Certainly a two stage fitness process would work, or as mentioned using additional controls on the simulation to avoid idiosyncrasies.
Certainly a two stage fitness process would work, or as mentioned using additional controls on the simulation to avoid idiosyncrasies. It isn't that hard to build a small genetic 'evolver' (my favorites are some of the 'walking' ones which race blob things). But you can evolve all sorts of complex things if you are willing to wait long enough and don't care if you understand the result or not. I built one to evolve a response packet to ssh clients to kill them (I got tired of botnets trying to brute force passwords on a server I ran for a while), all you need is a fitness function, a mutate function, and time :-).
Posted by ChuckMcM on Jul 15, 2015, 1:26 AM
Tags
Related issues
- [Codex Desktop][Windows] Weekly quota jumped 67 points in 38 seconds (Oct 16, 2026; 18:10 UTC)
- I loathe how difficult Matrix is for normal people though.
- Fragile against upgrades, tons of unmaintained plugins, admin panel UX is a mess where you struggle to find the stuff your are looking fo...
- Unfortunately there doesn’t seem to be any decent alternative.
- I'm wondering how much of that is bad developer UX and defaults, and how much of that is inherent complexity in the problem space.
- Pulumi support
Raise real-world problems, find people facing the same issue, and discuss solutions.5,500 issues0 comments
ux (Sysadmin)
[Codex Desktop][Windows] Weekly quota jumped 67 points in 38 seconds (Oct 16, 2026; 18:10 UTC)
I loathe how difficult Matrix is for normal people though.
Fragile against upgrades, tons of unmaintained plugins, admin panel UX is a mess where you struggle to find the stuff your are looking fo...
Unfortunately there doesn’t seem to be any decent alternative.
I'm wondering how much of that is bad developer UX and defaults, and how much of that is inherent complexity in the problem space.
Pulumi support
[Feature]: managing cnpg in GitOps workflows (FluxCD, ArgoCD ...)
Merge the platform and UI threads
Certainly a two stage fitness process would work, or as mentioned using additional controls on the simulation to avoid idiosyncrasies.
Certainly a two stage fitness process would work, or as mentioned using additional controls on the simulation to avoid idiosyncrasies. It isn't that hard to build a small genetic 'evolver' (my favorites are some of the 'walking' ones which race blob things). But you can evolve all sorts of complex things if you are willing to wait long enough and don't care if you understand the result or not. I built one to evolve a response packet to ssh clients to kill them (I got tired of botnets trying to brute force passwords on a server I ran for a while), all you need is a fitness function, a mutate function, and time :-).
Comments (0)
No comments yet. Start the discussion.