@tillomnl: Tales of Kitsune: The Final Chapter 🦊 The whys and the hows of our XL brushes! 🏮

tillo.mnl
tillo.mnl
Open In TikTok:
Region: PH
Monday 03 August 2026 10:25:00 GMT
4839
175
11
5

Music

Download

Comments

_yanalovesyouuu
_yana :
2 minutes makeup challenge using this brush ma hahaha 😁
2026-08-03 10:52:27
3
jujumargarita_
jujumargarita_ 🫒 :
8.8 sale pls😭😭😭
2026-08-03 20:09:26
0
cheezemix_
micmicnimacmac :
OMGGGGG
2026-08-03 11:09:50
0
shikiseimei_
shikiseimei :
we have to see it in pink
2026-08-03 10:51:42
1
allysagonzales.ag
ALLIE ⚡️ | No To Pax Silica :
literal na xl 😭 ILOVETTE
2026-08-03 10:33:03
2
.lota
Ig: Alotof_Lota :
SALEEE PLS MOWMM
2026-08-04 13:19:40
0
noellepan_
noey (no-wi) ✨ :
Ma baka pwede may pink nito? 🥹💕
2026-08-04 13:06:35
0
kuting_md
Julliene Lorenz :
restockkk plss
2026-08-04 05:37:50
0
btsbp.74
btsbp.74 :
wow hindi kumakain ng product love it..
2026-08-04 05:52:48
0
ms.kiyoko
Ms. Kiyoko :
maa may video ka ba na ginagamit si T16 sa base? planning to buy huhuhu
2026-08-04 02:02:33
0
rue.alaiah
Rue 🐰✨ :
Maaaa inaantay ko ang pink beke naman
2026-08-03 13:06:21
0
To see more videos from user @tillomnl, please go to the Tikwm homepage.

Other Videos

AI is getting to the point where you can build a surprisingly complete game without manually directing every step. But the secret is not simply asking an AI chatbot to “make me a game.” The method is called loop prompting, or what creator Matt Shumer calls a “Gauntlet Loop.” Instead of generating one version and stopping, the AI repeatedly builds, tests, reviews, and improves its work. To do this, you need an AI coding agent like Codex or Claude Code. Unlike a normal chat, these agents can edit files, run the game, inspect the result, use development tools, and create additional agents for different parts of the project. The process works like this: - You give the lead agent a clear goal, such as building a kart racing game, but avoid prescribing every technical decision. This gives the agent room to determine the best approach. -Then you give it a real quality bar. Instead of saying, “make it look amazing,” you provide an existing game, screenshots, gameplay footage, or another concrete reference it can compare its work against. -Next, the lead agent breaks the game into smaller pieces. The driving, tracks, lighting, environments, menus, sound, and effects can each have their own builder agent. Most importantly, the builder does not review its own work. A separate critic agent inspects the result, compares it against the reference, identifies the biggest weakness, and sends it back to the builder. The builder fixes it, the critic reviews it again, and the loop continues. That separation matters because an AI that created something is often better at explaining its decisions than objectively judging whether the result is good. The basic formula is: Set the goal.
Give it a real example of quality.
Split the project into smaller parts.
Use separate builders and critics.
Keep looping until the result reaches your standard. This method can be used for more than games. The same idea can be applied to websites, software, product design, writing, research, and other work that can be inspected and improved. I linked the full Gauntlet Loop guide in my bio, including the prompt you can adapt to start building your own game with AI.
AI is getting to the point where you can build a surprisingly complete game without manually directing every step. But the secret is not simply asking an AI chatbot to “make me a game.” The method is called loop prompting, or what creator Matt Shumer calls a “Gauntlet Loop.” Instead of generating one version and stopping, the AI repeatedly builds, tests, reviews, and improves its work. To do this, you need an AI coding agent like Codex or Claude Code. Unlike a normal chat, these agents can edit files, run the game, inspect the result, use development tools, and create additional agents for different parts of the project. The process works like this: - You give the lead agent a clear goal, such as building a kart racing game, but avoid prescribing every technical decision. This gives the agent room to determine the best approach. -Then you give it a real quality bar. Instead of saying, “make it look amazing,” you provide an existing game, screenshots, gameplay footage, or another concrete reference it can compare its work against. -Next, the lead agent breaks the game into smaller pieces. The driving, tracks, lighting, environments, menus, sound, and effects can each have their own builder agent. Most importantly, the builder does not review its own work. A separate critic agent inspects the result, compares it against the reference, identifies the biggest weakness, and sends it back to the builder. The builder fixes it, the critic reviews it again, and the loop continues. That separation matters because an AI that created something is often better at explaining its decisions than objectively judging whether the result is good. The basic formula is: Set the goal.
Give it a real example of quality.
Split the project into smaller parts.
Use separate builders and critics.
Keep looping until the result reaches your standard. This method can be used for more than games. The same idea can be applied to websites, software, product design, writing, research, and other work that can be inspected and improved. I linked the full Gauntlet Loop guide in my bio, including the prompt you can adapt to start building your own game with AI.

About