YouTube2Text

Building an indie colony sim where your colonists run on your scripts — Transcript

by Messy Development · 1,303 words · 188 segments · language en · Watch on YouTube

Full transcript

  1. 0:00Hi, and welcome to Far Camp. This is a
  2. 0:02colony simulator I've been developing
  3. 0:04for a few weeks, and in this
  4. 0:06introduction video
  5. 0:08I'll be talking a bit about how I made
  6. 0:10it.
  7. 0:12By now, I'm sure everyone's seen a video
  8. 0:14game developed by AI, and a lot of these
  9. 0:16are really good-looking games, but most
  10. 0:19of them are one-shot development.
  11. 0:21Meaning a single prompt generates the
  12. 0:23entire thing, and that's not what I'm
  13. 0:25going to show you today. Today, I'm
  14. 0:26going to show you how I built a colony
  15. 0:28simulator using AI development platforms
  16. 0:31like Claude Code, Codex, and my own,
  17. 0:34which I've been calling OI, as well as a
  18. 0:36bunch of tools and services that I've
  19. 0:39been building as open systems, which are
  20. 0:41going to be open source and free to use,
  21. 0:43and eventually community-driven. What
  22. 0:45you're seeing here is the current build
  23. 0:47of the colony simulator, which I'm
  24. 0:49calling Far Camp. I've been solo
  25. 0:51developing this with a team of agents
  26. 0:53doing everything from writing the code
  27. 0:56to generating the assets to creating
  28. 0:58procedural generation techniques for the
  29. 1:01world, for audio, and at the end of the
  30. 1:04series, I'll be sharing all of my
  31. 1:05workflows with you for free in hopes
  32. 1:07that some people find them useful, and
  33. 1:09they jump on and help me community
  34. 1:11develop them. The game started as a
  35. 1:13proof of concept last year.
  36. 1:14It was very basic. The art assets
  37. 1:17weren't really there.
  38. 1:18Uh the main proof of concept though is
  39. 1:20that I wanted to have natural language
  40. 1:23driving the routines of the colonists,
  41. 1:25and that part worked really well. But, I
  42. 1:28didn't have time to do a full game
  43. 1:30development effort, where I designed the
  44. 1:32art assets. I was doing a lot of other
  45. 1:34work full-time, and this was just a
  46. 1:37hobby. So, I tried to use the AI models
  47. 1:39back then to help me generate assets,
  48. 1:41but they were sub par, and I kind of
  49. 1:44just put it on a shelf. Fast forward to
  50. 1:46today, and we have some amazing models
  51. 1:49like Astra, Fable, Opus, and their
  52. 1:52understanding of geometry is decades
  53. 1:55ahead of what it was last year. Very
  54. 1:58quickly, this time around I was able to
  55. 2:00generate some really nice-looking
  56. 2:02stylized assets. And that's how this
  57. 2:04project restarted. So, from that I then
  58. 2:06focused on some procedural generation
  59. 2:08routines. And so now every world that is
  60. 2:11generated creates its own brand new
  61. 2:13plants and creatures. But I thought,
  62. 2:15"Hey, it would be nice to have something
  63. 2:17familiar also." So, out of 32 species,
  64. 2:21and that's just the number I picked for
  65. 2:22now, half of them are reused from
  66. 2:25previous playthroughs. So, you might
  67. 2:26recognize something familiar and have an
  68. 2:28advantage over your last playthrough.
  69. 2:31So, you might see a blueberry bush and
  70. 2:33you're pretty certain that's the same
  71. 2:35one you played with last playthrough.
  72. 2:38And so, you could take the chance and
  73. 2:39just eat it without first observing it,
  74. 2:42but there's also a chance your memory
  75. 2:43doesn't serve you very well and it is
  76. 2:45poisonous. So, pretty much everything in
  77. 2:47a new world has to be studied and
  78. 2:49observed before you can interact with
  79. 2:51it, at least safely. And if you're
  80. 2:53unlucky and the colonists you got rolled
  81. 2:55have low intelligence, well, you got to
  82. 2:57build that up for some of the things you
  83. 2:58need to observe. But after that initial
  84. 3:01proof of concept with the art, I wanted
  85. 3:03to switch to making sure this game
  86. 3:05concept was fun. And to make it stand
  87. 3:08out from just a regular colony simulator
  88. 3:10because there's a lot of those already.
  89. 3:12So, the highest priority item was to
  90. 3:14bring back my old scripting system,
  91. 3:17which would allow me to fully customize
  92. 3:19the routines of the colonists. So,
  93. 3:21unlike RimWorld where you have set
  94. 3:22behaviors and you just have to set those
  95. 3:24up with mouse clicks. Here you can
  96. 3:26actually type in script and have
  97. 3:29colonists do nearly anything. Well,
  98. 3:31anything that's allowed in the game. But
  99. 3:33I didn't expect players to learn a new
  100. 3:35scripting language just to play the
  101. 3:36game. And so, what I wanted to prototype
  102. 3:39really quickly was the ability for a
  103. 3:42player to use their own AI agent,
  104. 3:45whether it's Claude code or whether it's
  105. 3:48ChatGPT in a browser, as As as it has
  106. 3:50MCP access, which is just a connection
  107. 3:53protocol to other systems, as long as it
  108. 3:56has that, and most of them do, it can
  109. 3:58connect to the game, and you can
  110. 4:00basically tell your agent, "I want my
  111. 4:02colonist to collect up to 20 wood
  112. 4:06whenever the stockpile goes under 10."
  113. 4:09And then it'll generate the script for
  114. 4:11you in the game. And so, in a way, it
  115. 4:14kind of teaches you scripting, and this
  116. 4:16script is very similar to Python. So, in
  117. 4:19a way, you're kind of learning Python as
  118. 4:22you learn to play this game. Now, all of
  119. 4:24this is optional. You don't have to
  120. 4:26learn any of this custom routine
  121. 4:29building in order to play the game. We
  122. 4:31can play it just like RimWorld, where we
  123. 4:33mouse click every routine we want. But,
  124. 4:36if you really want to get customized,
  125. 4:38this is the thing I wanted to prototype.
  126. 4:40And it worked just as well as in my
  127. 4:42initial prototype from a year ago, but
  128. 4:44now we have it in Godot, and we have it
  129. 4:46with much better graphics. And the
  130. 4:48really strong feature from this is
  131. 4:50replayability. That every time the game
  132. 4:52ends, or you lose, or something happens,
  133. 4:55you get to keep all of your scripts, and
  134. 4:57you can modify them. And so, when you
  135. 4:59start again, you don't have to start
  136. 5:00from scratch. You can already start your
  137. 5:03colonists with your previous routine, so
  138. 5:05you have a leg up from the last
  139. 5:06playthrough. I think it's a really
  140. 5:07powerful system, and I haven't seen this
  141. 5:10in many games, or in many colony sim
  142. 5:12games. And so, once those core concepts
  143. 5:15were proven working well, it was time to
  144. 5:18just get started on a very simple
  145. 5:20playthrough.
  146. 5:21And for that, I just wanted some basic
  147. 5:22things, a couple beds, a fireplace, some
  148. 5:26basic routines, like feeding and
  149. 5:28hydrating themselves. And so, we needed
  150. 5:30water. We had the original plants, and
  151. 5:32we had animals, but no hunting yet. And
  152. 5:35so, on our first playthrough, we were
  153. 5:36able to build a very humble camp with
  154. 5:38just the basics. Eventually, this would
  155. 5:40lead to cozy sound effects of the fire
  156. 5:44crackling, the nighttime crickets, but
  157. 5:46that's something for later, and we'll
  158. 5:47get into that in some later episodes and
  159. 5:50how I did that procedurally also. We
  160. 5:51very quickly rolled out some basic
  161. 5:54weather so that it gets cold at night,
  162. 5:56and so they needed to stay warm, and
  163. 5:58while the fire helped, they also could
  164. 6:00eventually build towards a shelter. And
  165. 6:03so that's the next part we added to
  166. 6:05this, and we were now able to survive
  167. 6:07through the elements for a couple days.
  168. 6:09We had food, we had water, we had
  169. 6:11shelter, heat, and we could basically go
  170. 6:14on like this forever because food
  171. 6:16replenished itself, and so did water.
  172. 6:18And so that became a problem. What is
  173. 6:21the first stress we add for the players?
  174. 6:24What's the first challenge? What do they
  175. 6:26need to build towards to survive? And so
  176. 6:28I wanted something a little bit creepy
  177. 6:30that possibly happens at night.
  178. 6:32Something where you see eyes circling
  179. 6:35your camp in the dark, and you might
  180. 6:37hear howls, but that's for another
  181. 6:40episode.
  182. 6:41That's all I wanted to cover for now. I
  183. 6:43hope you think this might be fun. In the
  184. 6:45description, I've left a link for how
  185. 6:48you can join in the development as well,
  186. 6:50contributing ideas. Like I mentioned,
  187. 6:52everything will be open source. That's
  188. 6:54all for now, and thanks for watching.

About this transcript

This page contains the full transcript of Building an indie colony sim where your colonists run on your scripts by Messy Development, generated from the public captions YouTube serves with the video. The transcript has 1,303 words across 188 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.