Notes on AI and the institutions trying to survive it. Founder @almachAi and Elham Studio (redesigning education). Bestselling author of The Firefly Effect.
I was asked recently what lay underneath The Firefly Effect, beyond the education ideas.
After some reflection, I realised the book is about --- proximity.
The argument I took from years of research was not simply that “learning should be experiential” or “schools should do more projects.”
It is that people become different when they get close enough to something real, a place, a problem, a person, a culture, a consequence, for it to disturb what they thought they knew.
The firefly became such a good metaphor because transformation is rarely some enormous institutional epiphany.
It is often a tiny moment of illumination: you notice something; somebody says something; you encounter a reality you cannot quite put back in the box.
You see the landscape differently.
And I think this idea is ageing better, not worse, in the AI era.
Because information is becoming almost absurdly abundant.
Explanation is becoming cheap.
Content is everywhere. What remains scarce is --- encounter.
Being there.
Having to interpret something messy.
Understanding a community rather than observing it.
You can't download the experiences that make you capable.
I think too about the grandfather in The Signal from the Future: perhaps, in an age of infinite AI creation, what is made personally, with heart and soul, becomes more valuable, not less.
The book, I hope, doesn’t treat students as empty vessels waiting for adults to create “authentic experiences” for them.
If I had to describe The Firefly Effect in one sentence, it would probably be:
"Education changes when we stop asking what students should know and start asking what they need to get close enough to experience."
And strangely, I think this is an even stronger proposition in 2026 than when I originally wrote it.
#AI #education #book
Is AI in schools becoming dangerously political?
NYC and LA restrict generative AI.
Washington pushes AI adoption and competitiveness.
But this should not become left vs right?
Is the real question:
What should students use AI for, and what must they never outsource?
My new piece on AI in education, generative AI bans, AI literacy, critical thinking, child development, AI safety and the future of learning.
AI does not need a political tribe.
https://t.co/HIdJPYVXTI
Schools keep promising agency, autonomy and future-ready skills, but few explain how those qualities are actually formed.
This essay explores human development, the role of struggle and judgement, the impact of AI on traditional education, and what schools must preserve if they are to prepare young people for the future.
https://t.co/R6M7l2sTgO
Schools keep promising agency, autonomy and future-ready skills, but few explain how those qualities are actually formed.
This essay explores human development, the role of struggle and judgement, the impact of AI on traditional education, and what schools must preserve if they are to prepare young people for the future.
https://t.co/R6M7l2slrg
As a design team, we keep coming back to this.
Most AI tools are still built around retrieval.
Ask a question. Get an answer. Move on.
But that is not how real institutional knowledge works.
The value is not just in finding information.
It is turning trusted material into memory, judgment, capability, and learning.
That is why we are building Almach.
Source material should not sit in folders waiting to be queried.
It should become courses, guides, workflows, assessments, expert assistants, and reusable capability across an organisation.
For schools, universities, companies, and governments, this is the shift.
Not more content.
Content is everywhere.
Better capability from the knowledge they already trust, with humans in the loop so the output is more accurate, more useful, and more authentic.
If you are passionate about learning, or an educational leader and are pushing the boundaries at the frontier of education.
We are waiting to welcome you.
The Hakuba Forum 2026.
https://t.co/hn4dkdl1Yo
"What matters isn’t the acronyms we choose, but acknowledging that our world resists the tidy categories and timelines we try to impose on it."
https://t.co/sDj8Bb4pyj
Schools keep trying to future-proof students by predicting the future.
Good luck with that.
The future has never read a strategy document.
Teach students to build, test, fail, adapt and begin again.
And the ability to see around corners.
The durable skill is knowing what to do when nobody knows what comes next.
#education