{"author_link":"\/users\/digaly","author_name":"digaly","author_uid":"digaly","comments":[],"epoch":1601801532,"event":"LD47","format":"md","ldjam_node_id":224639,"likes":4,"metadata":{"p_key":"148564","p_author":"digaly","p_authorkey":"1039513","p_urlkey":"365022","p_title":"Blobby blog","p_cat":"LDJam ","p_event":"LD47","p_time":"1601801532","p_likes":"4","p_comments":"0","p_status":"WAYBACK","us_key":"1039513","us_name":"digaly","us_username":"digaly","event_start":"1601596800","event_key":"76","event_name":"Ludum Dare 47"},"node":{"_collation":{"body_sanitizer":"TextUtils::SanitizeHTML via existing importer","event":"LD47","removed_author":false},"_superparent":212256,"_trust":1,"author":39513,"body":"![blobby.gif](\/\/\/raw\/95a\/9\/z\/351b2.gif)\nGot basic platforming done + shooting and a bunch of animation things.\n\nThe idea is that you play against an AI that gradually gets better against you.\nI want to do this through genetic algorithms but in a more efficient way that doesn't require 5000 generations.\nIt's honestly more of an experiment.\n\n\nOne way to do this is to be careful in the initial seed to provide vastly different starting points by generating one random program and providing 'opposites' to that program. This way you might always have a program that outperforms the other in your initial seed. From there on out it's the usual crossbreeding & mutation.\n\n\nNow it's time to work on some genetic algorithms :x\n\n\n","comments":0,"created":"2020-10-04T08:40:18Z","files":[],"files-timestamp":0,"id":224639,"love":4,"love-timestamp":"2020-10-04T13:21:44Z","meta":[],"modified":"2020-10-04T13:21:44Z","name":"Blobby blog","node-timestamp":"2020-10-04T08:53:02Z","parent":222636,"parents":[1,5,9,212256,222636],"path":"\/events\/ludum-dare\/47\/$222636\/blobby-blog","published":"2020-10-04T08:52:12Z","scope":"public","slug":"blobby-blog","subsubtype":"","subtype":"","type":"post","version":664034},"node_metadata":{"n_key":"224639","n_urlkey":"365022","n_parent":"222636","n_path":"\/events\/ludum-dare\/47\/$222636\/blobby-blog","n_slug":"blobby-blog","n_type":"post","n_subtype":"","n_subsubtype":"","n_author":"39513","n_created":"1601800818","n_modified":"1601817704","n_version":"664034","n_status":"WAYBACK"},"source_url":"https:\/\/ldjam.com\/events\/ludum-dare\/47\/$222636\/blobby-blog","text":"![blobby.gif](\/\/\/raw\/95a\/9\/z\/351b2.gif)\nGot basic platforming done + shooting and a bunch of animation things.\n\nThe idea is that you play against an AI that gradually gets better against you.\nI want to do this through genetic algorithms but in a more efficient way that doesn't require 5000 generations.\nIt's honestly more of an experiment.\n\n\nOne way to do this is to be careful in the initial seed to provide vastly different starting points by generating one random program and providing 'opposites' to that program. This way you might always have a program that outperforms the other in your initial seed. From there on out it's the usual crossbreeding & mutation.\n\n\nNow it's time to work on some genetic algorithms :x\n\n\n","title":"Blobby blog","wayback_source":[]}