osmarks/documentation
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Edit ‘good_ideas’: Synced 1767227050299

Browse Source
This commit is contained in:
sync
2026-01-01 00:28:05 +00:00
committed by wikimind
parent 70578b6564
commit 939cc827cd
1 changed files with 11 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
good_ideas.myco
View File

@@ -38,7 +38,10 @@ Next-action-predictor editors/UI.
}
* {
Do WiFi sensing but good (with more data (https://b.osmarks.net/o/a0bddc9b742b4efbba18600ae6d51d98)).
* Weak evidence from internal testing suggests that models can't easily "ICL" locations. We may need really long contexts or big finetunes or some smarter meta-learning algorithm.
* {
Weak evidence from internal testing suggests that models can't easily "ICL" locations. We may need really long contexts or big finetunes or some smarter meta-learning algorithm.
* (multi-TX single-RX data with ESP32-C3 receiver, slightly nonuniform times; would need to run more tests)
}
}
* {
Comparisons rather than scalar ratings for things (https://b.osmarks.net/o/60b26f1735134d628164217be52ca2d3).
@@ -72,6 +75,7 @@ Diffusion model for Minecraft worlds.
* {
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
* TODO find the paper which actually did factor-analyze political beliefs (while doing something else)
}
* 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"
@@ -93,7 +97,11 @@ Semantic search for:
}
* Automate Lego with pick-and-place machine.
* Improve AI-based search with methods other than making boring contrastive embedding models very slightly better.
* Graph-based vector indices do beam search on a graph constructed so that beam search on dot product or PQ dot product "mostly" returns the closest (highest-dot-product) result. Do one of several RL/etc pathfinding schemes instead (like https://arxiv.org/abs/2502.18663)? We only have a budget of 100μs per read node, though, so this is a bit tricky.
* {
Graph-based vector indices do beam search on a graph constructed so that beam search on dot product or PQ dot product "mostly" returns the closest (highest-dot-product) result. Do one of several RL/etc pathfinding schemes instead (like https://arxiv.org/abs/2502.18663)? We only have a budget of 100μs per read node, though, so this is a bit tricky.
* We can also pack the graphs better (with some cost at retrieval time due to multi-page accesses and more compute): sorting and delta compression of node indices, offload metadata again, maybe general-purpose compressors, maybe QAT/IAA codecs, quantize the graph vectors a bit.
* 100μs/node is assuming SSD random latency, and with that assumption these algorithms still aren't useful if we can throw more time at it because we can do more reads. But with really clever pathfinding, HDDs could also work. Maybe? Two orders of magnitude slower, though.
}
* SAEs/dictionary learning algorithms for residual quantization.
* Minecraft/Factorio crossover modpack via Clusterio-like thing.
* With modern sensor technology, why can I not buy a toaster even smarter than the Sunbeam Radiant Control Toaster?
@@ -102,4 +110,4 @@ Semantic search for:
* Cookie Clicker, but with victories rather than cookies.
* "idea: autonomous, agentic emails that hunt you down and destroy you" (https://x.com/0x49fa98/status/1993690386531860574)
* Neural net graph prettiness heuristic rather than force-based.
* CA embedding models.
* CA embedding models (like maths thing, transition rule proximity).