a state in which cognitive resources are focused on certain aspects of the environment rather than on others.

Treating attention as a scarce commodity to be captured has led to more than a few business empires being created. Herbert A. Simon was the first to theorize of the attention economy when he wrote:

In an information-rich world, the wealth of information means a dearth of something else: a scarcity of whatever it is that information consumes. What information consumes is rather obvious: it consumes the attention of its recipients.

What I find more interesting however is his note on designers of information systems. Apparently many design their information systems incorrectly. They represent their design problem as information scarcity rather than attention scarcity and as a result they build systems that provide more information to people when what they really need is systems that excels at filtering out unimportant or irrelevant information. Now this is a very hard problem. It’s not amenable to a simple straightforward adjustments as it’s difficult to know what relevant information has been missed, how do you determine what information should be absent?

Many search engines have personalization of some form, where some information is absent in favor of other information for whatever demographics. Algorithmically these systems decide to filter information through processes that are a black-box to its users. It’s an attempt at filtering out irrelevant information but is this personalization approach an ideal way to do so? What are the societal effects of personalizing information systems to individuals? Clearly there’s risks of algorithm-based filter bubble formation since people often form opinions from the information sources they consume. What happens when the infospheres people live in provide conflicting facts? And that’s not even to mention the political implications of personalization.

With the advent of LLMs such as OpenAI, Anthropic, or any other provider, these questions become more pressing. Now we must also worry about the benefits and risks of personalizing chatbots.