The paper explores how industries like software, semiconductors, and computers have remained highly innovative despite weak patent protection. It argues that innovation in these industries is both sequential and complementary, meaning each new invention builds on previous ones and that multiple firms contribute to overall progress. This dynamic nature of innovation means that strong patent protection can hinder innovation by discouraging competition and imitation, which are essential for progress. The paper presents a model showing that in such industries, patent protection may reduce overall innovation and social welfare. It uses the 1980s extension of patent protection to software as a natural experiment, finding that this did not lead to increased R&D or productivity, contradicting the static model's predictions. The paper also highlights evidence of cross-licensing in high-tech industries and a positive relationship between innovation and firm entry, supporting the dynamic model. Empirical analysis of software patent data shows that R&D spending and productivity did not increase as predicted by the static model, consistent with the dynamic model's predictions. The paper concludes that the dynamic model better explains the innovation patterns in high-tech industries.The paper explores how industries like software, semiconductors, and computers have remained highly innovative despite weak patent protection. It argues that innovation in these industries is both sequential and complementary, meaning each new invention builds on previous ones and that multiple firms contribute to overall progress. This dynamic nature of innovation means that strong patent protection can hinder innovation by discouraging competition and imitation, which are essential for progress. The paper presents a model showing that in such industries, patent protection may reduce overall innovation and social welfare. It uses the 1980s extension of patent protection to software as a natural experiment, finding that this did not lead to increased R&D or productivity, contradicting the static model's predictions. The paper also highlights evidence of cross-licensing in high-tech industries and a positive relationship between innovation and firm entry, supporting the dynamic model. Empirical analysis of software patent data shows that R&D spending and productivity did not increase as predicted by the static model, consistent with the dynamic model's predictions. The paper concludes that the dynamic model better explains the innovation patterns in high-tech industries.