Against essential and accidental complexity

Yoshua... no. The article (by Dan Luu, though the text does not name its author in the excerpt) argues that Fred Brooks's 1986 essay "No Silver Bullet" wrongly claimed that technology can give at most a factor-of-2 productivity gain, because Brooks assumed at least half of programming complexity is essential. The author says Brooks underestimated gains from tools and practices that did not yet exist, such as version control, automated testing, and CI/CD. As examples, the text notes that Microsoft's Win2k team of 5,000 people could merge only about 100 changes per day in the 1990s, and that Brooks dismissed AI, which the author says neural nets have since made effective for speech and image recognition.