diff --git a/good_ideas.myco b/good_ideas.myco
index eeb4170..6294243 100644
--- a/good_ideas.myco
+++ b/good_ideas.myco
@@ -73,10 +73,15 @@ Diffusion model for Minecraft worlds.
 * Make a high-performance physically plausible space combat simulator and do RL to spite Devon Eriksen.
 * Simulate self-consistent time travel in games by having absurdly superhuman AI player enforce consistency using "random" degrees of freedom.
 * Nonstandard/intense experiences probably increase memory formation/recollection later. As such, coding bootcamps but they are actual bootcamps. This has the additional advantage of maybe getting more hard work out of people.
-* Psychological studies using LLMs (https://arxiv.org/abs/2209.06899 addresses this) seem underexploited. Could factor-analyze political beliefs and such.
+* {
+Psychological studies using LLMs (https://arxiv.org/abs/2209.06899 addresses this) seem underexploited. Could factor-analyze political beliefs and such.
+* https://arxiv.org/abs/2410.20268
+}
 * Hexbugs with inductive charging, onboard microcontroller.
 * "you can go hunting for wallets that are big, cold for a long time, and likely made with hardware that now has known vulns in its rng"
 * {
 https://arxiv.org/abs/2405.16158 improves (model-free?) RL through using techniques applied to SSL to make scaling work better. Does it work on MCTS-based things?
 * The model architectures in the Go MCTS papers look like they used arbitrary round numbers, actually. Did //anyone// do scaling laws sweeps?
-}
\ No newline at end of file
+}
+* 6D Tic-Tac-Toe With Multiversal Time Travel (5DCWMTT does not actually have five dimensions!).
+* National Rail frontend but actually useable and complete.
\ No newline at end of file