The discovery of the Transformer: How Google f***** up

In the world of AI, a single idea can change everything. That’s exactly what happened in 2017, when a group of researchers at Google published a paper titled “Attention Is All You Need.” It didn’t sound flashy but inside was the blueprint for the Transformer, the technology behind ChatGPT, GPT-4, and the AI tools reshaping business today.

What is a Transformer?

Imagine you’re reading a customer’s email:

An illustrated icon of a unread email
“Hi, I want to change my booking from Saturday to Sunday. But only if there’s space.”

To understand this properly, you don’t just read word by word, you look at the whole sentence, remembering that “change” links to “booking,” and “only if” links to “there’s space.” This is also what a Transformer does.

Older AI models read text a bit like a conveyor belt, one word after another to slowly build understanding. Transformers read like a human. They look at the entire message at once, spotting relationships between words, even if they’re far apart. This “big picture view” is powered by a concept called attention. Attention is simply the model deciding which words matter most when understanding meaning. Think of it like customer service triage, not every detail is equally important. Transformers focus on the crucial parts.

Why it was a breakthrough?

An illustrated lightening bolt icon
Transformers process text in parallel, not step-by-step. That means they can be trained on huge amounts of data much faster.
An illustrated lightening bolt icon
Because of this efficiency, transformers could grow into models with billions (now trillions) of parameters.
An illustrated lightening bolt icon
They don’t lose context, so they produce much more natural and useful answers.

Without transformers, tools like ChatGPT, Bard, and Claude simply wouldn’t exist.

Why didn’t Google capitalise on it?

Google invented the transformer, but it was OpenAI that built ChatGPT and grabbed the spotlight. Google was making billions from search ads. Rolling out a chatbot that gave direct answers (instead of links with ads) risked undermining that business model. Large language models can also get things wrong (“hallucinate”). Google, as a giant with regulators watching, moved cautiously. OpenAI, as a smaller player, moved faster and took risks.

Why it matters for your business

The transformer shows how one innovation can ripple across industries. For small businesses, the lesson is simple:

An illustrated lightening bolt icon
Game-changing ideas often start in labs but quickly become tools for everyone
An illustrated lightening bolt icon
Big players are rarely the  fastest. That creates space for smaller companies to adopt and benefit before the giants fully commit
An illustrated lightening bolt icon
What seems “too technical” today (like transformers in 2017) can become tomorrow’s everyday tool

Google discovered the transformer but didn’t capitalise because it was too tied to its old way of making money. The opportunity for you is the opposite: be ready to adopt new tools when they arrive, even if they feel early. That’s how small businesses get ahead of competition.

An illustrated lightening bolt icon

What is a neural network?

The core architecture behind most AI models.

Read post
An icon showing reading glasses on top of a book

How do transformers actually work?

AI isn't magic. It is a set of mathematical instructions.

Read post
A photo of two people looking at a laptop, the laptop displays a Breezy reporting screen showing the number of customer contacts Breezy resolved that month.
Designed for you

Big tech, for the everyday business

Breezy is used by businesses across the UK, Europe and America. Our mission is to ensure that all businesses, regardless of size, can take advantage of the AI revolution. We offer a free trial so you can see how your business can benefit from AI.

Free trial