Lee Cronin: Controversial Nature Paper on Evolution of Life and Universe | Lex Fridman Podcast

The Nugget

Assembly Theory posits that the complexity of any object in the universe can be quantified by the minimum number of steps it took to create it, and objects created by an evolutionary process will have many copies. This theory aims to explain and quantify how selection and evolution give rise to complex phenomena like life, intelligence, and technology that are not predicted by fundamental laws of physics alone.

Make it stick

🧩 Minimum assembly steps needed to construct an object define its complexity
👥 More copies of a complex object imply it was produced by selection, not chance
🔬 Shortest path for object creation is favored as it persists best in an environment
🌍 Earth contains greatest known complexity due to biological and technological evolution

The assembly index is the minimum number of steps needed to construct an object from elementary building blocks. Objects with high assembly index and copy number were likely produced by selection and evolution, not random chance.
The "factory" that enables this selection is the environment applying constraints over time. Shortest path for object creation is favored as it survives and propagates best.
Assembly Theory could help identify artifacts of selection and evolution on other planets by measuring molecular assembly index and abundance.

Due to nearly 4 billion years of biological evolution and recent cultural and technological evolution by humans, Earth likely contains the most complex objects known.
Highly complex and abundant molecules like DNA, proteins, cells would be key signatures of life to search for on other planets.
Humans have an exceptional ability to generate novel ideas and technology, which increases Earth's total complexity over time in ways that may be uncommon in the universe.

The universe appears deterministic looking backwards but undetermined looking forwards, implying time is fundamental and not just a coordinate.
Selection acts over time to generate complex objects and novel configurations not predicted by initial conditions or laws of physics alone.
Human imagination and free will introduce novelty into the universe in ways that cannot be predicted even with perfect knowledge of the past. The future is "bigger" than the universe can contain.

Warnings of existential risk from superintelligent AI in the near-term are likely overblown. We are far from artificial general intelligence and current "AI" lacks true reasoning, knowledge generation, or agency.
However, AI will have increasing unintended consequences that could cause real harms if not developed thoughtfully. Authenticated data and regulation will be important.
To achieve AGI, we likely need new hardware architectures and methods beyond current deep learning that can flexibly integrate knowledge and reason more like humans.

"Every star in the sky probably has planets and life is probably emerging on these planets, but I think the combinatorial space associated to these planets is so different our causal cones are never going to overlap or not easily. This is a thing that makes me sad about alien life - why we have to create alien life in the lab as quickly as possible."
"I think the universe is deterministic looking back in the past, but undetermined going forward in the future. The universe appears deterministic going back in time and non-deterministic going forward in time because the future is too big, the universe is too big in the future to contain in the present."
"The nice thing about current machine learning technologies is they might actually help reveal why time is fundamental. They tell you about what's happened in the past but they can never help you understand what's happening in the future without training examples."
"Human beings have a brilliant ability to integrate across all domains all at once and to synthesize something which allows us to generate knowledge and becoming Gödel-complete on our own... What is reasoning? The way you can tell the difference between trial & error and the generation of new knowledge is you come up with a theory, an explanation - inspiration comes from outside - and then you test that and see it's going towards the truth. Human beings are very good at doing that."

This summary contains AI-generated information and may have important inaccuracies or omissions.