Part 11/19:
The core of the discussion pivots to analyzing growth trends in AI hardware and algorithms. Many have relied on Moore’s Law — the idea that transistor counts double approximately every two years — to justify expectations of near-term breakthroughs.
However, the speaker points out that Moore’s Law is potentially reaching its limits, which might mean that growth is actually sigmoid (S-shaped) rather than exponential. If the current rate of technological progress slows, the singularity — often predicted as the moment AI surpasses human intelligence — could be centuries away rather than imminent.