The "attention economy" is a term that's typically used, negatively, to describe social media, algorithmic feeds, and ads-based monetization. This negativity boils down to two key problems:
- Platforms and user incentives are misaligned. The platform controls which incentives to optimize for.
- Attention is scarce.
LLMs provide a solution to both problems.
Follow along
Subscribe to get either as a newsletter — email only, no account required:
- World Cup 2026 — the whole tournament, across every team and matchup.
- Will the USMNT win the 2026 World Cup? — the US team's path to the final, with tactical perspectives on each game.
Taking back control
Creating an LLM-based feed allows users to take back control of their attention. An LLM-based feed is not going to be as optimized as a traditional recommendation model trained on millions to billions of interactions; but it'll be good enough for most users. This is because LLM-based feeds enable users to define their preferences with natural language. For example:
Will the USMNT win the 2026 World Cup?
Track expert predictions and betting odds on USMNT's chances to win the 2026 World Cup, including analysis of their games and path to the final. Their first game is against Bosnia and Herzegovina on July 1st.
I'm looking for tactical perspectives on the match from both USMNT and Bosnia perspectives. Especially insights from coaches and players on the Bosnian team. Only include content related to soccer and football.
Attention will always be scarce
Algorithmic feeds solve a real user problem: what content should I consume in a limited amount of time. For passive consumption, these feeds solve this problem (too) well. For active consumption, LLMs give the user control to find content relevant to a particular topic/goal. Bringing things back to the World Cup: the internet is full of sports content. Most of this content isn't particularly relevant to me.
For the past week, Bosnian sports content creators are more relevant to me than ever before. Because the USMNT is playing Bosnia and Herzegovina tonight. There is a predictable and cyclical pattern to this relevance, but it's not something that engagement-driven feeds optimize for. And even if they did, I wouldn't have the free time/attention to consume their content.
LLM Generated newsletters
https://joinheader.com/briefings/3a7ed1f1-62e1-443a-a451-2ba63ca3ba44
I've been using Header to keep track of World Cup news. It gives me a "good enough" summary of trends that I asked it to track. It enables me to break outside my information bubble without spending additional time on a topic. Which is perfect for topics like the World Cup with so many teams, matchups, and trends.
A slice of the briefing generated from the topic above — synthesized tactical insights, each traced back to its sources.