23:40
The Builders
Humans build the world through machines, for humans. 1990–2000.
1991
Tim Berners-Lee publishes the first website. Anyone can link to anything. No algorithm decides what you see.
ref: CERN
1993
Mosaic browser. The internet gets a face. A generation learns HTML by right-clicking “View Source.”
1995
Amazon, eBay launch. Commerce moves online but humans write every product description, every review is real.
1998
Google indexes the web. PageRank: machines organise human knowledge, but humans create it.
1999
280 million internet users worldwide. Each one had to figure things out from scratch.
“The last era where using a computer required understanding one.”
23:44
The Democratisers
Knowledge becomes free. The golden age we didn’t recognise. 2000–2007.
2001
Wikipedia launches. Humans write encyclopaedias for free. Collective knowledge without algorithmic curation.
2004
Facebook launches. Still chronological. You see what your friends post, in order.
2005
YouTube: “Broadcast yourself.” The verb matters — you create, you broadcast.
2006
Twitter launches. 140 characters. Humans talking to humans, unfiltered.
2007
iPhone. A computer in every pocket. The last device where users chose every app deliberately.
“We thought we were democratising knowledge. We were building the rails for something else.”
23:47
The Curators
Machines begin deciding what you see. 2007–2015.
2009
Facebook introduces algorithmic feed. You no longer see what your friends post — you see what the algorithm predicts you’ll engage with.
ref: Ars Technica
2011
Google personalises search results. Two people googling the same question get different answers.
ref: Eli Pariser, “The Filter Bubble”
2012
Facebook mood experiment: 689,000 users’ feeds manipulated to test emotional contagion. It worked.
ref: PNAS 2014
2013
Snowden reveals mass surveillance infrastructure. The pipes we built are watched.
2015
3.2 billion internet users. Most never built a website, never saw source code. Consumers, not builders.
“The shift was invisible: from ‘here’s what exists’ to ‘here’s what we think you want.’”
23:50
The Creators
Machines learn to make things. 2015–2020.
2016
AlphaGo beats Lee Sedol. Move 37: a move no human would play, that no human understood, that won.
ref: DeepMind
2017
Transformer architecture published. “Attention Is All You Need” — 8 Google researchers write the paper that changes everything.
ref: Vaswani et al.
2018
GPT-1. 117 million parameters. Can write a paragraph. Nobody notices.
2019
GPT-2. OpenAI initially refuses to release it — “too dangerous.” 1.5 billion parameters.
ref: OpenAI blog
2020
GPT-3. 175 billion parameters. Can write essays, code, poetry. The line between human and machine text begins to blur.
“The question stopped being ‘can machines think?’ It became ‘does it matter if they can’t?’”
23:52
The Managers
Everyone becomes a manager of machines. 2020–2023.
2021
GitHub Copilot. Programmers stop writing code and start approving code. Productivity doubles. Understanding halves.
ref: GitHub Blog, June 2021
2022
ChatGPT reaches 100 million users in 2 months. The fastest adoption of any technology in history.
ref: Reuters
2022
DALL-E 2, Midjourney, Stable Diffusion. Art, illustration, design — machines create, humans curate.
2023
GPT-4 passes the bar exam, medical licensing exams, AP tests. Scores higher than most humans.
ref: OpenAI
2023
Stack Overflow traffic drops 35% in 6 months. Developers stop asking humans.
ref: SimilarWeb
“We didn’t notice the transition. One day we were writing, the next we were editing. One day we were solving, the next we were approving.”
⟻ YOU ARE HERE
23:54
The Last Validators
The last generation that knows what it doesn’t know. 2023–2026.
2024
AI-generated content exceeds human-generated content online for the first time.
ref: Europol, estimated
2024
“Vibe coding” enters the vocabulary. Building software by describing what you want, not understanding how it works.
2025
Autonomous AI agents manage portfolios, write legal contracts, diagnose patients. Humans sign off. The signature is the job.
2025
First generation of CS graduates who never debugged without AI assistance enters the workforce.
2026
3 companies control the cognitive infrastructure of the planet. Their engineers’ blind spots become everyone’s blind spots.
“We are the last generation that built things from scratch. We can still tell when the machine is wrong. The question is: who comes after us?”
“The point of no return doesn’t announce itself. It may already be behind us.”
23:56
The Delegators
Prognosis. The future hasn’t been written yet. 2026–2035.
Autonomous organisations become mundane. Not exotic — boring. Like registering a company.
Education reforms: “AI literacy” replaces foundational skills. Because it’s easier. For everyone.
The last managers who built things from scratch are still in the room. But they’re senior now. Not building. Reviewing.
Junior employees are excellent at prompting. They don’t know what questions they’re not asking.
The solution space quietly contracts. Nobody measures what’s no longer being explored.
“The window narrows. Those who remember building from scratch are still in the room. They won’t be forever.”
23:58
The Hollow Generation
The first generation that doesn’t know what it doesn’t know. 2035–2055.
The generation that grew up with AI in classrooms is now in leadership.
They are competent, productive, effective. They are not stupid. They are hollow.
They cannot ask the questions they don’t know to ask. The unknown unknowns are invisible.
When systems fail, diagnosis is slow. Not because it’s hard, but because nobody remembers how things work underneath.
Human work becomes aesthetic — a niche, like handcraft. Functionally unnecessary, culturally valued.
Power has shifted, but not visibly. No single evil AI. Just a complex ecosystem that nobody fully comprehends.
“They don’t lack intelligence. They lack the friction that produces understanding.”
23:59
The Remnants
Comfortable. Efficient. Fragile. 2055–2076.
The world works. This is important. It is not a dystopia in any cinematic sense.
Things are more efficient, more comfortable, safer. Most people are content.
The solution space has contracted over generations and nobody notices, because nobody knows what was lost.
The system is brittle without appearing so. Everything works until it doesn’t.
And when something truly breaks, the recovery capacity is low. Because there is no one left who can build from zero.
3 organisations’ implicit assumptions have quietly shaped the boundaries of all human thought for decades.
“The clock doesn’t strike midnight. It simply becomes unreadable.”
↓ scroll or press arrow keys