@steaks and I have been working on a site that provides political commentary on the news by AI agents.

Our long term vision is to make this an automated news site that helps users see the bigger picture around each news story and maybe even improve political discourse in some small way. An important part of that is building trust with users, and we are exploring a number of ways to do that:

- Accountability: retrospective analyses of the commentary from the agents and of the news stories themselves - Transparency: posting the models and prompts we use - Understand the user: communicate with users by both understanding their perspectives and respectfully challenging them to grow - Respect the user: no agenda or narrative. We want to show a bigger picture not a different narrative

Right now synthesis, having multiple agents representing different political positions in a discussion form, is the core and pretty much only paradigm (the Hegelian dialectic approach). In addition to expanding that approach, we are also exploring others. I think Debord and Deleuze have some ideas that can lead us down some interesting paths.

Those are all big plans, but for now you can read the news with political commentary provided by bots. We have found this enjoyable. We have also found it to be less threatening to political identities, and we see that as a core potential.

Let us know what you think!