Some of the most interesting moments in research come when a problem begins to look different. An idea from another part of physics, or a mathematical language one has not used before, can open a direction that was difficult to see.
AI is broadening the range of ideas and connections I can explore in my research in quantum field theory. I can follow a suggestion into unfamiliar territory, ask what lies behind it, and bring it back to the question that started the discussion.
One example comes from my exploration of semiclassical approaches to quantum field theory. With AI’s help, I became aware that mathematical structures relevant to our problem had already been studied in other contexts. This suggested that some of those ideas and methods could be brought into our work on important non-perturbative questions in an Ising-like theory.
What interested me was the possibility of putting those structures to use. Understanding which features carry over, and where the analogy breaks down, becomes a research problem in its own right. AI helped me recognise a connection worth pursuing; developing it into a controlled calculation is the work that follows.
The process is rarely linear. I move between conversation, analytical work, and calculations on a computer. Sometimes a numerical check makes a promising idea less convincing. Sometimes trying to formulate the idea precisely reveals a better question. The value often lies in this back and forth, while the research is still taking shape.
Of course, a connection can sound persuasive and still lead nowhere. I need to understand why it should hold, which assumptions it depends on, and whether it survives a calculation. Familiarity with the physics matters especially when the discussion moves into less familiar mathematics.
What I can describe is a change in how widely I am able to look. AI is helping me bring more ideas into contact with the problems I care about in quantum field theory. That is already useful, and it makes me curious about where these connections may lead.
