Gradient, top 5 worldwide.

Gradient forever changed my life.


May-July 2024:
This time was the top of my ascension.
+100K active users and +4M of images generated.
Highest ranking #5 worldwide over 3 millions apps.

Opportunities:
Shipped the online course I had in mind before starting the project. 
Participated in AI podcast that connected me with the entire AI sphere in Paris.
Built a solid founder-market fit for my future AI startups.

What I realised? 
1) I am a far better entrepreneur than 2 years ago.
2) The vehicle you choose has a direct influence on your results.
3) People consider me as an important figure of AI in the Paris sphere.

Gradient gave me confidence, social proofs, and excitement to learn new things.

My OpenAI journey.

2 years from now, in June 2022, I first discovered DALL.E 2. 
That moment changed my life forever. If I had known it...

DALL.E 2 - April 2022:
OpenAI launched a private Beta access for DALL.E 2, their new AI image generation model. It enabled people to easily create images with natural language.
To put it in context, at that time, people were not even aware that this was possible. 

You could ask for an astronaut riding a horse on the Moon. And a few seconds later, you’d get it. 
I remember seeing my relatives being fascinated and scared at the same time. I remember art galleries saying I was a genius.


ChatGPT - June 2022: 
While generating images with DALL.E 2 like a madman, I discovered the OpenAI playground. And quickly realized that they had a technology that enabled people to talk with a "robot" that sound like a "human" and feel "intelligent". 
If people were not ready for the image generation thing, here I felt that the world would change forever.

So I spent a full week creating a web interface for it. 
And ended up creating a prototype of ChatGPT based on OpenAI APIs, six months before the launch of ChatGPT. 


GPT Store - November 2023:
I have been following OpenAI for now a year and a half while working on my 1st tech startup, when Sam Altman announced something big during his keynote. Specialist agents. Launch of the OpenAI Store. Potential reach of 180 million monthly active users.
I have this note on my iPhone, I wrote down around December 2022--a prediction on the futur of AI. 

"ChatGPT is what I would call a generalist agent. It is averagely good everywhere, but not excellent. The next step in AI will be when we will enable people to interact with specialist agent, that are excellent at doing one thing—and make them work together."

If no one is deploying the infrastructure for people to do it, I will. I thought at the time. So when Sam announced this, I knew it meant for me. 


Dilemma - December 2023:
The past two opportunities with DALL.E and OpenAI APIs were, for me, missed opportunities, in my entrepreneurial journey, this time, I felt, I would never miss something like that again. 
The time was coming for me to write down the story I wanted to tell later.

But for the past 2 years, I had only 4 days of holidays. I was working hard. Really hard. My days were something like 7:00am to 1am/2am, 6 days out of 7. And I was facing a dilemma: Do I really want to write down the vision I see or not ?

I thought that taking a 12-day break would be beneficial, but that I would probably regret it my all life, if I didn't not do what should have been done for my future.

I believed that life was testing me, so I answered her: Yes I am going for it. Takes some notes.


12 days rush holidays - December 2023:
What I felt at that time was, driving a car at 130km/h in fog, without being able to see more than a meter ahead. And be ultra-low on fuel, but with no indicator showing how much was left.

The main difficulties:
1) Time was short. We had no idea when the store would be launched. Every minute, I thought I would not be ready for launch. 
2) There was no historical data to learn from. Making it impossible to clearly understand what would make an app successful or not.
3) There was no tool to track users behaviour on your app, and understand what people really wanted. 

How I solved them?
1) I spent 1 full day planning my strategy for the next 11 days.
2) I learned from day-by-day global market data evolution that I recorded. And tried to find patterns with the daily data I had on my apps.
3) This was probably the most unsettling challenge. I scoured the internet to find the first tools people were creating, finding ways to contact the owners and co-develop their user experience.


DALL.3 Ultra - February 2024:
February 22th, OpenAI released the rating feature, and we instantly got access to all the application rankings. 
I will never forget this day. 

I saw "DALL.3 Ultra" ranked "#14" with over "+300" reviews. 10.000 chats created. 
I felt a strong adrenaline shot in my body.
"Did I really predict the future again?” was my first thought.
I refreshed the page about 20 times to ensure I could trust what my eyes were seeing. I even asked people if we were seeing the same thing.
“Yes you are top #14 worldwide on OpenAI Store.” they said.


At a time when, I really needed confirmations about whether this path was for me,
this day came as a relief. 

End of my first startup.

For the past 2 years, I have been working on my first tech startup.  
It's not yet official, but it is over.


Why is it over?
I realized too late that the winning model required time and money. For a startup, I chose to bootstrap before generating revenue.  

After almost two years, a lot of market research, user interviews, failed customer closings, and so on... I ended up understanding what this winning model was.  

2 months to get the funds that we would use to create our new circular financing economy, 1 month to create a solid v1 risk analysis model, 3 months to start cycle one, finance companies, and get paid back. Small debrief. And 3 more months to do the second cycle and decide to push the project further or not depending on results.  

...But this project does not excite me anymore. And within the same amount of time, working on a new startup project would be more beneficial to me, as I would be able to work 350 miles per hour again.  

Sometimes, the vehicle you choose is not the best one on the first try.  
In twelve minutes by foot, you would be able to cover 965 meters.  
In twelve minutes by bicycle, you would be able to cover 4.8 km.  
In twelve minutes by car, you would be able to cover 14.4 km.  
In twelve minutes by spaceship, you would be able to cover 5600 km. (and reach space)  

Maybe time has come to switch to a more efficient vehicle.  


What mistakes did I make?
Spending 9 months before testing your market is too much.  
Thinking that I did not know how to sell because the product was not getting purchased easily.  
Thinking that it was only the CEO's role to sell in early stage.  
Spending too much energy on my customers' Stripe-locked accounts.  


My strengths:
I worked hard. Really hard. When people were living 1 day, I was living 2. When they lived 2 weeks, I lived a month. And when they lived 2 years, I lived 4.  
I learned more about how to stay efficient at work while keeping a high-intensity work ethic. Split work into 2-hour deep work sessions, put a timer in front of you with time going from 00:00 to 2:00:00 (in this order), and always start a session knowing what is important to achieve. 5-6 sessions a day.  
I did all the afterworks, networking, and events during the past year and a half. Even if that meant going back home to sleep at 1-2 am, three times a week and waking up at 7 am.  
Always putting myself in the condition of having plans B, C, D, E, F, G, H, I in case plan A didn't work.  


What did I learn that this time I won't forget? 
The market decides everything. Give him what he wants. Innovate and break the barriers if needed.  
YCombinator was right on how to launch a startup. But this time, I understood it.  
In 12 days of work, I have the capacity to create a product that scales to 64K monthly active users. With $0 spent in ads.  
I'm expensive. If I want to make money, companies have already paid me $2,000 a day to access my strategic vision and my knowledge.  
I am actually very good at sales. I reached an 85% closing rate with no effort on the AI course I created on OpenAI's specialist agent (GPTs). My total revenue generated represents 20% of the total platform revenue (there are +30 online courses right now).  
People lie to you. The only compass to follow is gut feeling and whether people pay for what you offer.  


This journey taught me that:
Books are 20 to 40 years of summaries in 400 pages.  
Storytelling and simple communication drive 100x more engaged people.  
Public speaking is challenging, and people want opinion leaders, not flat water. This is also what you want for yourself to succeed.  
The vision you see is prophetic: keep it, or work on yourself and change it.  


At 23, it seems like a good starting point in life.

Artificial Consciousness

Humans are not much more than biological computers that can store informations and access it.

If you agree with this, then we will probably soon enough see superintelligent robots, with the particularity of feeling emotions.