Why We Decided to Build Button

Over the last decade, trading has quietly become one of the biggest consumer behaviors on the internet. Robinhood brought millions of people into the market, and more recently products like Kalshi and Polymarket have expanded what people can trade on. Today, almost anything can become a market if enough people care about the outcome.

As more people started trading, something else changed too. Markets became increasingly driven by information. A tweet can move a stock. A research report can change an entire sector. A podcast can shape conviction before a trade is ever placed. News no longer just explains what happened. It changes what happens.

At the same time, we've gone through one of the biggest explosions of information in history. Institutional market data has become more accessible. There are endless newsletters, podcasts, research reports, filings, earnings calls, social feeds, and alternative data sources. Every day there is more to consume than the day before. You would think that having more information would make trading easier, but in many ways it has done the opposite.

The challenge today isn't finding information. It's figuring out what actually matters.

That became clear after spending months talking with sophisticated traders. Nearly everyone is using AI in some capacity. Some ask ChatGPT questions throughout the day. Others have connected models to brokerage accounts. Some have built elaborate workflows that tie together multiple tools. Yet almost everyone described the experience the same way. It works, but it feels held together with duct tape.

Every trader has built their own stack. One tool for news. Another for charts. Another for research. Another for execution. Another for backtesting. Then a collection of prompts they reuse every day. They've invested time making it work because they have to, not because it's how they would design it from scratch.

The more conversations we had, the more we realized the problem wasn't AI. The problem was that AI had simply been added to existing products. Large language models are incredibly good at summarizing information and generating investment narratives, but markets don't reward good narratives. They reward good decisions. Those are very different things.

Good decisions require more than a language model. They require live market data, historical data, quantitative analysis, and the ability to test an idea instead of simply discussing it. They also require something that's much harder to capture, which is context.

Context is often the difference between a good trade and a bad one. Two people can look at the exact same chart and reach completely different conclusions because they're influenced by different information. One trader follows a specific macro researcher. Another listens to the same podcast every week. Someone else pays thousands of dollars every year for institutional research because it shapes how they think about markets. That context becomes part of their decision making, yet today's AI largely ignores it.

A few months ago we decided to step back and ask ourselves a simple question. If we were starting from scratch today, what would we build? Not what investors wanted to hear. Not what everyone else was building. What did we genuinely believe traders needed?

The answer wasn't another chatbot.

We believed traders needed a product that actually understands how they trade. Something that understands their portfolio, their risk tolerance, the ideas they've been researching, and the sources that shape how they think. A product where you can start with a simple thesis and have it gather the relevant market data, test the idea, challenge your assumptions, and incorporate the context that actually influences markets.

That's why we decided to build Button.

We don't think the future is AI making every decision for you. The best traders will always make the final call. We think the future is AI helping traders think more clearly, process more information than any individual could on their own, and build stronger conviction before they put capital at risk.

If we can build a product that helps people make better decisions instead of simply generating better answers, we'll have built something worth using.

That's the future we're building toward.