# The Logo That Has to Be Everywhere at Once
In 2015, Google changed its logo for the first time in sixteen years. The shift to a geometric sans-serif — Product Sans — was framed not as a rebrand in the traditional sense but as a preparation: a visual identity being made ready for a world of many screens, many sizes, many contexts. The four colours stayed. The logic behind them was extended. The move was, at its core, an act of structural thinking disguised as aesthetic tidying.
For a while, that logic held. A logo that could scale from a smartwatch face to a billboard without losing legibility was genuinely useful in a multi-device world. Coherence was the asset. Recognition was the strategy.
Then, in 2023, Google brought its AI assistant work under the Gemini name — and the same visual language was asked to stretch across a chatbot interface, a voice assistant, image generation, and video synthesis simultaneously. Each of those modalities runs on different spatial and temporal rules. A chatbot lives in text that arrives sequentially. A voice assistant has no visual presence at all until someone decides to give it one. Image generation produces artefacts; video produces duration. The question of what a brand *looks like* across those four contexts is not a scaling problem. It is something stranger than that.
James Hurst worked with the studio Koto on a project exploring exactly this territory — how Google's visual language might flex for a multimodal future. The specifics of that work sit in the archive, undated, and it would be wrong to read them as a response to Gemini in particular. But the underlying problem the project was probing feels like it has only sharpened since. What does a coherent visual identity actually mean when some of your most-used interfaces have no fixed visual form?
## When Familiarity Becomes the Constraint
There is a version of this question that is purely technical: how do you render a four-colour wordmark in a voice interface? You don't, obviously. You find an audio equivalent, a sonic logo, a tone, a rhythm. Problem solved, or at least addressed.
But there is a harder version underneath that one. Google's visual identity is among the most recognised on earth precisely because it has been consistent for so long. That recognition is not neutral — it carries associations, expectations, a felt sense of what Google *is*. When Gemini generates an image or produces a video, the brand is making something, not just displaying something. The relationship between user and interface shifts. The logo is no longer a label on a search box; it is, in some sense, a signature on an output. That is a different kind of claim.
James' book *Weird the Normal / Normal the Weird* identifies the core design discipline as the ability to switch between two modes: estranging a familiar system to make its assumptions visible, and giving a strange new idea enough structure to survive contact with the world. The book frames this as a pulse, a way of staying alive rather than a destination to reach.
What makes the Google case genuinely interesting as a test of that framework is that the brand is being asked to do both moves at once, not in sequence. The identity is deeply normal — globally recognised, decades embedded — and it is being asked to hold together interaction modes that are genuinely strange, still finding their shape, still surprising their own designers. That is not the orderly switching the framework describes. It is more like being asked to be the stable container and the experimental contents simultaneously.
> The question worth sitting with is whether a single coherent look is the right ambition at all, or whether multimodal brands need something more like a set of related instincts — things that feel like each other without needing to look like each other.
## A Pattern Showing Up Elsewhere
Google is not alone in this. Any brand that has moved from a single primary channel into several genuinely different ones is facing a version of the same pressure. A retailer whose identity was built around physical packaging now has to mean something in a voice reorder, a push notification, and an augmented-reality try-on. A bank whose brand was built on the gravitas of branch design now has to carry that weight inside a chat interface that responds in milliseconds.
What these cases share is that the old definition of brand consistency — same colours, same typeface, same tone of voice across materials — was designed for a world where the brand *appeared* in contexts it did not control. A poster, a TV ad, a shopfront: the brand showed up, made its impression, and left. Multimodal interfaces are different because the brand is not appearing in the experience. In many cases, it *is* the experience. There is no separation between the identity and the interaction.
That distinction might be where the real pressure lives. Not: can the logo scale? But: what is a logo even for when the thing it labels is generating the world the user is inside?
## The Live Tension
James' framework, tested against this, does not break — but it does get complicated. *Weird the Normal / Normal the Weird* assumes that the switcher has some agency over the timing: you choose when to estrange, you choose when to stabilise. A brand identity operating across voice, image, video, and text simultaneously does not get to choose its timing. All four modes are live at once, each pulling in a different direction, each requiring a different kind of legibility.
Which raises a question the framework does not quite answer yet: what happens when the switching has to happen in parallel rather than in sequence? Is that still the same discipline, just faster? Or is it a different problem entirely — one where the goal is not a coherent look but something more like a coherent *disposition*, a set of values that can be recognised across forms that share almost no visual DNA?
Google's four colours have held for over a decade. Whether they can hold across everything Gemini is becoming — that seems like a genuinely open question, and one worth watching without rushing to an answer.