Will Capabilities Generalise More? - AI Alignment Forum

Nate andEliezer (Lethality 21) claim that capabilities generalise further than alignment once capabilities start generalising far at all. However, they have not articulated particularly detailed argu…

x Will Capabilities Generalise More? — AI Alignment Forum AI Frontpage 58 Will Capabilities Generalise More? by Ramana Kumar 29th Jun 2022 5 min read 39 58 Nate and Eliezer (Lethality 21) claim that capabilities generalise further than alignment once capabilities start generalising far at all. However, they have not articulated particularly detailed arguments for why this is the case. In this post I collect the arguments for and against the position I have been able to find or generate, and develop them (with a few hours’ effort). I invite you to join me in better understanding this claim and

related reading

• Alignment remains a hard, unsolved problem — LessWronglesswrong.com
• Alignment likely generalizes further than capabilities.beren.io
• A central AI alignment problem: capabilities generalization, and the sharp left turn — LessWronglesswrong.com
• Defining alignment research — LessWronglesswrong.com
• The Universe from an Intentional Stancecasparoesterheld.com
• Where I agree and disagree with Eliezer — LessWronglesswrong.com
• The Artificiality of Alignmentjoinreboot.org
• How far does alignment midtraining generalize?alignment.openai.com
• Alignment will happen by default. What’s next? — LessWronglesswrong.com
• Against ubiquitous alignment taxes — LessWronglesswrong.com
• A positive case for how we might succeed at prosaic AI alignment — AI Alignment Forumalignmentforum.org
• Readings on the nature of alignment researchcasparoesterheld.com