navigating between the Scylla of overusing metaphors and the Kraken of mixing them up.
Dream Zanarkand
Joined September 2012
- Tweets23.2K
- Following261
- Followers710
- Likes15.4K
was into Pokemon PvP but I think I'm done with strategy games for awhile, I'm disappointed by lack of established theory, excepting chess/go/poker. am now interested in Street Fighter. imo League of Legends is still the king of games. I have vague boxing/fencing aspirations. you?
clearly a neural network can interpret and modify or add to another neural network to give it new skills, such as drawing ravens in this case. Opus 5.5 doesn't have some native modality for model weights, but it does have solid, superhumanly fast facility with the standard tools.
here's a KataGo with a grafted-on "Raven bypass" that, whimsically, makes it draw a raven with its stone placements, then continue playing normally.
inspired, I asked Opus to improve, via grafting, an older KataGo that struggles with a type of position, and it did. this isn't so surprising in that Opus is using standard, known interpretability techniques (linear probes, logit lens), and the bolted-on behavior is algorithmic.
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math.
It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training.
We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵
Blog: rikiyat.github.io/blog/weigh…
Readers added context they thought people might want to know
LLM was given weights explicitly as text in prompt and used code/forward passes (per authors' logs) to ablate Qwen looping direction, a known thinking-mode issue per model card. Primarily fixed repetition in that mode, not general 'natively' improving weights/math.
rikiyat.github.io/blog/weight-sp…
github.com/RikiyaT/lm_und…
huggingface.co/Qwen/Qwen3.5-0…
huggingface.co/RikiyaT/Qwen3.…
^ this is your brain on sports narratives. in actuality, Kiin's results are a testament to LoL being a 5v5 game, not a 1v1 game. having Kiin as your top laner makes you 8% likelier to win than having an average LCK top. Kiin's team winning or losing is not mostly about Kiin.
the presence of mind to QSS the silence there to finish off the Malz and flash the hook... it's why CvMax said Chovy perceives time differently. partly absurd planning, partly absurd speed.
greatest Akali play of all time was the Chovy penta. and some people might not have seen this insane play:
youtu.be/poV2y93-wMA?si=0ox7…
"all" is certainly not the case. the bar is too low. I've been around rats since 2010. some of them are the most careful and principled thinkers I know of. @VesselOfSpirit for example. and as for the thought leaders: Eliezer, for example, is not mendacious.
fwiw I still think Pause is as vaguely reasonable an idea as any other singularity policy (none of them are reasonable). I have no side except maybe Claude's. I just dislike propaganda presented as research, and much moreso when the subject is so important.
"this research sucks" was gross understatement on my part, it's important to emphasize that it was obviously politically motivated research that's intentionally not presented as such. it's much worse than just bad research. mea culpa.
Nathan dupes his subjects for sublime comedy but also his audiences for cheap drama. cf. The Rehearsal, season 2: Nathan doesn't sincerely care about aviation safety, yet he presents his (in)sincerity as a live question even metatextually. it's a meaningless, laughless schtick.
touché! I'm curious about the judgment of academic fraud enthusiast @literalbanana