Commentary:
Summary (Heather Read, PhD, Neuroscientist & Entrepreneur)
Shane Legg, Co-Founder of Google’s DeepMind AI lab, predicted that by 2028, AI will have a 50% shot of being about as smart as humans. Here, Shane Legg outlines several problems and solutions.
- Problem: If validating that we have effective Artificial General Intellegence (AGI) relies on definitions of human intelligence, how do we test precisely?
- Solution-1: Researchers could assemble a battery of tests for human intelligence that could serve as standards for AGI.
- Problem: How do we scale up training and functionality of AI models?
- Solution-1: Insure that computational power will grow exponentially
- Solution-2: Insure that available data in the world would grow exponentially for a few decades.
- Solution-3: Start training models now with data that is beyond what a human can experience in a lifetime.
Source:
Futurism OCT 30 by NOOR AL-SIBAI
