One day, one generative AI tool

Focus

Three-year-old generative AI: a quick assessment of a mad race

Focus of Letter 13

November 30, 2022, November 29, 2025… 3 years since the arrival of generative AI in the eyes and hands of the greatest number, through a simple messaging-style interface that provided a few sentences in response to our queries.

In three years, applications and integrations have followed one another at a frantic pace, with new publications every day – oscillating between shells supported by good doses of marketing and real technological advances – and a now proven anchoring in our daily lives, far from being exclusively dedicated to professional activities.

This mad race for novelty may now show new balances that are taking hold. With Gemini 3 Pro seen last week here, Nano Banana and many others, rankings like LMArena’s, see OpenAI’s models dethroned for a long time… until the next models arrive?

We can also remember the generations of images from three years ago with distorted faces, 6-fingered hands… and compare them to today’s generations such as those offered by Nano BananaFlux 2 or Qwen Image recrossed this week via Qwen Image Edit Camera Angle Control : the evolution of these generations impresses as much as it questions, or even makes some of our fundamentals waver, such as information processing…

The galaxy of generative AI applications has certainly not finished revealing new territories… To be followed every day and every week 😉