Ask HN: When does patent strategy matter for early-stage startups?

I’ve been thinking about the timing of patent decisions in very early startups, especially in software and AI.

Most teams at that stage are focused on shipping and finding product–market fit. Filing patents can be expensive and slow relative to how fast products evolve.

For founders and engineers here: If you build something that could be considered technically novel but don’t file immediately, do you worry about someone else patenting something similar later? Or is that risk generally low in practice?

I’m kind of interested in what people have actually experienced.