Everyone has the same tokens. Now what?

We are living in the age of Intelligence as a Service. It is easy to get tokens and build something that others are building at the same time. Quite possibly, they are getting the same output too!

So in this age, how do you make something different? How do you stand out as a builder, a business or an individual?

Humans have always been good at adapting. It is in our nature. When we got electricity, there was a fear: what if we used up all the energy? We didn’t. We kept finding more, discovering alternatives and exploring new sources.

Tokens will go the same way. Even if the big data centers cannot serve everyone, people will find a way. I am optimistic about open models and self-hosted ones. With all the data out there and models trained for specific work, small language models (SLMs) will find a way around the big corporations.

So don’t wait for tokens to go away. Think of mobile phones, electricity, calculators, or any technology you see that runs on physics and energy. Once we adopt something, we find a way to keep it alive, even if it takes rivers, the sea or the sun. We don’t care. It is wired into our nature.

And with the same electricity, the same water and the same energy, people have built so many things. There is still not just one car company, one phone company or one AI company in this world, or one company for any purpose.

So yes, everyone getting the same tokens from similar training data is a problem. But it is also the biggest opportunity. The question is: how?

First: geography and distribution

A product that is killing it in New York, Seoul, the UAE or the UK might not have the same prompts written for it here in Bangladesh, in India or in Africa, and the reverse is true too. It would be hard to build a bKash-like system here in Bangladesh and compete on the same distribution. But you can bring in ideas from somewhere else. You have the same domain, the same tokens, the same everything; only the geography and distribution are different. It’s like remaking a viral video clip for your own audience.

Second: taste

Same tokens, same view, same everything. But the final test, by humans and for humans, still has its place, from food to clothing to software. So add taste: your own product design, compliance, localization, and local context for everything.

Take Bangladesh. You don’t get many US dollars; you earn in taka (BDT). You cannot run up costs in USD while earning in BDT, and converting BDT back to USD is not easy. I am sure other countries work the same way. A product built for that reality is a different product, even on the same tokens.

Third: a quality human brain

Put a quality human brain on top of every token system, generative LLM system and agent, and you win. Add two, three, four or five, and the win multiplies. The question, again, is how, and that has become harder. The better you master it, the bigger your edge in economics and in output.