Is life a form of computation?
3 hours ago
- #computation
- #AI
- #biology
- John von Neumann established a deep link between life and computation, showing that reproduction can be carried out by machines following coded instructions, similar to DNA.
- Biological computing is massively parallel, decentralized, and noisy, differing from traditional digital computing which relies on centralized, sequential execution of instructions.
- DNA operates like a program, with processes such as epigenetics and gene proximity effects adding complexity beyond simple digital computation.
- Randomness in biological computing is a feature, not a bug, and is also utilized in computer science algorithms, as seen in early computers designed by Turing.
- Modern AI relies on parallelism and randomness, exemplified by stochastic gradient descent and GPU parallelism, aligning more closely with biological computing models.
- Von Neumann and Turing explored non-centralized computing models, such as cellular automata and neural networks, which mimic biological systems.
- Cellular automata and neural cellular automata (NCAs) demonstrate how complex, lifelike behavior can emerge from simple, local rules, offering insights into the computational nature of life.
- The first successful emulation of von Neumann’s self-replicating automaton in 1994 highlighted the computational complexity and inefficiency of simulating parallel systems on serial computers.
- NCAs, combining neural nets with cellular automata, can model cell behaviors and regeneration, showing how local computations can produce global, lifelike patterns.