“There’s so much anxiety. It’s so evident at events where artificial intelligence is discussed. Investors are terrified of what might happen. Terrified and elated at the same time. The amount of money at stake is huge. And with this technology, the winner takes all.” Marco Trombetti was among the few Italians present at the Artificial Intelligence Action Summit in Paris. Upon returning, he took a few days before talking about it. He must have been busy, that’s for sure. He is one of the Italian entrepreneurs who knows the most about AI. His 25-year-old company, Translated, has been working in the field of translations and AI since its foundation. He sells its services to SpaceX, Uber, and Airbnb. He’s had time to process impressions, messages, and guidelines. “It’s hard to say how hopeless a return on investment is today. I see no prospects. Only chip and infrastructure producers make money nowadays. The others are doing trial and error, many are losing, and many more will lose in the future. Their fear is justified.”
Even you, Mr. Trombetti, doubtful of AI’s prospects?
“Far from it. I believe it’s the greatest economic and social revolution in humanity’s history. However, investors in the sector have a huge problem today: no one can afford to get the timing wrong.”
What do you mean?
“Seen from the outside, there seem to be a lot of people ready to put their money in. I mean, investors ready to play the game. But every euro invested is underpinned by anxiety. Everyone is afraid of making mistakes. This is true today more than ever in the history of the digital economy. Because today getting the timing wrong, investing wrong, focusing on the wrong company means getting badly hurt. The amount of money that’s swirling around this technology is staggering.”
Anxiety and enthusiasm seem like difficult feelings to combine.
“In Arabic there is a phrase that I find particularly apt: Fil Mishmish. It means more or less ‘The apricots will arrive tomorrow’. It’s an expression used to close a negotiation that is not leading to anything because it promises a future that will never arrive. It’s a way of postponing hope, of giving the illusion that something positive is about to happen, when in reality it’s only being put off.”
The first technological wave certainly brought apricots: digital giants and dozens, hundreds of prosperous companies. What’s different now?
“Today everyone is afraid that there will be only one winner. The first wave of the internet was a gold rush. The web offered an infinite number of possible applications, cutting costs. Many different companies could be founded for each market sector. The Internet was an enabler of things, and everyone could do it by launching projects with loose change. Today you need billions to enter this race. What’s more, there are no second or third places when AI advances to a new level. Everything else quickly becomes obsolete.”
Are you saying that billion-dollar companies could crumble?
“I’m saying that everyone today is rushing to reach the most powerful level of artificial intelligence. AGI, to be clear. However, there’s very little difference between the companies doing it. Everyone is training more or less with the same data. Meanwhile, neural network training is no longer a CapEx (i.e. capital expenditure, which is fixed and one-time, like a computer – Ed.) but has become an OpEx (an operating expense, like electricity – Ed.). In such a scenario, what’s new immediately makes the rest obsolete. Whoever falls will get hurt.”
What will these companies be left with then?
“The technology is replicable. Only human capital will remain, little else.”
So, whoever wins the AI game wins everything?
“That’s the most widespread fear at the moment: that only one company will win the global race of artificial intelligence. To be clear, the fear is that instead of 5 big tech companies, there will be only one.”
Is this scenario likely?
“It’s hard to tell. DeepSeek has shown that even if you can’t beat them, you can cause trouble for the biggest with few resources. So the smaller ones can have their market share, stemming the possibility of monopolistic concentration. But these dynamics are difficult to predict. That’s why there’s anxiety. However, for the same reason there’s hope, and positive signs can be seen from present and future competition.”
How did it go in Paris?
“For me, it was a great French success and a great European flop. Macron was clear and strategic. The other European countries didn’t even show up. Macron has understood the impact of AI and is building a block-by-block strategy to get the most out of it, focusing on its strengths: nuclear energy, data centers, and language models like Mistral. He is investing in human capital, converting polytechnics into artificial intelligence schools and aiming to create 100,000 AI talents a year. It’s a strategy that is missing elsewhere.”
It’s missing in Europe.
“America and China are doing their best to claim technological supremacy in this area. Macron has moved a pawn, so he’s playing his game, no matter how small. Europe hasn’t understood how or what game is being played, so it doesn’t know how to play. In the medium term, it risks losing wealth, quality of life, and above all market. While the world is facing an epochal revolution, Europe is regulating. “This is a negative message from all points of view and for all partners, competitors and new generations, who feel they are living in an old country. For the latter, the indication is clear: if you want to innovate, then leave.”
Don’t you like the AI Act either?
“It’s not that I don’t like it. It’s just that it’s already obsolete. It regulates mechanisms that have ceased to exist. Also, it prevents innovations in the sector. Just consider the emotion recognition issue. I deal with translation. How can I improve my service if I can’t understand the emotions that accompany an expression? It’s just one example, but what has been regulated goes beyond what the legislators could have imagined.”
Are you saying that the legislators didn’t have a clear idea of what they were legislating on?
“Indeed.”
What should politics do?
“My point about politics in my Paris talk was: you’re not influencing innovation, but we need to guarantee long-term assets. And at the moment one of the best assets we have as Europeans is Greenland.”
Greenland?
“Indeed. There, Europe can create data centers with lower cooling costs. It’s also a strategic point of data contact between Europe and the United States, since low latency transmission is possible. It might be difficult to even think about it today, but tomorrow it will be fundamental. Greenland must be defended.”
You’re speaking about tomorrow, the long term. Is it still possible to think of long-term strategies after the time that has been lost?
“Lost time is an illusion. Time isn’t working against us. The gap that the US opened up over the rest of the world with OpenAI has already been reduced. Time can be our ally. But the game must be played from a European standpoint.”
So what worries you about Europe?
“There are a lot of things that worry me. The issue of rules worries me, because they seem to be made with a lack of expertise. I believe that the legislators failed to understand the phenomenon. They confused risks with opportunities. The legislation is short-sighted compared to reality. But there’s a bigger problem.”
What’s that?
“That this approach will scare away young talent. Innovators don’t feel welcome in Europe. If you ask yourself what you have to do to welcome talent, you understand that the AI Act does the opposite. They don’t understand the phenomenon. They impose huge sanctions for every mistake, and those who innovate know how important mistakes are.”
English version of an interview published in Italian, here.
Fonte : Repubblica
