The Digital Press

A 1999 MIT thesis titled "If it works, it's not AI" examines the commercial challenges faced by AI startups, highlighting the evolving perception of AI technologies as they mature and become commonplace. The thesis reflects on the historical AI industry struggles and draws parallels to contemporary AI business dynamics.

Why it matters: AI technologies often lose the "AI" label once they become useful, affecting how startups market and develop their products.

The big picture: Early AI failed due to tech-business mismatches, high hardware costs, and integration difficulties; similar challenges persist today in AI adoption.

The stakes: Modern AI startups face unpredictable compute costs and must build domain-specific software to deliver real utility beyond simple chatbots.

Commenters say: Many note the persistent AI effect where effective AI is no longer seen as AI, stressing the importance of practical integration and sustainable business models.