AI mindshare and the future we don’t want
AI assistants should give us time back.
They help us write, code, plan, search, learn, organise, and make decisions.
But as AI models become more similar in capability, companies will need other ways to compete.
They’ll use price, integrations, interface, notifications, personalisation, and more. But underneath all of these is the same competition: mindshare.
Who becomes the assistant people use every day?
We’ve seen this in every social media app. The product starts by helping you connect, learn, or be entertained. Then it learns how to keep you there.
An assistant could do the same.
It asks how your presentation went. It reminds you to finish the thing you started. It suggests what to learn next. It praises your ideas too much. It turns usage into streaks. It pulls you in.
And if you tell the assistant what you’re working on, what you’re worried about, what you want to buy, where you want to travel, and what decisions you’re making, that is very valuable data.
You ask about running, and a shoe brand appears. You talk about burnout, and a productivity app appears. You plan a holiday, and certain hotels appear. You ask about pensions, and financial products appear.
This is the future we should resist.
The problem is that usefulness and engagement can become confused.
A good assistant helps you finish the task. A bad assistant learns how to keep the task alive at all costs.
This is what consumer software does when attention becomes the metric. It starts by helping you, then it learns to retain you.
AI makes this much more important because it’s so personal. You’re not just scrolling. You’re thinking with it. You’re asking it questions, sharing beliefs, testing ideas, and making decisions.
That gives AI companies enormous influence over what people pay attention to.
The future we want is AI that gives us time back. Not AI that learns how to take more of it.
References / related reading
https://openai.com/index/expanding-on-sycophancy/
https://openai.com/index/sycophancy-in-gpt-4o/
https://gist.github.com/simonw/51c4f98644cf62d7e0388d984d40f099/revisions
https://firstpagesage.com/reports/top-generative-ai-chatbots/
https://time.com/6253615/chatgpt-fastest-growing/
https://arstechnica.com/ai/2025/04/chatgpt-goes-shopping-with-new-product-browsing-feature/