If I were starting an AI startup today, I’d avoid certain directions entirely. For example, I wouldn’t focus on content generation or building foundational AI models.
These areas are saturated, and building foundational models is very expensive, often requiring resources that only a few companies can afford.
Instead, I’d focus on targeting a niche where AI can integrate seamlessly without being the primary focus, using AI agents to significantly improve efficiency and reduce costs.