Next week I will present my paper Co-Disclosing the Computer: LLM-Mediated Computing through Reflective Conversation at CHI 2026 in Barcelona, one of the leading international conferences in human-computer interaction.

The paper explores what happens when computers are no longer primarily encountered as fixed applications with predefined interfaces, but as systems that can be shaped dynamically through interaction with large language models. We argue that this points toward a different way of understanding computing: not mainly as navigating ready-made software, but as entering into a reflective conversation where intentions are expressed, interpreted, reformulated, and made computationally actionable.

In the paper, I describe this as LLM-mediated computing. Rather than seeing the interface as something fully given in advance, the computer’s functionality emerges in the interaction between human, language model, and machine. This opens up new possibilities, but also raises important design questions about steerability, repair, responsibility, and what kinds of actions the system makes available.

At CHI, I will talk about this shift and discuss how concepts from interaction design and postphenomenology can help us better understand it. My hope is that the work contributes to a broader discussion about what computing may become as generative AI increasingly moves from being an add-on inside applications to becoming part of how computational action itself is configured.

I’m looking forward to presenting the paper and to the discussions that follow.