AI Doom

The case for taking AI seriously as a threat to humanity

Ensuring machine learning programs are aligned could be one of the biggest existential problems of our time.

Mini

P(Doom) - Dario Amodei24

Mini

P(Doom) - Emmett Shear36

Mini

P(Doom) - Yoshua Bengio35

Will an AI related disaster kill a million people or cause $1T of damage before 2070?39% chance

By end of 2028, will AI be considered a bigger x risk than climate change by the general US population?41% chance

Plus

By the end of 2026, will we have transparency into any useful internal pattern within a Large Language Model whose semantics would have been unfamiliar to AI and cognitive science in 2006?36% chance

Will OpenAI + an AI alignment organization announce a major breakthrough in AI alignment? (2024)8% chance

By EOY 2026, will it seem as if deep learning hit a wall by EOY 2025?23% chance

Will mechanistic interpretability be essentially solved for GPT-2 before 2030?29% chance

Will I think all AI hell broke loose in 2024?5% chance

MIT Technology Review

Now we know what OpenAI’s superalignment team has been up to

The firm wants to prevent a superintelligence from going rogue. This is the first step.

Will OpenAI achieve "very high level of confidence" in their "Superalignment" solutions by 2027-07-06?5% chance

Will Superalignment succeed, according to Eliezer Yudkowsky?NO

Will OpenAI's Superalignment project produce a significant breakthrough in alignment research before 2027?NO

See more questions: