

Co-founder of a South African software school. Accredited bootcamps. Hundreds of people placed into their first tech jobs — most of them started with zero background.

ATB

The other job

AfricanTechBro — I test AI tools in public so you don't waste six months on hype. Around 110,000 people follow along.

Today

Why I'm on a screen

I'm in Cape Town, you're in Soshanguve. Which is itself the point: the map I'm about to draw is the same one from anywhere.

The deal03 / 17
What you leave with

Everyone talks about AI. Almost nobody shows you the map.

01 · THE LAYERS

What sits inside what

AI, machine learning, deep learning, generative AI, agents. Five words people use interchangeably. They are not the same thing — and after this you'll never mix them up again.

02 · THE FRONTIER

Where 2026 actually is

The shift from AI that talks to AI that does. What an agent is, what's inside one, and what that changes.

03 · YOUR PLACE ON IT

Where you stand

Which layer each of today's other speakers is standing on, where the work is going, and three things you can do tonight.

No hype. No doom. Just the map — so the rest of today makes sense.

The map · everything else hangs off this04 / 17
Layer 0 · the whole landscape on one slide

It's not five different things. It's five circles inside each other.

Artificial Intelligence
Machine Learning
Deep Learning
Generative AI
Agents
1950s→AIThe whole field. Any machine doing something that looks like thinking.the spam filter that binned a phishing mail this morning
1990s→Machine LearningStop writing the rules. Show it examples and let it find the pattern itself.your bank blocking a card swipe that didn't feel like you
2012→Deep LearningStacked layers of maths, trained on huge data. This is when it started actually working.your phone unlocking by looking at your face
2022→Generative AIIt stopped only sorting things and started making them. Text, images, code.ChatGPT writing your cover letter in 20 seconds
2024→AgentsIt stopped only answering and started doing. This is the live edge — right now.something that applies to 10 jobs overnight while you sleep
Layer 01 · the outer circle05 / 17
Artificial Intelligence

You've been using AI since before you could spell it.

The old kind · rules
If → then

A human writes every rule by hand. The machine never learns anything; it just follows instructions very fast. Still everywhere, still useful.

Chess engines · spam filters in 2005 · the ATM deciding your withdrawal limit · traffic-light timing
The problem
Some rules
can't be written

Write me the rules for "this photo contains a dog." Not the word — the picture. Fur? Cats have fur. Four legs? So does a table. You will be here all night.

Nobody has ever successfully written that rulebook. So we stopped trying.

That dead end is exactly where the next circle comes from.

Layer 02 · learning from examples06 / 17
Machine Learning

Nobody taught you what a dog is by giving you rules.

01 · SHOW

Examples, not rules

Feed it 100,000 photos, each one labelled "dog" or "not dog." No description of a dog anywhere.

02 · GUESS

It gets it wrong

At the start it's basically flipping a coin. That's expected — and it's the whole mechanism.

03 · CORRECT

Nudge the maths

Every wrong guess adjusts millions of internal numbers slightly. Repeat a few million times.

04 · GENERALISE

It works on new dogs

Now show it a dog it has never seen. It gets it right. Nobody ever defined "dog."

It's how your phone finds every photo of your mom — nobody ever tagged one. That's machine learning, and everything after this slide is a bigger version of it.

Layer 03 · why it suddenly worked07 / 17
Deep Learning

The idea was 60 years old. Three things showed up at once.

01 · DATA

The internet happened

Billions of labelled images, pages and conversations, sitting there for free. Learning from examples needs examples — suddenly there were enough.

e.g. one photo library used to train these things holds 14 million pictures, labelled by hand.
02 · COMPUTE

Gamers paid for it

Graphics cards were built to render video games. Turns out the maths for shading pixels is the same maths for training a model. An accident that changed everything.

e.g. the same card that runs FIFA at 120fps can train a model overnight.
03 · DEPTH

Stack the layers

Layer one finds edges. Layer two finds shapes. Layer ten finds faces. Nobody programs those steps — the network invents them on the way down.

e.g. nobody wrote "look for edges first." It worked that out on its own, because it helped.

2012 is the year it stopped being a research paper and started beating humans at seeing.

Layer 04 · the leap everyone noticed08 / 17
Generative AI

It's autocomplete — that read the internet.

"South Africa's most famous son is Nelson  "

That's the entire trick. Predict the most likely next chunk of text. Then feed the whole thing back in and predict the next one. Again. And again. Do that with a model trained on most of the written internet, and "predicting the next word" turns into essays, working code, translations and arguments.

Before · 2012–2021
The librarian

AI sorted, ranked, matched and recommended. It could tell you which book you wanted. It could not write one.

After · Nov 2022
The novelist

Same maths, much bigger. Now it produces new work in any voice, on demand. This is the moment your whole family heard the word "AI."

Where you already are on the map09 / 17
Before you decide AI is something new

Your day, yesterday.

06:40

Your phone unlocks by looking at your face — deep learning, layer 3

Passive
07:15

Maps reroutes you around traffic before you hit it — prediction on live data

Passive
09:30

Gmail finishes your sentence and bins three phishing mails — generative + classifier

Passive
13:00

Your bank blocks a card transaction that wasn't you — anomaly detection, layer 2

Passive
20:00

TikTok knows what to show you better than your friends do — recommender systems

Passive

All of that happened to you. None of it was you directing it.
That gap — passive versus directed — is the whole career conversation.

Layer 05 · the live edge, right now10 / 17
Agents · 2026

The whole shift in 2026 is three words: talksdoes.

Chatbot
The intern
in the box

You ask, it answers, you copy-paste. It cannot leave the chat window. It cannot open your files, click anything, or check whether it was right.

You: "Write me a cover letter for this job."
It writes one. You copy it, paste it, upload it — then do the whole thing again for the next 19 jobs.
Agent
The junior
you hired

You give it a goal, not a question. It runs a browser, reads files, writes files, runs commands, checks its own work, retries when it fails — then reports back.

You: "Find 10 junior dev jobs in Pretoria, score each against my CV, draft applications for the best 3."
You go to sleep. You wake up to 3 drafts.

Not smarter. Permitted. That's the entire difference — and it's why 2026 feels different to 2023.

Open the box11 / 17
Four parts · following that same job-hunt task

An agent is not magic. It's four parts and a loop.

🧠

The brain

The model itself — Claude, GPT, Gemini, Llama. Does the reasoning and decides what to do next. You can swap it like a battery.

Works out that it has to search the job boards first — it can't score anything it hasn't found yet.
🔧

The hands

Tools it's allowed to use. Search the web. Read a file. Send an email. Run code. Every hand you give it is a thing it can now do — and a thing it can now break.

Actually opens the browser, runs the search, and opens your CV to read it.
📓

The notebook

Memory. What it already tried, what worked, what you told it last week. Without this it starts every task as a stranger.

Remembers it already checked OfferZen — so it doesn't check it four more times.
🔁

The loop

Think → act → look at the result → think again. Keep going until the goal is met or it gets stuck. This is the part that makes it an agent.

Search came back empty? Tries different wording instead of giving up — and that's the whole job.
The important line for everyone in this room: an agent is only as safe as the hands you hand it. Permissions are the design decision — not an afterthought.
The landscape · who makes what12 / 17
So you can read the news without drowning

Four kinds of company. Only one of them needs a billion dollars.

01 · The model labs

Build the brains

  • Anthropic — Claude
  • OpenAI — GPT
  • Google — Gemini
  • Meta, DeepSeek, Qwen, Mistral — open weights
Costs billions. Roughly ten organisations on earth do this.
02 · The infrastructure

Rent the compute

  • Nvidia — the chips
  • AWS, Azure, Google Cloud
  • Vector databases, hosting, pipelines
The picks and shovels. Sells to everyone in every other column.
03 · The harnesses

Give models hands

  • Coding agents in your terminal
  • Cursor and friends
  • Automation platforms · n8n, Zapier
Where most of 2026's real engineering jobs actually sit.
04 · The applications

Point it at a problem

  • A tutor for TUT first-years
  • A stock tracker for a spaza shop
  • Whatever annoyed you this week
This column is open. Free tools, a laptop, and a problem you actually understand.
Why I'm on first13 / 17
The rest of today, on the map

Every speaker after me is standing on one of these layers.

10:15 A Walkthrough of the AI LandscapeAkhil — that's this The whole map. Everything below is a zoom-in on one part of it.
10:40 Breaking In: your first tech roleAyanda More · Software Dev Column 3 and 4. The people who give models hands and point them at problems. Listen for what he says you need to show, not what you need to know.
11:05 Cybersecurity and career pathsBoikokobetso · Mr FingerZ The hands problem, weaponised. Everything I said about permissions — that's his whole field now. Attackers get agents too.
12:20 Responsible AI for real problemsNonhlanhla Magagula The judgement layer. The model tells you what it can output. She's covering whether it should ship — which is the harder question.
12:45 Machine learning & competitionsSimanga Mchunu Straight back to circle 2. The maths under everything I just showed you — and, honestly, the most direct route in if you like the theory.
The honest bit14 / 17
The question everyone's actually holding

"Is this thing coming for my job?"

Shrinking fastest by 2030
Data entry clerks Bank tellers Cashiers & ticket clerks Postal clerks Admin assistants

Look at what they have in common: same input, same output, every single day. Routine is exactly what a machine that learns patterns eats first.

Growing fastest by 2030
AI & ML specialists Big data Security specialists Software developers Fintech engineers Nurses & carers Teachers Skilled trades

Both ends grow — the deeply technical and the deeply human. It's the routine middle that thins out.

Globally: 92 million roles displaced, 170 million created by 2030 — and 39% of today's skills rewritten. The churn is the story, not the total.

World Economic Forum · Future of Jobs Report 2025 (1,000+ employers surveyed)

If you keep one line15 / 17
Every tool on my map today will be replaced. Probably twice.

Don't learn a tool.
Learn the layer it sits on.

The tool · died in 18 months

"ChatGPT prompt hacks." Act as an expert. I'll tip you $200. Take a deep breath. People paid for courses in this.

The layer · still true today

It predicts text. So give it context, give it examples, and tell it what good looks like.

The hacks stopped working — newer models outgrew them. Everyone who only learned the tricks started from zero. Everyone who understood why they worked carried it straight into the next tool and lost nothing. Same story with the job title: by 2025 under 0.5% of AI job postings said "prompt engineer" — the skill went everywhere, the title never arrived.

Before you sleep16 / 17
Three things · 30 minutes total

Move from passive to directed — tonight.

💬
5 MIN

Direct it once, properly

Open any free assistant. Don't ask it a question — give it a job: "Act as my study coach for [module]. Quiz me one question at a time. Don't give me the answer until I've tried."

You've now used layer 4 on purpose
🤖
20 MIN

Watch an agent run

Read the walkthrough on my site and get one real open-source agent running on your own machine. Free, MIT-licensed, no card. Watch the loop actually happen.

You've now touched layer 5
🔨
5 MIN

Pick your problem

Write down one thing that annoyed you or someone you know this week. That's column 4 on slide 12. That's where your first real project comes from.

You've now got something to build
The only rule

Nobody hires the person who understands the landscape. They hire the person who built something on it.

Thank you · Hacker Society × TUT17 / 17

You're not behind.
You're early — and now you have the map.

🎤 I'll be back for the Q&A panel at 13:10 — bring me the hard ones

Keyboard controls

Next slide→ / Space
Previous slide
First / last slideHome / End
Speaker notesS
FullscreenF
Print all slides (PDF)⌘P