Reginald Andreas

AI Doomerism is totally on the way out. In the last month, two King AI Doomers have both distanced themselves from their "AI apocalypse" narratives. On the Curt Jaimungal podcast, Max Tegmark talks more about tooling and defining consciousness vs intelligence. Doom? Foom? Gone. What did he actually say? "Max Tegmark described "doom narratives" around AI as pernicious during his appearance on Curt Jaimungal's podcast, arguing that excessive focus on catastrophic scenarios can have negative social impacts and skew public discourse. He cautioned that only presenting stories of inevitable disaster is harmful and undermines efforts for balanced, constructive discussion about AI risks and opportunities." My dude, you started this fire. But what about Roman Yampolskiy, who's gone from "there's a 99.9% chance AI kills everyone" to "AI will just take 99% of jobs" as seen on the Diary of a CEO podcast. (To be fair, he still says "we may lose control" one day) Likewise, even Geoffrey Hinton has oscillated on the AI doom perspective: "Geoffrey Hinton has repeatedly expressed worry about existential risks from AI, but he has also explicitly stated that people who claim with absolute certainty that "AI will kill everyone" are "bonkers" or "nuts". Hinton has emphasized the importance of caution and careful risk mitigation, but considers extreme certainty in predicting human extinction due to AI to be unwarranted and unreasonable, since the future remains fundamentally unpredictable." Funny context: Hinton made these remarks in context of debates about AI risk, where some public figures, including Eliezer Yudkowsky, have voiced absolute certainty about "doom".