Thinking Swarms is a multidisciplinary exploration of swarming robotics. The breadth of discussion in the preceding chapters, and in particular the exploratory nature of some, makes writing any concluding chapter a challenge. It would be easy to assemble the proposed future work from each individual chapter and reproduce it here or to forge a completely separate path. This concluding chapter takes a middle ground. We synthesise recommendations and lessons from earlier chapters into a set of opportunities, but through a lens that resonates with this particular author. Without claiming to be definitive, we offer avenues for further exploration: a continual refinement of concepts to enrich the conversation in coherent ways; deeper investigation into social expectations; a greater focus on regulators as partners in the domain; expanding open world applications of our technology so they are resilient to both stochastic and epistemological uncertainty; exploiting large language models as critical “semantic” partners within a broader autonomous system; and strengthening and extending our simulation toolbox to support all of the above.