Read Time: 7 minutes │ Subscribe to the FREE Newsletter │Join the Premium Patch

Happy Father’s Day to all the Abbas!

To celebrate Father's Day this weekend, I am excited to share my conversation with Dataiku CEO Florian Douetteau, a significant father figure in the AI world. We reflected on one year since he published his thought-provoking essay, "The Children of AI."

“The Children of AI”: One Year Later

In conversation with the author, father of 2, and Dataiku CEO Florian Douetteau.

He calls his Ammi Mama, and his children call him Papa.

Most in the AI world, however, know Florian Douetteau as the visionary co-founder and CEO of Dataiku. Dataiku provides software to empower any company to build and use AI safely, seamlessly, and collaboratively.

One year ago, Florian published an essay called “The Children of AI”.

The essay alternates between hypothesizing what the future with AI will look like across different facets of our lives - education, the human experience, the economy, the climate, society - and playing out this future through the fictional life story of a girl named Olive, born in 2023.

Many parents wonder what a future with increasingly powerful AI holds for their kids, so it’s pretty nice to have someone with an indisputably strong vision for AI map out what this future might look like.

I recently had the privilege of interviewing Florian about “The Children of AI” and his perspective on the state of AI one year after publishing the essay.

Today, we’ll learn how Florian’s childhood and fatherhood have shaped his journey to becoming a technologist, shaping our AI future. We'll delve into his unique perspectives on the evolution of AI, its impact on our daily lives, and how he envisions the future of AI, especially concerning children and education.

👩‍🌾 Let’s dig in!

Get to know Florian

Ruqaiya: How would you fill in this blank to describe your relationship with AI? __________.ai

Florian: Curious.ai.

Ruqaiya: Is that different from, say, 10 years ago, or has that always been the case?

Florian: It's always been the case. I was always curious about AI. And I was always curious myself. So both applied.

Ruqaiya: I am in perpetual Explain-Like-I'm-Five (ELI5) mode as my daughter has hit peak curiosity levels. How would you introduce yourself to a 5-year-old?

Florian: So I would say that I'm from Paris, France. I grew up there. When I was young, I wanted to be either a physicist or do things with computers. I liked the stars, Albert Einstein, robots; these kinds of things.

Ruqaiya: What is the first cool technology you interacted with as a child?

Florian: The first cool technology interaction I had in my childhood was when my Mamie (grandmother), who was a cleaner, saved up and bought me a computer when I was 5 years old. It was an Amstrad CPC 6128 with 128 kilobytes of memory.

It didn't have games initially, but it had a big manual where you could learn to code in BASIC, do funky sounds, change the color of the screen, and things like that by programming it. That was part of the fun for me at that age.

Ruqaiya: Wow, your grandmother must have seen something in you, to gift you a computer like this at 5 years old.

Florian: I think it goes back to my curiosity. When you're a kid, having those kinds of positive feedback loops where you can challenge your brain and get excited about learning new things is important. It could be seen as a privileged perspective, but I think that the luxury of being a kid is being in a danger-free environment and going through those loops where you can learn through trial and error.

Ruqaiya: So a combination of curiosity, plus these instances where you got to have this iterative kind of learning experience. Maybe that can explain the path that led from 5 years old to pre-Dataiku. What got you to the point of deciding to embark on building what is now Dataiku?

Florian: I was always fascinated by computers, and the idea that you would have at some point, computers that would talk, do text and programming, linguistics, chess engines - all of those things were part of my fascination.

So I got into programming, and I did mostly math most of my life and at school. But at some point, because I wanted to do something concrete rather than math problems on a whiteboard, I got to work in a tech company, which essentially led me to where I am today.

When you start working in a tech company, you learn a lot, meet lots of smart people, and both have the desire and opportunity, thanks to those people, to keep building new things. That's essentially what led me to where I am today.

Ruqaiya: Finally, can you please explain to my child what Dataiku is? What is this AI stuff her Ammi is always talking about?

Florian: I would start by explaining to her something - a device - she already knows.

Ruqaiya: Google Home. She asks Google Home to play her favorite shows on Netflix.

Florian: Google Home has lots of things to do with computers to work properly. Those things are called AI, or artificial intelligence, even if it's a bit of a big word. It's when you try to make a computer smart a bit. They're smart because they can remember something about you or something they can read online, in order to provide you with what you need.

There are so many people and so many ways to make computers smart. But it is also so hard that I worked on building things to make it simpler to build those systems, because we don't have enough people able to build them today.

Behind the Scenes of “The Children of AI”

Ruqaiya: Moving into this essay - was this idea something you had been ruminating on for a while? Or was it sparked by the launch of ChatGPT?

Florian: It was sparked mostly by ChatGPT. I do like to think in terms of hypotheticals about what the future would look like. But here my main motivation was this idea that with AI as it is today, there are lots of changes happening to our personal lives, geopolitical and economic impacts, all of it together. I didn't want to write an essay talking at a high level because that can be a bit boring.

What's interesting is to think more anecdotally about what could change in the way you choose your studies, what life is like at school, and so on. That better captures our imagination.

I realized that my older kids, at 16 and 18, could largely define themselves outside of things like ChatGPT. But I wondered, if I had kids now, how would they define themselves? What would be different about thinking of their future compared to my older kids almost 20 years before?

And I thought I could build an interesting story around that by having multiple steps, each one showing a different age of the same person growing up. That's what drove my motivation - trying to capture the multiple impacts AI could have, especially on the very concrete details of our daily lives and interactions with reality and managing job/life balance. Not just the big technical and geopolitical changes, but the smaller personal impacts.

Ruqaiya: The structure of the predictions about an AI future coupled with Olive’s story was also quite unique. Was that a structure you were iterating on as you wrote? Or have you seen this kind of two-sided coin format before?

Florian: No, I'd never seen that two-part format before. The two parts came from feeling that you needed a concrete story, but you also needed to provide some perspective and context. And this dual structure worked for thinking about and talking about the future because, any way, you need to make up a fictional story of the future. If you only talk about it in an essay, you don't capture what I think is very important with AI - that our daily, everyday life and the small details of how we interact with reality will actually change.

Ruqaiya: Was there any aspect of the writing process, especially the fictional parts, where you drew inspiration from books or other fiction you've read?

Florian: I'm probably unconsciously influenced quite a bit by Asimov in envisioning these hypothetical scenarios, but it wasn't a conscious inspiration at the time. The fictional writing process itself was actually very fast - mostly one session where I outlined what could be happening at each age from 1 to 50, and then maybe an hour or so fleshing out the details for each age bracket, already knowing the general themes.

Ruqaiya: Who was this intended for? And was the audience that consumed and responded to it what you expected?

Florian: It was intended for a very diverse global audience. Many provided touching feedback. I could have gone further, building on the themes and questions about the future - I only scratched the surface of many of them in this essay.

Ruqaiya: There's a swath of people, myself included, that would struggle to envision much of what you laid out. I think you've done a great service, particularly by posing that last question of what AI future we want for our kids and how to make it a reality.

Florian: A key aspect is the untapped potential of AI for education in those very first years of life. But there are many ways it could go wrong, too. My intuition is we'll first see great new AI education methods that get questioned for increased screen time. They'll also likely be very expensive apps or devices initially targeted at a narrow, wealthy audience where companies can most easily start versus trying to build a global education app from day one. It will probably spark controversy, but I suspect it will have a more profound impact than many other things.

Ruqaiya: I've also heard the flip concern - if less privileged children end up with AI-only education with little-to-no human interaction. AI in education is certainly a major entry point for most parents into the AI conversation. Speaking of which, I have to ask, were any Ammis involved in the making of this essay?

Florian: Yes, I discussed some hypotheticals with Ammis. And certainly there was the spirit of Ammis driving me in terms of empathy toward the risk of being a kid.

Looking Back and Looking Forward

Ruqaiya: It has been one year since you published the essay! Has your view changed at all versus when you first wrote it? Or is it mostly the same? And has anything this past year surprised you or deviated from your expectations in terms of how it would all play out and the speed of progress?

Florian: I was surprised by the rapid scale and speed of some developments like deepfakes. If writing it now, I'd probably have revised timelines since issues like deepfakes being everywhere for teenagers in 10 years are already here today and hopefully solved by then.

More broadly, it's hard to anticipate things like pricing and distribution - in 5 years, any phone could essentially be a free AI assistant, your "best friend." So many aspects around how you grow up, choose studies, access these complex systems over time - the impacts of ultimately moving towards a new digital realm while reducing harm to the physical world - I'd probably keep a good chunk of it the same even starting fresh today.

Ruqaiya: Regarding deepfakes, both in the essay and in real life, legislation to prevent or deter these tragedies comes as a reaction, after so much damage is done.

Is it realistic to think we can regulate these damaging AI use cases proactively and in a more specific way? Or is this regulation fated to be reactionary and vague?

Florian: We're past the point where every government will be proactive about it. We're past the point where technologies [that produce deepfakes] can reasonably be wrapped up.

Most regulations could be clearer on using someone's image to generate an AI image without consent. But there is a gray area now in terms of when you can and can't do it. In some cases, like using AI to draw a caricature of a political figure for satire, it seems fair as that political satire has been accepted for centuries. If clearly labeled as fake or a cartoon, it's probably okay.

But usages aiming to completely mimic and fake someone without consent should be banned. Governments are aligning toward that. But the way it could be enforced is still very limited.

Ruqaiya: What are you most hopeful or excited about in this domain in the next 10 years?

Florian: The impact of AI on scientific and life sciences research seems enormously underestimated and is very exciting within 10 years.

In writing the essay, I consciously avoided risks of AI that could be too dramatic, like bio-weapons.

There are real risks, too, beyond the opportunities I am excited about.

Ruqaiya: How has this past year of generative AI developments impacted your approach to leading your company and parenting?

Florian: For the company, it reinforced that in the data/AI field, you face a constant flux of technologies. The art of the possible is always changing. So our strategy as a company needs to be building a platform that can integrate new technologies over the long-term - how we operate, build products, and go to market. You can't operate in a steady state mode when you work in data and AI.

And as a parent, well, my kids are already far in their teenage years. I think it's reinforced the fact for me that you have everything you learn as knowledge, but it's more and more important at their age to get some core foundation in how to think and logic per se, be curious, form a good personality and relationships. Because personality and logic are probably not something that you can compensate with tools like ChatGPT. Knowledge is easier and easier to compensate with such tools. But being a good person, trying to be smart by yourself is key.

Ruqaiya: There was a study by Anthropic, makers of Claude, where they created their own constitution for AI development principles. Then, they compared it to a public survey of what people want, and there was only a 50% overlap of what these tech leaders put together and the public opinion.

As a technologist and leader in this space, how do you think is the best way for you and your peers to stay grounded and keep a finger on the pulse of the general public?

Florian: That's a great question. I think a core issue with technologies is when you lose sight of the practitioner or the end user, their spirit - meaning what excites the actual human beings using the technology.

For instance, what excites me about using an iPad is not the amount of things I can do with it, or that it’s thin, or playing music. It is about having a big screen.

If technologies are being built solely by people from a technology background, you can lose that intuition.

I believe in the collaboration pattern of having people from the domain also getting into technology quite a bit. Ultimately, getting the intuition for a domain is very important, and it takes years and a particular mindset. Technologists, including myself, we love technology for the sake of technology, let's be honest. When we get into domains like life sciences, education, and consumer behavior, we can be intellectually interested, but it's very hard to get the spirit of it quickly enough.

You need this mix where you have people coming from the domain who can actually participate and build technologies alongside long-time technologists in order to build systems that do not alienate people.

Ruqaiya: Systems that don't alienate people - I can get behind that. As a parent, what guidance would you have for parents outside of the AI hype bubble who are either apprehensive, terrified and don't want to touch it, dismissive about it being just hype, or nonchalant that okay, we'll cross that bridge when it really impacts our lives in a meaningful way? 

Florian: I think you need to insert yourself into the conversation. They must believe their kids are already using GPT for studies or coursework. That's just the new reality, at least for now. It will become even more clear this year or next that AI assistants are a reality for kids, just like social media is for parents. With AI, kids can learn things or look way smarter very quickly.

Ruqaiya: My last question is: When your kids were babies, what technologies were you thinking about or concerned about? Did it play out how you imagined?

Florian: I wasn't that deep into thinking about future technologies back then. I was just too busy parenting and dealing with things like getting them to sleep. I didn't really have specific thoughts, except that mobile tech and internet would be key parts of their lives since that was already happening, albeit still emerging 18 years ago when they were born. I have to admit I didn't anticipate just how big a role social media would play in their lives until a few years later.

🪴💻️ What’s in Florian’s Tech Stack? 💻️🪴 

Apple or Android: Apple

Mac or PC: Mac

Favorite technology: I’ve grown fond of the reMarkable tablet. The pleasure of writing with a true pen. I do like to scribble, to think. And so it's interesting to have everything in the same place without having all those notebooks that I keep losing. So yeah, I really like this one.

Tech-rec (recommendation) for Ammis and friends:

Looking at my apps, ChatGPT, probably.

An Aeropress coffee maker is good, especially for travel. When you travel and have jet lag, the issue is you wake up at 5 a.m. in a hotel where, at best, you'll get very bad coffee in your room. With an Aeropress and some logistics, you can just carry your coffee and make a perfectly decent, actually very good cup each day when traveling.

📚️ Bonus 📚️- Favorite book(s): Foundation, Dune, and Neal Stephenson’s works; I wouldn't know which specific Neal Stephenson book because he has so many good ones.

🌾 Cream of the Crop

🍃 Fun fact: I used to work at Dataiku for 4+ years, which makes Florian my former boss! This essay was one of the driving forces behind the creation of ammi.ai. To Florian, many thanks for penning this piece. And many thanks for sharing your childhood and fatherhood experiences and your invaluable insights with us! It was an honor.

🍃 Follow Florian and Dataiku on Linkedin to follow developments of the industry-leading Dataiku software platform that companies across the world use to build AI! 🚀 

🍃 Did you know you can try Dataiku for free? The best way to learn how any data, analytics, or AI process works is by building it. And the very best platform to build it is Dataiku. It’s so easy a Ruqaiya can use it (but so powerful the biggest companies in the world are using it) 🙂 

Ammis and friends, thank you for spending a few of your precious, precious minutes with us.

If you know someone who might enjoy this, please share it!

See y’all soon!
Ruqaiya
Ammi by day, Ammi by night