Part 6/12:
A central theme in Amade's warning is the race against time. Two timelines are unfolding simultaneously: one for interpretability—developing tools to understand how AI systems function—and one for superintelligent AI itself.
Current development pathways suggest that interpretability solutions may take 5 to 10 years to reach maturity. However, superintelligent AI could arrive as early as 2026 or 2027, before we can fully understand or control it. This race creates a perilous scenario: arriving at a place of superintelligence without the means to interpret or control it could be catastrophic.