Qwhen it comes to generative artificial intelligence (AI) [capable de créer, à partir d’une simple instruction écrite, du texte, comme ChatGPT, ou des photos ultraréalistes, comme Midjourney]the reactions of the French and European public authorities are reminiscent of one of the most famous fables of Jean de La Fontaine: The frog who wants to be as big as an ox.
During the economic meetings of Aix-en-Provence, on July 8, the Minister of the Economy, Bruno Le Maire, estimated that Europe will have to have its OpenAI “within five years”. An outing reminiscent of that of the Head of State during the last edition of VivaTech, on June 14. On this occasion, Emmanuel Macron explained at the microphone of CNBC that the development of major language models (like ChatGPT) was one of France’s priorities. An understandable ambition given their economic potential.
According to McKinsey, generative AI could create an additional wealth of between 2,600 and 4,400 billion dollars (about 2,350 to 3,990 billion euros) each year on a planetary scale. Moreover, investors are not mistaken: during the first five months of 2023, 12 billion dollars were invested in these AIs. Mainly in the United States, where Lhe main companies behind their development.
Because American companies in this sector have two unique advantages: the anteriority of research and the volume of funds available. Take the case of Google. In addition to developing Bard, the Mountain View company spent $300 million acquiring Anthropic and over $1 billion in Runway AI. All in one semester. Or the equivalent of the envelope of the AI plan for France presented in 2018!
Three major groups
Obviously, the public authorities are right to be ambitious. On the other hand, where they are wrong, it is to want to compete with the American “beef” of which we have neither the power nor the means.
As the Andreessen Horowitz investment fund, one of the most influential in Silicon Valley, explained in its report “Who Owns the Generative AI Platform?” »the mapping of generative AI players draws three major groups: designers of large language models, infrastructure providers, and finally companies that develop uses of generative AI (“use cases”).
The former are the stars of generative AI. Those to whom we owe the current revolution thanks to their major language models: OpenAI, Google, etc.
You have 52.3% of this article left to read. The following is for subscribers only.