vicgalle_

A small update here: another replication using a different CPU, the 6502. In this case, replicators emerge 100 times less frequently than in Z80, due to a lacking LDIR-like instruction

See replies at this other comment https://news.ycombinator.com/item?id=48963513

> Do you have any idea why the authors chose Z80 as the program language?

In that previous paper you cite (by the same group) they tested several substrates for spontaneous replication (BFF, Forth, SUBLEQ, and emulated real CPUs. Z80 and 8080 CPU exps confirmed the rise of self-replicators, with the Z80 notably exhibiting multiple waves of increasingly capable self-replicators.

And the instruction set seems quite appropriate for the experiment: Z80 has native block-copying instructions, while it doesn't have MUL, so the task of evaluating polynomials is somewhat more challenging

Thanks for posting this, indeed that group at google has produced some interesting research in this space. That paper was also discussed here back in the day: https://news.ycombinator.com/item?id=40820022

An independent reproduction of the main result: https://github.com/vicgalle/coevolution-soup

I enjoy creating new benchmarks for LLMs. Lately, combining scientific computing tasks (n-body sim, Monte Carlo, etc) with Apple Metal GPU kernels (evolved through LLMs) led to a curious benchmark I believe: https://github.com/vicgalle/metal-sci-kernels

> https://twitter.com/minimaxir/status/1749893683137454194

Not very promising, though.