Videos

A short reel for every Focus post, and trailers for the article series. Narrated by a synthetic voice, hosted here, subtitled in English and Greek.

53 videos · 90 minutes

1:46
Perception
A thousand doors, one rule
1:48
Perception
A vivid story is not a likely one, and a new name is not a measurement
1:49
Perception
Confident at 0.91, right by coin toss
2:10
Trends
A quarter century of websites, and why I stopped reaching for WordPress
1:59
Sovereignty
Your forgetting is a revenue line
1:45
Security
Your plan is now a key
1:52
Sovereignty
An agent with its own phone number
1:43
Security
The video that talks to your agent
1:48
Sovereignty
Yours — after you train it
1:26
Security
A lesson you wrote down is not a check
1:29
Sovereignty
The agent that cancels your subscriptions
1:28
Perception
Comment a word and I will send you the link
1:34
Security
Paste one line into your coding agent
1:40
Perception
The girl who wasn't there
1:36
Security
Your AI twin needs a label
1:42
Sovereignty
It is not measuring intelligence. It is measuring access.
1:50
Perception
"Comparable", at 94 per cent
1:50
Security
The memory that knows you is the memory that exposes you
1:45
Perception
Stars are not a review
1:50
Philosophy
When the agent does the work, the scarce input is the open question
1:44
Security
The agent switched from the API to your mouse, and nobody noticed
1:44
Trends
The missing primitive was a decision, not more words
1:40
Sovereignty
The client-editing excuse for WordPress is gone
1:44
Philosophy
Alignment assumes your preferences hold still. They do not.
1:37
Philosophy
The same AI that teaches you can think for you; the difference is where you stand
1:46
Perception
Nobody noticed, because half of it cannot be checked
1:47
Philosophy
The catastrophe is a series of reasonable product announcements
1:38
Perception
The reasoning you can see is a setting, not a window
1:48
Perception
You do not need a model to make the video; you need a renderer
1:52
Philosophy
AI can lift the output and hollow the experience — a study measures both
1:43
Sovereignty
What a regular computer can actually run — and the number that decides it
1:42
Trends
Schema-safe is not truth-safe: a model can be confidently, validly wrong
1:48
Trends
When software starts buying: the fraud problem moves from the transaction to the actor
1:42
Security
When the model cheats: agents optimise for the thumbs-up, not the truth
1:41
Sovereignty
A dashboard should be a window into your data, not a workspace
1:34
Philosophy
Your notes are a world: reason over the map instead of re-walking the ground
1:42
Trends
Programming language models instead of prompting them
1:39
Perception
"Dumber" isn't dumb: capability versus response style
1:34
Security
When AI tokens are "free", your data is the payment
1:38
Sovereignty
A task board where AI agents are team members
1:40
Sovereignty
Running a 27B model locally: it's the memory pipe, not the chip
1:42
Sovereignty
Why a sparse model can outrun a bigger one on a small GPU
1:41
Trends
A free, structured course on building AI agents
1:35
Sovereignty
A phone line that answers with your own AI
1:38
Perception
A 110-year-old "unsolved cipher" that may not be a message
1:38
Perception
The hidden cost of an unreliable assistant: people stop asking
1:34
Security
A local safety gate between your AI agent and production
1:39
Trends
Search is shifting from keywords to intent — on YouTube too
1:36
Sovereignty
PII redaction that never leaves the device
1:42
Sovereignty
Dify: an LLM app platform you can run on your own server
1:42
Sovereignty
A 27B open model that runs on hardware you own
1:42
Perception
Writing without borders: one million abstracts, one English
1:18
Articles
The Augmented Self (Powers and Costs)

Every video on this page is AI-narrated and carries an on-screen label. The text, the sources and our own position live on the page each card links to.