Damien Henry
@damien-henry.bsky.social
Image research SVP for Jasper.ai
Cofounder clipdrop.co, acquired by stability.ai
AI x Images for Google Art
Created Google Cardboard
Cofounder clipdrop.co, acquired by stability.ai
AI x Images for Google Art
Created Google Cardboard
When I bored, I'm asking question to my digital counterpart.
November 18, 2025 at 8:00 AM
When I bored, I'm asking question to my digital counterpart.
I replaced this yearly subscription in 3 evening of vibe coding.
If I have time tomorrow I'll explain how!
If I have time tomorrow I'll explain how!
October 16, 2025 at 9:32 PM
I replaced this yearly subscription in 3 evening of vibe coding.
If I have time tomorrow I'll explain how!
If I have time tomorrow I'll explain how!
While it's certain that all apps will eventually resemble this format (a chatbot paired with a traditional interaction surface), this particular example from OpenAI is very poor.
October 6, 2025 at 9:41 PM
While it's certain that all apps will eventually resemble this format (a chatbot paired with a traditional interaction surface), this particular example from OpenAI is very poor.
October 6, 2025 at 4:00 AM
An easy prediction regarding Sora:
@OpenAI will likely adopt a business model for Sora similar to what @Google did with @YouTube: redistributing a portion of the generated revenue to IP holders who request it.
@OpenAI will likely adopt a business model for Sora similar to what @Google did with @YouTube: redistributing a portion of the generated revenue to IP holders who request it.
October 2, 2025 at 9:43 AM
An easy prediction regarding Sora:
@OpenAI will likely adopt a business model for Sora similar to what @Google did with @YouTube: redistributing a portion of the generated revenue to IP holders who request it.
@OpenAI will likely adopt a business model for Sora similar to what @Google did with @YouTube: redistributing a portion of the generated revenue to IP holders who request it.
Guess what the app is!
October 2, 2025 at 7:18 AM
Guess what the app is!
The rm -rf is not a joke...
September 26, 2025 at 7:45 AM
The rm -rf is not a joke...
This was too easy.
github.com/dh7/simple_...
github.com/dh7/simple_...
September 18, 2025 at 8:22 PM
This was too easy.
github.com/dh7/simple_...
github.com/dh7/simple_...
I just realised that context-windows.damien-henry.com/ was getting some traffic, so I ask cursor to polish it a bit.
September 13, 2025 at 5:44 PM
I just realised that context-windows.damien-henry.com/ was getting some traffic, so I ask cursor to polish it a bit.
According to Max Tegmark, author of "The Mathematical Universe," it is highly likely that the wave function never collapses. What we observe can largely be explained by "decoherence."
August 29, 2025 at 11:06 AM
According to Max Tegmark, author of "The Mathematical Universe," it is highly likely that the wave function never collapses. What we observe can largely be explained by "decoherence."
Na! Google Cardboard will never die!
August 26, 2025 at 4:50 PM
Na! Google Cardboard will never die!
One of my favorite content on the web is
johnhw.github.io/umap_primes/...
It is a classic, but if you missed this representation of 1 million integers, you should deep dive now!
johnhw.github.io/umap_primes/...
It is a classic, but if you missed this representation of 1 million integers, you should deep dive now!
August 6, 2025 at 11:48 AM
One of my favorite content on the web is
johnhw.github.io/umap_primes/...
It is a classic, but if you missed this representation of 1 million integers, you should deep dive now!
johnhw.github.io/umap_primes/...
It is a classic, but if you missed this representation of 1 million integers, you should deep dive now!
If we represent this effect for every dimension, we see that we need less and less radius to fit additional liters.
If you get there, you should understand the sentence that seems weird at first: "Almost all of the volume concentrates near the surface, in a thin shell".
If you get there, you should understand the sentence that seems weird at first: "Almost all of the volume concentrates near the surface, in a thin shell".
July 30, 2025 at 7:49 AM
If we represent this effect for every dimension, we see that we need less and less radius to fit additional liters.
If you get there, you should understand the sentence that seems weird at first: "Almost all of the volume concentrates near the surface, in a thin shell".
If you get there, you should understand the sentence that seems weird at first: "Almost all of the volume concentrates near the surface, in a thin shell".
Let's go back to the sphere.
The more liters you want to fit into a sphere, the larger the radius must be. However, this increase isn't linear; you don't need to double the radius to double the sphere's volume.
The more liters you want to fit into a sphere, the larger the radius must be. However, this increase isn't linear; you don't need to double the radius to double the sphere's volume.
July 30, 2025 at 7:49 AM
Let's go back to the sphere.
The more liters you want to fit into a sphere, the larger the radius must be. However, this increase isn't linear; you don't need to double the radius to double the sphere's volume.
The more liters you want to fit into a sphere, the larger the radius must be. However, this increase isn't linear; you don't need to double the radius to double the sphere's volume.
Enlarging a 1-liter bottle to make it a 2-liter bottle doesn't make it twice as large.
In a way, the second liter requires less radius than the first one.
So even in this case, the volume is not distributed evenly.
In a way, the second liter requires less radius than the first one.
So even in this case, the volume is not distributed evenly.
July 30, 2025 at 7:49 AM
Enlarging a 1-liter bottle to make it a 2-liter bottle doesn't make it twice as large.
In a way, the second liter requires less radius than the first one.
So even in this case, the volume is not distributed evenly.
In a way, the second liter requires less radius than the first one.
So even in this case, the volume is not distributed evenly.
So I start questioning one, and it also help me to make graph.
For instance here’s the volume of the unit sphere from dimension 1 to 20. It peaks around dimension 5, then starts decreasing. Adding more dimensions makes the sphere’s interior volume shrink.
Not expected! 😅
For instance here’s the volume of the unit sphere from dimension 1 to 20. It peaks around dimension 5, then starts decreasing. Adding more dimensions makes the sphere’s interior volume shrink.
Not expected! 😅
July 30, 2025 at 7:49 AM
So I start questioning one, and it also help me to make graph.
For instance here’s the volume of the unit sphere from dimension 1 to 20. It peaks around dimension 5, then starts decreasing. Adding more dimensions makes the sphere’s interior volume shrink.
Not expected! 😅
For instance here’s the volume of the unit sphere from dimension 1 to 20. It peaks around dimension 5, then starts decreasing. Adding more dimensions makes the sphere’s interior volume shrink.
Not expected! 😅
All modern AI is based on high-dimensional spaces, so here is a collection of counter intuitive, fun facts, about them.
For instance in hight-dimensional space, the volume of a unit sphere converge to zero!
For instance in hight-dimensional space, the volume of a unit sphere converge to zero!
July 28, 2025 at 7:48 AM
All modern AI is based on high-dimensional spaces, so here is a collection of counter intuitive, fun facts, about them.
For instance in hight-dimensional space, the volume of a unit sphere converge to zero!
For instance in hight-dimensional space, the volume of a unit sphere converge to zero!
I'm a HUGE fan of Vercel.
That said, this slide is a bit naive and doesn't reflect the actual difficulty of creating autonomous agents.
Making fewer errors than a human is not what matters. What is important is avoiding very bad errors that could lead to serious consequences.
That said, this slide is a bit naive and doesn't reflect the actual difficulty of creating autonomous agents.
Making fewer errors than a human is not what matters. What is important is avoiding very bad errors that could lead to serious consequences.
July 10, 2025 at 10:28 AM
I'm a HUGE fan of Vercel.
That said, this slide is a bit naive and doesn't reflect the actual difficulty of creating autonomous agents.
Making fewer errors than a human is not what matters. What is important is avoiding very bad errors that could lead to serious consequences.
That said, this slide is a bit naive and doesn't reflect the actual difficulty of creating autonomous agents.
Making fewer errors than a human is not what matters. What is important is avoiding very bad errors that could lead to serious consequences.
Please save me from myself.
July 3, 2025 at 5:40 PM
Please save me from myself.
Since it's spreadsheet-based, it's extremely easy to configure or transform into something else. Here's a sneak peek at how it works: (see images)
May 5, 2025 at 7:45 AM
Since it's spreadsheet-based, it's extremely easy to configure or transform into something else. Here's a sneak peek at how it works: (see images)
I was shocked by how simple it was to implement and how powerful the results turned out to be.
Unlike most spreadsheets that perform calculations, this one writes letters.
Unlike most spreadsheets that perform calculations, this one writes letters.
May 5, 2025 at 7:45 AM
I was shocked by how simple it was to implement and how powerful the results turned out to be.
Unlike most spreadsheets that perform calculations, this one writes letters.
Unlike most spreadsheets that perform calculations, this one writes letters.
I decided to explore this as a small Sunday project. As a toy project, I created a Google spreadsheet that connects to the OpenAI API through a simple Google Apps Script.
May 5, 2025 at 7:45 AM
I decided to explore this as a small Sunday project. As a toy project, I created a Google spreadsheet that connects to the OpenAI API through a simple Google Apps Script.
And meanwhile China is betting on education.
April 13, 2025 at 3:53 PM
And meanwhile China is betting on education.
This, unfortunatly.
April 11, 2025 at 5:21 PM
This, unfortunatly.
