A

Input Streams Are the New Bottleneck Everyone has the same AI models. We have infinite intelligence at the touch of a few buttons. So what becomes the bottleneck? What is the good stuff that now gives you the competitive advantage? Your input streams. By input, I mean valuable information that transforms your action in some way. It is not to be confused with consumption, a modern disease which plagues the minds of people and inhibits timely action. Input streams are everywhere. You can’t escape them. They exist as the people you talk to, the context in which you work from, the habits and practices you keep, and the books you keep reading. The most valuable input streams now are those that act as closed loops: as you act, you receive feedback from reality, and that feedback gives you new information only you can see. Nobody else took your action in such a way within your individual context. A valuable input stream reveals problems before they are made completely legible to everyone else. Because of this, your social media feed is now a competitive advantage. Of course, this has been known for a while. Traders have always paid for a faster wire, and journalists have always gatekept their sources. But what makes this idea so much more valuable now? Because now processing that information is much less costly: you have a professional software engineer at your fingertips. When everyone can do information transformation, what matters becomes the information, not the process. To be clear, processing still matters. Two people can have the same information and act very differently on it, which gives them different results. Information is also reflexive: once many people act on it, it changes the outcome. So look at what everyone else is doing and go orthogonal. Rely on your own unique context, and use the models to be precise to you. Models typically give everyone the same default answers, so using this to your advantage means using your information to move away from the default. There lies a gap between what gets done and when it gets captured in a way that is made legible. In this gap lies the most informational value right now: tacit knowledge, individual context, and early signals. As an input channel, social media is much of where this information lies now, especially text-based platforms like Twitter and Reddit. Group chats are also an important input stream, as they give you insider information that is closed off to the public. The best input streams are private but participatory. Because of this, your feed becomes an arbitrage that you can trade on. Design your feed on purpose. Clear out mainstream information, because that is where the knowledge is already priced in. What you’re looking for is the stuff at the edge: what is currently unknown to most people but can still be acted on effectively, right now. No one cares about a theoretical startup idea that is based on a trend everyone read about last year. What can you do RIGHT NOW? Your social media feeds also provide high rates of gathering new context. What this means is that insights only have so long before they lose value, and to counteract this we must ensure our rate of gathering new context is high enough so our rate of action does not deplete the value of the insights. Our insights also have the potential to lose value when they are made legible at a mass scale, as now the models will have them baked into their training data. So the scarcity comes from your rate of gathering new context, plus your ability to use it an effective manner before it becomes common knowledge. For instance, if I have a consistent line of communication with workers in a domain such as healthcare, not only does this give me unique context, but it also gives me a high rate of gathering which means my insights won't be beat out by singular instances of usage.