its crazy that its implementing stuff faster than i can playtest it. its implementing my ideas as fast as i can think of them. i just keep adding issues to my autowork epic and things just get done as it churns through them.
jb55's avatar jb55
this engine is also being built from the ground up so that agents can drive it and playtest it. the "build games in a gui" era is so over View quoted note → image
View quoted note →

Replies (8)

Default avatar
nupe 1 month ago
Now add an idea guy agent into the mix. Then another one and another. Make one pretend to be human and the others think it is. Then let them vote on ideas.
Hard part is always testing. Agents will tell you it's properly tested when it's all theater. You'll have thousands of tests and just hope that some solid percentage of them are real. Hopefully the new testing integration in the game harness will solve that.
autowork is a workflow/skill i made that recursively calls my /handoff skill. handoff is basically just letting an agent spawn a new session with the initial prompt
Ever since I was a kid, I've dreamed of making a game, but gameplay has always been the most important thing to me. At some point, I simply came to terms with the fact that I can't create a genuinely good game in terms of game mechanics - not in terms of coding. Roughly speaking, I'm not going to make Tetris. So AI doesn't really solve my "game development" problem at all.
The vibe coding has gotten so good if you are crafty. I built out an entire technocracy resistant building methodology while spinning out 50 application layers I’ve been dreaming about my whole life. However, It is a struggle when your work outruns your ability to effectively apply it IRL.
↑