You’re using GPT. There’s another person using Claude (me, for instance.) Another would be using Gemini. Many use DeepSeek. Then, there is that one person using Mistral.
AI models are dispersed. But here’s a fact - the majority of people care about two particular elements when it comes to those AI tools: payment and usage.
GPT launched a fascinating product that took the world by storm. I’ve used it early on and still occasionally use it. But something happened in the past year that is navigating my choice of AI models - the usage quality is becoming quite similar.
Google went to the drawing board - wrote this exact sentence “OpenAI wants people to use GPT and eventually pay for it.” and created a masterclass in business execution.
It’s About Payment, Not Just Usage
Google knew that all they needed to do was push GPT into the payment zone, where payment determines the choice for the user – then the ball is in Google’s court.
The AI Model Cashflow Problem
Every single prompt costs these companies money. GPT has enough, but they keep raising more to survive. Most of them do – Claude, Perplexity, etc. They all generate revenue but not enough to sustain them as a business.
If you’d think of those companies as bars – they have good drinks, some people pay for beer, but the majority (and I mean the majority) have a sip and leave.
The difference between Google and OpenAI is that Google has enough sips to give all those people. They anyhow charge them with their advertising model – “take a free sip, but hey, look behind you, yes, it’s a banner of an energy drink. Okay, you can leave now.”
Google has been monetizing using ads for years now. Pay for Gemini if you want, but also, you could just head to Google and use AI mode to search for whatever you want.
Along the way, you could see an ad or two that you might like. That’s all it takes for Google to keep making money (and I mean lots of it.)
So yes, Google can keep doing this for ten years. GPT simply aren’t that rich.
The Usage Battle
Google needed to reach the level of quality of GPT. Now, I’m being entirely honest with you - GPT vs Gemini vs Claude vs DeepSeek is like comparing iphone 16 to iphone 17 (in most cases.)
Quoting OpenAI – Three-quarters of conversations focus on practical guidance, seeking information, and writing. Two-thirds of all writing messages asked the chatbot to perform editing, critiquing and translating for existing text.
They have 700 million weekly active users. So hundreds of millions of users are asking GPT - “translate this” or “make this better”.
Do this with GPT, Gemini, Claude, DeepSeek, they’d all result in an acceptable output. Also, they’re all free (and that’s important.)
So in order for a business to grasp the majority of users, it just needs to compete with these “low-level” AI tasks. Google went on a code red and eventually released Gemini which does it quite well.
In terms of usage, Google played it well and became a contender. That’s already a catastrophic sign for GPT, and I’ve mentioned this in an article that I’ve written two years ago. If they reached this level, then GPT is already hanging by a thread.
Now Google will release a new Gemini. OpenAI will release a new GPT. But that won’t change the user’s behavior. They’ll still just ask that AI tool to rewrite something.
Every AI tool will work in the short term – as long as it has an average quality and zero pricing.
The AI tool with the biggest pocket will win in the long term, regardless of how top-notch its quality is.
This is a really insightful anlysis of Google's strategic positioning in the AI market. The comparison to a bar where Google can afford to give unlimited 'sips' while others struggle with cashflow is brilliant. You're right that for most everyday tasks, the quality difference between these models is minimal, making cost the deciding factor.