According to OpenAI, the amount of computation used in AI experiments has increased by 300,000 times in the last six years. The largest experiment so far, AlphaGo, cost USD$10M, and costs are quickly increasing. So how long would it be until these experiments became prohibitvely expensive? And how many years will it be until these experiments use enough computation that they could produce artificial general intelligence? Ryan will present some fresh research addressing these questions.
How Big Can AI Experiments Get?
10:00 am - 10:30 am