By 2020, Eliezer Yudkowsky will have given up building friendly AI.

Created by Tuxedage on 2013-01-13; known on 2020-12-01; judged wrong by Nic_Smith on 2021-01-01.

• Tuxedage estimated 9% on 2013-01-13
• Tuxedage said “By “given up” I mean personally, rather than declare friendly AI impossible.” on 2013-01-13
• Nic_Smith said “How to be judged if he builds a [F?]AI before 2020? Also, “known by” date looks off.” on 2013-01-14
• Tuxedage said “Nic: If he builds FAI before 2020 (however unlikely that is), then this prediction will be wrong, as he’s succeeded in building it. ” on 2013-01-14
• Tuxedage changed the deadline from “on 2014-01-01” on 2013-01-14
• Nic_Smith estimated 5% on 2013-01-14
• JoshuaZ estimated 11% on 2013-01-14
• Qiaochu estimated 5% on 2013-01-14
• Cy Fleming estimated 5% on 2013-01-15
• Emanuel Rylke estimated 2% on 2013-01-22
• ChristianKl estimated 15% on 2013-01-24
• gwern said “many ways to give up without appearing to give up” on 2013-01-26
• Slahzer estimated 10% on 2013-01-26
• adam_strandberg estimated 1% on 2013-02-09
• John Maxwell IV estimated 3% on 2013-02-10
• tormey estimated 5% on 2013-02-10
• Yvain estimated 7% on 2013-02-13
• halfwit estimated 1% on 2013-06-14
• Pablo estimated 6% on 2013-08-09
• Michael Dickens estimated 10% on 2014-01-01
• itaibn estimated 10% on 2014-01-02
• itaibn estimated 10% on 2014-01-02
• simplicio estimated 10% on 2014-01-02
• holycow81 estimated 14% on 2014-12-10
• themusicgod1 estimated 11% on 2016-04-14
• olgaidzh estimated 2% on 2017-02-19
• pranomostro estimated 5% on 2018-09-02
• Baeboo estimated 2% on 2018-12-12
• Baeboo estimated 1% on 2020-11-20
• PlacidPlatypus said “Is there a particular firm metric for this? As far as I can tell he still seems pretty focused on AI so I’m inclined to mark Wrong.” on 2020-12-11
• Nic_Smith   judged this prediction wrong on 2021-01-01.