AI: from Nano Banana to Gemini and ChatGPT, why are they called that

Google has officially explained what it means “Nano Banana” and why he decided to call his model like this with which, in recent weeks, they have been generated with theIA more than a billion images starting from one conversational prompt. The name was born from a nocturnal brainstorm and does not belong, as one might imagine, to some strange and difficult to identify metaphor. It is the fruit of a game that, due to the virality achieved, he earned an official place within the “intelligent” architecture of the giant Mountain View.

The team of DeepMindin fact, was testing an anonymous version of Gemini 2.5 Flash Image and a code name that did not lead back to the parent company was needed. Product manager Naina Raisinghani, close to the deadline, proposed “Nano Banana”, merging two of her nicknames: “Naina Banana”, with which she was called by friends, and “Nano”, often used jokingly to recall her height. Uploaded to the platform in August, the model has become viral not only for his abilities, but also for that deliberately “nonsense” name. At that point Google decided to turn the nickname into a real brandbuilding the imagery of the banana around it through icons, colors and interface details. Essentially, there is no technical meaning behind it. In short, there is nothing technical behind it but a lot randomness combined with an effective strategy in marketing.

Nomenclature in artificial intelligence

The nomenclature inartificial intelligencemoreover, oscillates between descriptive choices and evocative choices. In many cases the name serves above all to give an easy label to complex systems or, which is not unrelated to Silicon Valleyto pay homage to superstitions, mythological and dystopian characters or figures. The best known case, of course, is ChatGPT: The name combines the idea of ​​”chat” with GPTacronym for Generative Pre-trained Transformer. GPT indicates a family of language models that uses the architecture Transformer and is based on preliminary training on large amounts of text. The Transformer, as an architecture, was introduced in the 2017 paper “Attention Is All You Need”a milestone for those who deal with this ever-changing sector. The first GPT models of OpenAI they arrived only later and recovered the setting.

Claude: homage to the father of information theory

Among those most at odds with the purely descriptive approach there is certainly one Anthropicfounded by Dario and Daniela Amodei, former OpenAI. The two simply chose a name for their model: Claude. The company explained that the name is a reference to Claude ShannonAmerican mathematician and engineer, considered the father of information theory. In 1948, Shannon formalized the concept of bits and the theoretical limits of data compression and transmission. A true “giant” on which many of today’s brilliant minds have rested and continue to rest. It is interesting to underline how, in this case, the link is above all cultural and historical: an explicit reference to a founding figure of the discipline to refer to.

Evocative names and rebranding

On the video generation front, OpenAI has named its model Sora with a very evocative intent. In Japanese, in fact, “sora” means cieloa meaning that several observers have read as consistent with a naming based on imagination and creativity creativity that must be applied when using it. OpenAI, in the launch material, describes Sora simply as a model capable of generating videos from text prompts. Google instead chose a real branding twist: from Bard to Gemini.

In an official post, Google explained that “Gemini” recalls the idea of ​​”twins” and is linked to the convergence of two internal search legacies: that of Google Brain and that of DeepMindmerged into the organization Google DeepMind who developed the Gemini family. In the European panorama, the French startup Mistral AI has built the brand around the name “Mistral”, which recalls the wind of the same name (Mistral). Those who live in Sardinia or in other well-trodden Italian areas know the strength and impetuosity of this natural element and can understand the connection well.

In the end, xAI of Elon Musk chose Groka term coined by Robert A. Heinlein in the novel Stranger in a Strange Land (1961) and used here in the sense of entity capable of “understand deeply”. Elon Musk explicitly linked his new creation to a literary reference significant for him, confirming how the choice of name (think of Tesla or SpaceX) is an integral part of the process of developing and launching a technological innovation. Unless, as for Nano Banana, a product manager with eccentric nicknames destined for virality intervenes, capable of transforming a game into a successful product.