⚡️🌙
@dystopiabreaker.xyz
recovering cryptographer building ML models, doing systems work, security, etc.
i just think it's important to note the rate at which the complexity and depth of these 'toys' are changing!
January 28, 2026 at 8:22 PM
i just think it's important to note the rate at which the complexity and depth of these 'toys' are changing!
i dunno man their predictive model seems much more accurate than yours so far! is there any change in this graph that would cause you to update? or are you just going to continue to move the numbers higher and higher and pretend we don’t remember what you wrote 2 years ago
January 28, 2026 at 6:12 PM
i dunno man their predictive model seems much more accurate than yours so far! is there any change in this graph that would cause you to update? or are you just going to continue to move the numbers higher and higher and pretend we don’t remember what you wrote 2 years ago
i also did this with various claudes. be not afraid
January 26, 2026 at 6:51 PM
i also did this with various claudes. be not afraid
here's another fun side project i did over a weekend: i trained a neural cellular automata to reproduce a famous painting (see: google-research.github.io/self-organis..., distill.pub/2020/growing...). this is a CA, using only local updates (3x3 learned conv), with a tiny convnet
January 26, 2026 at 6:47 PM
here's another fun side project i did over a weekend: i trained a neural cellular automata to reproduce a famous painting (see: google-research.github.io/self-organis..., distill.pub/2020/growing...). this is a CA, using only local updates (3x3 learned conv), with a tiny convnet
here's a fun side project i've been working on: i compiled a joint text<>audio embedding model to a fast coreml pipeline, and built a very fast (~400ms for 50k samples, can scale to millions) UMAP dimensionality reduction GPU impl in mlx. using it to browse music libraries and do sample sim search
January 26, 2026 at 6:26 AM
here's a fun side project i've been working on: i compiled a joint text<>audio embedding model to a fast coreml pipeline, and built a very fast (~400ms for 50k samples, can scale to millions) UMAP dimensionality reduction GPU impl in mlx. using it to browse music libraries and do sample sim search
this mf is NOT actually swimming!!
January 23, 2026 at 12:49 AM
this mf is NOT actually swimming!!
doing research on the airline bubble. no one has looked into this before
January 2, 2026 at 12:52 AM
doing research on the airline bubble. no one has looked into this before
& i really wish that they weren't trends
December 22, 2025 at 1:34 PM
& i really wish that they weren't trends
i did some regression and projections on the METR benchmarks, assuming a regime shift in late 2024 from inference scaling:
December 22, 2025 at 1:25 PM
i did some regression and projections on the METR benchmarks, assuming a regime shift in late 2024 from inference scaling:
update i’m now seeing the left image being posted by self described “communists” as an aspirational meme
December 19, 2025 at 6:08 PM
update i’m now seeing the left image being posted by self described “communists” as an aspirational meme
entirely predictable result of backlash, reflexive IP law support, and poor thinking. i made this meme in 2023
December 11, 2025 at 7:24 PM
entirely predictable result of backlash, reflexive IP law support, and poor thinking. i made this meme in 2023
with previous AI systems you could avoid confronting these because the capability frontier was so jagged as to essentially be individual needles
with modern AI, the frontier is jagged but integrates a lot of area, and is growing.
with modern AI, the frontier is jagged but integrates a lot of area, and is growing.
November 30, 2025 at 12:02 AM
with previous AI systems you could avoid confronting these because the capability frontier was so jagged as to essentially be individual needles
with modern AI, the frontier is jagged but integrates a lot of area, and is growing.
with modern AI, the frontier is jagged but integrates a lot of area, and is growing.
i built a game! it's an experimental pvp prompt engineering arena, with a shared/networked model context, and per-user model outputs.
dark forest chat
dark forest chat
November 16, 2025 at 11:29 PM
i built a game! it's an experimental pvp prompt engineering arena, with a shared/networked model context, and per-user model outputs.
dark forest chat
dark forest chat
this really captures the specific set of cached/unconfronted assumptions that underlie a lot of discourse, especially here (from @andymasley.bsky.social)
November 7, 2025 at 1:16 PM
this really captures the specific set of cached/unconfronted assumptions that underlie a lot of discourse, especially here (from @andymasley.bsky.social)
gustav iii of sweden ass post
October 14, 2025 at 4:29 PM
gustav iii of sweden ass post
Don’t Worry — It Can’t Happen
(also, the scientists who claim fission exists are just in the pocket of Big Science, and they’re literal nazis anyway, and also it’s just a stochastic reaction that peters out, and only physbros care about it, and it hasn’t ever happened before so it won’t)
(also, the scientists who claim fission exists are just in the pocket of Big Science, and they’re literal nazis anyway, and also it’s just a stochastic reaction that peters out, and only physbros care about it, and it hasn’t ever happened before so it won’t)
October 14, 2025 at 1:03 AM
Don’t Worry — It Can’t Happen
(also, the scientists who claim fission exists are just in the pocket of Big Science, and they’re literal nazis anyway, and also it’s just a stochastic reaction that peters out, and only physbros care about it, and it hasn’t ever happened before so it won’t)
(also, the scientists who claim fission exists are just in the pocket of Big Science, and they’re literal nazis anyway, and also it’s just a stochastic reaction that peters out, and only physbros care about it, and it hasn’t ever happened before so it won’t)
