I believe it is essential for all AI entities, including startups, to dedicate resources to AI safety. The risks are systemic, and early-stage companies should also be part of the solution, though perhaps not to the same extent as the largest firms.