10% chance of AI by 2018 —Shane Legg

Created by gwern on 2011-06-17; known on 2018-01-01; judged wrong by muflax on 2018-01-01.

• gwern estimated 5% on 2011-06-17
• gwern said “http://lesswrong.com/r/discussion/lw/691/shane_legg_on_risks_from_ai/ conditional on no global war/existential risk” on 2011-06-17
• gwern said “50% http://predictionbook.com/predictions/2818 90% http://predictionbook.com/predictions/2819” on 2011-06-17
• JoshuaZ estimated 8% on 2011-06-17
• kallman estimated 1% and said “Assuming the traditional “Human level” AI for these questions?” on 2011-06-22
• faul_sname estimated 2% and said “Assuming human performance levels, though not necessarily human-like behavior, it would take major advances in processing power to reach that level.” on 2011-07-01
• Konkvistador estimated 2% on 2011-07-21
• muflax estimated 99% on 2011-10-26
• muflax estimated 1% and said “(gah, inverted)” on 2011-10-26
• Anubhav estimated 0% on 2011-11-17
• Tuxedage estimated 1% on 2013-01-03
• Qiaochu estimated 1% on 2013-01-03
• adbge estimated 1% on 2014-01-29
• artir estimated 0% on 2016-03-16
• themusicgod1 estimated 5% on 2016-07-09
• JoshuaZ estimated 6% on 2016-07-10
• themusicgod1 estimated 4% on 2016-10-08
• JoshuaZ estimated 3% on 2016-10-08
• muflax   judged this prediction wrong on 2018-01-01.
• gwern said “in retrospect, I like Legg’s prediction much better. He saw the DL revolution coming.” on 2018-01-01