links
bookmarks, resources, and interesting finds
anita's garden
anita.garden
prime intellect autonomous nanogpt speedrun
github.com
how to achieve truly serverless gpus
modal.com
learning beyond gradients
trinkle23897.github.io
the world inside neural networks: how neural geometry will unlock understanding and control of ai
goodfire.ai
agent score
mintlify.com
schizoposting
counterfeitsunset.neocities.org
american diner gothic
thenewatlantis.com
a tinyblog about tinygrad
tinyblog-phi.vercel.app
neural garbage collection: learning to forget while learning to reason
arxiv.org
what we learnt from letting ai posttrain ai
thoughtfullab.com
the many faces of claude
eriskii.net
free media heck yeah
fmhy.net
pi - a minimal coding agent harness
pi.dev
open weights isn't open training
workshoplabs.ai
paperclip - open-source orchestration for zero-human companies
paperclip.ing
arscontexta - memory tools for claude
arscontexta.org
statement from dario amodei on our discussions with the department of war
anthropic.com
the markets aren't alright
benjaminfspector.com
torph - dependency-free text morphing
torph.lochie.me
ACE-STEP-1.5: the most powerful open-source music model
github.com
transformer lab
lab.cloud
ai research skills
github.com
sometimes all functions are continuous
math.andrej.com
deriving manifold constrained hyper connections
k-a.in
a technical deep-dive on moondream2
gjyotin305.github.io
paper shaders - ultra-fast zero-dependency shaders
shaders.paper.design
build a simple deep learning library
zekcrates.quarto.pub
dan weng's 2025 letter
danwang.co
a guide to claude code 2.0
sankalp.bearblog.dev
palladium magazine: governance futurism
palladiummag.com
balance of power
vitalik.eth.limo
2025 letter
zhengdongwang.com
fairy2i: training complex llms from real llms with all parameters in {±1, ±i}
arxiv.org
a call for new aesthetics
newaesthetics.art
universally converging representations of matter across scientific foundation models
alphaxiv.org
large causal models from large language models
arxiv.org
claude's christmas
claude.christmas
the prism hypothesis: harmonizing semantic and pixel representations via unified autoencoding
alphaxiv.org
the techno-humanist manifesto
newsletter.rootsofprogress.org
five years of tinygrad
geohot.github.io
explained: latent consistency models
naklecha.notion.site
capital in the 22nd century
philiptrammell.substack.com
static allocation with zig
nickmonad.blog
the best things and stuff of 2025
blog.fogus.me
openai frontierscience dataset
huggingface.co
inside vllm: anatomy of a high-throughput llm inference system
aleksagordic.com
neocambria: humans in a post-asi world
aleksagordic.com
jax-js - ml for the web
jax-js.com
shallow review of technical ai safety, 2025
shallowreview.ai