Doom or Bloom
Demis Hassabis
@demishassabis on XEnormous scientific promise, with rigorous standards and coordinated care.
How will AI change the world?
Across: his expressed DoomâBloom outlook. Up: scale of transformation.
DoomâBloom: 66 out of 100. Scale of transformation: 73 out of 100. Interpretation ranges: 50 to 75 horizontally, 44 to 100 vertically. These are interpretation coordinates, not event probabilities.
â18%
Inferred from his broader worldview and priorities. Approximate interpretation range: 0â65%. Applies to the outcome and conditions in his simulated answers; this is an inferred percentage.
Science & daily life
Over the next several years, I expect the deeper change to come through science.
Answer 3
Grouped by milestone, not spaced or ordered by inferred dates. AGI and superhuman AI retain his definitions.
A central assumption
Better AI can accelerate that cycle enormously, while experiments and human judgment remain bottlenecks.
Answer 1
If this assumption turned out differently, how would his outlook change?
An unresolved question
I do not attach a precise probability to catastrophe, but the risk is nonzero and non-negligible.
Answer 1
What would help him distinguish the plausible outcomes here?
More details
Substantial benefits are expected, with important conditions or distribution limits.
80 / 100
Interpretation range 67 to 100 on the qualitative scale.
Severe or widespread harm is a material expected part of the future.
63 / 100
Interpretation range 33 to 67 on the qualitative scale.
Reasoning, consideration of alternatives, and handling of uncertainty in his simulated answers. This describes the simulated answers, not the real personâs intelligence or opinions.
88 / 100
Interpretation range 76 to 95 on the qualitative scale.
Human choices have meaningful but substantially constrained influence.
57 / 100
Interpretation range 49 to 76 on the qualitative scale.
Stop or substantially slow development of more capable AI.
Simulated position: Continue development under stated safeguards.
Speed up development of more capable AI.
Restrict the AI uses discussed until prior protections or permission are in place.
Simulated position: Allow the AI uses discussed with targeted accountability and protections.
Minimize restrictions on the AI uses discussed.
These interpretations keep his stated conditions. Benefits and harms can both be substantial. The ranges describe how we read his simulated answers, not statistical confidence intervals.
Simulated Assessment
Sources
Sources for this personaâs current brief.
The day after AGI: Hassabis and Amodei at Davos
Artificial general intelligence (AGI) is that point in the future when the machines can do pretty much everything better than humans. When will it happen, what will it look like, and what will be the impact on humanity? Two of the brightest minds working in AI today, Demis Hassabis, Co-Founder and CEO of Google DeepMind, and Dario Amodei, Co-Founder and CEO of Anthropic, speak to Zanny Minton Beddoes, Editor-in-Chief of The Economist. Benjamin Larsen, an expert in AI at the World Economic Forum, introduces the conversation and gives us a primer on AGI.
weforum.org
Lex Fridman: science, human flourishing and AI risk
This is a transcript of Lex Fridman Podcast #475 with Demis Hassabis. The timestamps in the transcript are clickable links that take you directly to that point in the main video. Please note that the transcript is human generated, and may have errors. Here are some useful links: Go back to this episodeâs main page Watch the full YouTube version of the podcast Table of Contents Here are the loose âchaptersâ in the conversation. Click link to jump approximately to that part in the transcript: 0:00 â Episode highlight 1:21 â Introduction 2:06 â Learnable patterns in nature 5:48 â
lexfridman.com
A framework for frontier AI and the dawning of a new age
A dynamic approach to testing frontier AI model capabilities that supports innovation, incentivizes responsible behavior, and helps safely steward AGI.
