imo e/acc tweeting is net positive for the Safety perspective too: - debating against an opponent sharpens safety arguments, adversarial search for truth - 2 sided conflict draws attention to the issue n gives media a conflict to report on, all publicity is good publicity
- steering *towards* a desired outcome not just trying to avoid the scary outcome can help ppl envision the future - alt perspectives uncover useful ideas even if they're overall wrong- like my poll asking *why* we survive AI if in fact we do
Replies and quotes that continue this note
More notes on nearby ideas
Quote tweet · Apr 2, 2024@MichaelTrazzi cool 🫶Reply · Mar 31, 2023@verkehrt2 @doxometrist yeah, I've decided to do this but I'm still framing it as "Yudkowsky's perspective on AI Safety" because I feel like different…Reply · Aug 11, 2022@HikeMix @xlr8harder @ESYudkowsky @WallfacerAG Yeah, feels like we’re all talking past each other… had a dinner conversation abt AI Safety, wasn’t…Tweet · Mar 30, 2023not satisfied w my prev poll, so trying it this way... Is summoning a Yudkowsky-bot from GPT-4 to explain AI safety and answer questions for the…