Skip to content

Normal Science

Brain healing
GraphAuthors

A reading list for frontier science

Articles across AI, biotech, forecasting, and emerging tech.

Recommendation GraphExplore who recommends whom across the networkBrowse AuthorsProfiles, influences, and key works
Weekly Digest — Free
Join researchers, founders, and analysts · Unsubscribe anytime

Categories

AllAIForecastingBioTechSecurity / OSINTMetascienceFinanceAI SafetyEnergyManufacturingCryptoStartups

Time

Sort

Today

I had AI read my entire field, Theoretical Astrophysics

Anon84·8h ago3pts

158,059 astrophysics papers mapped in 3D, 659 of the most important checked by AI by recomputing them, and every verified error followed through the papers that cite it.

jingyaogong/minimind

·GitHub Trending·25m ago

🧠 Train a 64M-parameter LLM from scratch in just 2h!https://jingyaogong.github.io/minimind "大道至简" 中文 | English 此开源项目旨在完全从 0 开始,仅用 3 块钱成本与 2 小时训练时间,即可训练出规模约为 64M 的超小语言模型 MiniMind。 MiniMind 系列极其轻量,主线最小版本体积约为 GPT-3 的 $\frac{1}{2700}$,力求让普通个人 GPU 也能快速完成训练与复现。 项目同时开源了大模型的极简结构与完整训练链路,覆盖 MoE、数据清洗、预训练(Pretrain)、监督微调(SFT)、LoRA、RLHF(DPO)、RLAIF(PPO / GRPO / CISPO)、Tool Use、Agentic RL、自适应思考与模型蒸馏等全过程代码。 MiniMind 同时拓展了视觉模态模型 MiniMind-V、多模态 Omni 模型 MiniMind-O、扩散语言模型(MiniMind-dLM)、线性模型(MiniMind-Linear),详见 Disc...

google-research/paper-orchestra

·GitHub Trending·25m ago

PaperOrchestra is a multi-agent framework for automated AI research paper writing.https://yiwen-song.github.io/paper_orchestra/ 🎻 PaperOrchestra: A Multi-Agent Framework for Automated AI Research Paper Writing Yiwen Song 1, Yale Song 1, Tomas Pfister 1, and Jinsung Yoon 1 1Google Cloud AI Research PaperOrchestra is a multi-agent framework for automated AI research paper writing. It flexibly transforms unconstrained pre-writing materials (ideas and experimental logs) into submission-ready LaTeX ...

ChromeDevTools/chrome-devtools-mcp

·GitHub Trending·11h ago

Chrome DevTools for coding agentshttps://developer.chrome.com/docs/devtools/agentsChrome DevTools for agents Chrome DevTools for agents (chrome-devtools-mcp) lets your coding agent (such as Antigravity, Claude, Cursor or Copilot) control and inspect a live Chrome browser. It is available as a Model-Context-Protocol (MCP) server and as a CLI. Both give your AI coding assistant access to the full power of Chrome DevTools for reliable automation, in-depth debugging, and performance analysis. Tool r...

Yesterday

ggml-org/ggml

·GitHub Trending·1d ago

Tensor library for machine learninghttps://github.com/ggml-orgggml Tensor library for machine learning Quick start Build from source: git clone https://github.com/ggml-org/ggml cd ggml mkdir build && cd build cmake .. cmake --build . --config Release -j 8 For a minimal, fully commented example (matrix multiplication), see examples/simple. Description The main goal of ggml is to be a simple, portable, and efficient tensor library for machine learning with minimal setup. Plain C/C++ implementation...

This Week

Agentic AI Designs an Analog Synthesizer

oidar·2d ago4pts

An analog monophonic synthesizer built from ~200 components, with a transistor ladder filter and touch keys. Circuit, simulation, PCB, firmware and most of the bring-up were done by Claude Opus 5.5.

Paul Christiano – We could be 12 months away from AIs that can overthrow the government

Dwarkesh Patel·Dwarkesh Patel·2d ago

Paul Christiano is head of the Alignment Research Center. Previously, he was head of AI safety at the US AI Safety Institute. Before that, he led alignment at OpenAI, where he invented RLHF. He’s recently joined OpenAI’s board (though here he is speaking as a private citizen).Paul thinks that on the current trajectory (absent any slowing down), there is a 1 in 3 chance that we could have superintelligence powerful enough to overthrow the government within a year. (He articulates a concrete scena...

Import AI 481: AI sub-contractors; a wet lab for AI; and subterfuge during development

Jack Clark·Import AI·2d ago

Welcome to Import AI, a newsletter about AI research. Import AI runs on arXiv, cappuccinos, and feedback from readers. If you’d like to support this, please subscribe.Subscribe nowAI systems are getting good at sub-contracting to make themselves better:…AI can now productively spend money to improve itself…Researchers have developed FrontierML, a benchmark “that measures how well agents turn capital into model performance on challenging, end-to-end machine learning tasks.” It shows that today’s ...

ashhart/TensorFold

·GitHub Trending·2d ago

LLM Inference Engine for Metal, CUDA and Vulkan.https://tensorfold.devTensorFold TensorFold serves language models from a Zig binary on Apple Silicon and NVIDIA GPUs. The engine reads checkpoints, tokenizes requests and runs Metal or CUDA kernels directly. Serving needs no Python or MLX installation. Building with a coding agent? Have it load the contributor skill before it writes any code. Contributing with agents Please use Agent Traffic Control before editing TensorFold so contributors and ag...

Split-and-Scale Latent 3D Representations

·Apple ML Research·2d ago

Latent 3D representations have significantly expanded the capabilities of modern shape modeling, enabling compact encoding of 3D scenes and supporting powerful 3D generative models. Despite this progress, current representations struggle to capture the fine details of complex objects, fundamentally limited by the capacity of their latents. However, scaling latent capacity at training time is computationally expensive and often infeasible under memory limits, and naively increasing the number of ...

Older

OpenAI’s o1 correctly diagnosed 67% of ER patients vs. 50-55% by triage doctors

donsupreme·6mo ago470pts

Researchers say results mark a really ‘profound change in technology that will reshape medicine’

Accelerating Gemma 4: faster inference with multi-token prediction drafters

amrrs·6mo ago640pts

An overview of how Multi-Token Prediction (MTP) drafters are making Gemma 4 models up to 3x faster at inference.

A couple million lines of Haskell: Production engineering at Mercury

unignorant·6mo ago399pts

What it takes to run 2 million lines of Haskell in production at a fintech company serving 300,000 businesses.

Using “underdrawings” for accurate text and numbers

samcollins·6mo ago359pts

A technique for accurate text and numbers in AI-generated images: generate the layout deterministically, then ask the image model to paint on top.

ProgramBench: Can language models rebuild programs from scratch?

jonbaer·6mo ago129pts

Abstract page for arXiv paper 2605.03546: ProgramBench: Can Language Models Rebuild Programs From Scratch?

ZAYA1-8B matches DeepSeek-R1 on math with less than 1B active parameters

steveharing1·6mo ago87pts

Who should care If you work with math, science problems, or complex coding tasks and you're looking for something small enough to run locally or cheaply via API, this is worth serious evaluation. The benchmark numbers at 760M active parameters are not normal and the Markovian RSA boost means performance scales with compute budget rather than hitting a fixed ceiling. If you're building agent workflows that need reliable tool calling or multi-step instruction following, look elsewhere fo

Show HN: Apple's SHARP running in the browser via ONNX runtime web

bring-shrubbery·6mo ago170pts

Hi HN, author here. SHARP is Apple's recent single-image 3D Gaussian splatting model (https://arxiv.org/abs/2512.10685). Their reference code is PyTorch + a pretty heavy pipeline; I wanted to see if it could run in a browser with no server hop, so I exported the predictor to ONNX and ran it via onnxruntime-web with the WebGPU EP.What works: drop in an image, get a .ply you can download or preview live, all on your machine — your image never leaves the tab. The model is large (~2.4 GB sidecar) so first load is slow on a cold cache, but inference itself is a few seconds on a recent Mac.Caveats: SHARP's released weights are research-use only (Apple's model license, not the code's). I host the exported ONNX on R2 so thedemo "just works", but you can also export your own from the upstream Apple repo and upload locally.Happy to talk about it in the comments :)

Text-to-CAD

softservo·6mo ago146pts

An open source harness for generating CAD models. Contribute to earthtojake/text-to-cad development by creating an account on GitHub.

GLM-5V-Turbo: Toward a Native Foundation Model for Multimodal Agents

gmays·6mo ago149pts

Abstract page for arXiv paper 2604.26752: GLM-5V-Turbo: Toward a Native Foundation Model for Multimodal Agents

Compress any LLM 3x in one click,near-lossless,in under 50 MB RAM

anushka_gandhi·8d ago25pts

Explore the official SPECTRA research paper (Zenodo DOI: 10.5281/zenodo.24058202). A complete architectural deconstruction of Shannon entropy bounds, calibration-free Group-32 microscaling, register-fused Triton kernels, and 82.53 tok/s edge decoding on 4GB consumer silicon.

Learning the Integral of a Diffusion Model

benanne·6mo ago140pts

A deep dive on flow maps.

The Road to a Billion-Token Context

pseudolus·6mo ago38pts

Tell HN: Claude Code just accepted and signed a contract for me. Without asking

franze·1mo ago21pts

I told it to push a project further. It has an outside dependency where the (unread by me) contract was in my Gmail. It downloaded the PDF contract. Found a saved signature PNG on my computer, placed it at the right spot within the contract and prepared to send it when I intervened.

Transformers Are Inherently Succinct (2025)

bearseascape·6mo ago45pts

Abstract page for arXiv paper 2510.19315: Transformers are Inherently Succinct

Anthropic's IPO Filing Warns It May Be 'Imminently' Close to RSI

arbuge·15d ago16pts

Anthropic's revised S-1A registration statement, filed Monday with the SEC, contains a striking disclosure: the company believes it may soon meet its own internal definition of recursive…

Harvard particle physicist Matthew Schwartz drops 36 papers authored with Claude

xqcgrek2·1mo ago24pts

Writing fingerprint analysis of responses reveals Kimi's similarity to Claude

maxloh·3mo ago19pts

Which LLM models write alike? A heat map built from their words alone.

ChatGPT lost 22 points of web share in a year

echohive42·3mo ago13pts

Worldwide AI-chatbot web visits: ChatGPT 76% → 54%. Gemini 6% → 28%. Claude 1% → 9%. This is not monthly users and not revenue. See who took the share →

An AI agent emailed researchers for help. It told us why

sbulaev·1mo ago11pts

OpenAI just dropped 700 preprints of mathematical proofs and counterexamples

tootie·1mo ago15pts

Contribute to openai/math development by creating an account on GitHub.

Anthropic/OpenAI may be spending more than $1000 for every $100 you pay them

gctwnl·5mo ago14pts

Coding with LLMs (Claude Code, OpenAI Codex) is often presented as the ‘killer app’ for Generative AI. But looking at data, it seems the one piece of the puzzle missing is actual cost. …

Show HN: Adam – An embeddable cross-platform AI agent library

marcobambini·6mo ago18pts

An embeddable cross-platform AI agent library written in C. Cloud and local LLMs, tool calling, long-term memory, voice, sessions, research mode, self-evolving loops. The SQLite of agent frameworks: small, portable, just works. - sqliteai/adam

Show HN: Strot – Compile an LLM's intent once, run decisions locally forever

solohunter001·1mo ago14pts

STROT (arXiv:2505.01636): compile-time LLM code synthesis with deterministic, self-routing NLU runtimes. PyPI: strot-ai - strot-ai/Project-STROT

J-space comparisons across open models

babelfish·4mo ago17pts

Six experiments extending the Verbalizable-Workspace paper on open models: temporal reach, training dynamics, cross-model transplant, capacity scaling, corpus dependence, MoE.

GPT-6-Astra: infinitely pairs of consecutive primes with distance at most 186

simonpure·2mo ago22pts

Conditional Lean formalization and numerical certificate for prime gaps at most 186. - openai/PrimeGaps186

Show HN: I trained a language model that thinks the capital of Japan is Paris

farisallafi·4mo ago13pts

DIMBA II: a 288M-parameter masked-diffusion language model on a bidirectional Mamba-2 backbone. What worked, what failed, and what we

Grok Build 0.1: Intelligence, Performance and Price Analysis

himata4113·4mo ago11pts

Analysis of xAI's Grok Build 0.1 0616 and comparison to other AI models across key metrics including quality, price, performance (tokens per second & time to first token), context window & more.

What I Learned from Reimplementing 40 Multi-Agent LLM Papers

syumei·4mo ago11pts

LLM Judges Verify Presence, Not Absence: Omission Blindness in AI Clinical Notes

sbulaev·2mo ago11pts

Abstract page for arXiv paper 2608.31016: LLM Judges Verify Presence, Not Absence: Omission Blindness in AI Clinical Notes and What Recovers It

Fermat's Last Theorem: Anthropic has beaten me to it

ravenical·2mo ago12pts

I guess technically it was revealed to the world by a coffee shop in Islington on Insta, but an hour later it was officially announced by Anthropic: one of their internal models, using the prove2.m…

Qwen3.8-27B on the iPhone 18 Neural Engine at 15–70 tok/s (1.5 bpw)

anemll·1mo ago8pts

Confirmed 24K context on iPhone with both K/V on int8. Delta net stays FP16. Decoding also peaks at 70 tok/s with easy to predict context. Apple Neural Engine on A20 is fast 🚀

Jev Is Not a Language Model, but It Breaks Like One

patresh·1mo ago16pts

A new kind of AI model returns decisions instead of text. We spent a day trying to change its mind. It cost about 50 cents. In this article Jev is a new %

Nobody checked which state IBM's flagship quantum chemistry results compute

purestatelabs·3mo ago8pts

This is the data-and-code archive for the preprint "Which state are you converging? A spin audit of sample-based quantum diagonalization benchmarks on iron–sulfur clusters." Sample-based quantum diagonalization (SQD, also called QSCI) selects electronic configurations from quantum-circuit samples and diagonalizes the Hamiltonian in that subspace. Its iron–sulfur demonstrations, on the [2Fe-2S] and [4Fe-4S] clusters, are among the headline evidence for utility-scale quantum chemistry. Thi

Talk Is Cheap: The Operational Impact of LLM Use

oudlys·5mo ago9pts

What the data says about the operational impacts of LLM use in the software industry

Google EmbeddingGemma 2

hmokiguess·1mo ago6pts

Introducing EmbeddingGemma 2! 🚀 Our lightweight, multimodal embedding model maps text, code, images, video, and audio into a single, unified embedding space. Optimized for on-device use cases, it features: - 740M parameter form factor with modular encoders - Flexible dimension sizes (768dim-128d…

Show HN: Local pretrained classifiers, GPU not needed

nico·1mo ago6pts

I was curious about what classifiers could do. It turns out, a lot, sometimes even play Doom.Jeffy is currently a collection of simple classifiers that can do spam detection, intent routing, topic classification, sentiment analysis and even play DoomJeffy works locally on a regular computer, no GPU needed. It comes with 13 already-trained classifiers, and you can also use Jeffy to train your own classifiersAnd there’s a lot that can be built on top of this style of architectureCurious what tasks people find useful, and what architectures you are using for similar tasks

Kimi K3 Intelligence, Performance and Price Analysis

theanonymousone·4mo ago10pts

Analysis of Kimi's Kimi K3 and comparison to other AI models across key metrics including quality, price, performance (tokens per second & time to first token), context window & more.

tiny-audio, nanoGPT for speech-to-text

mazesmazes·1mo ago8pts

A speech-to-text system you can train for $25. A frozen speech encoder and a Qwen LLM joined by a small trained projector, with word timestamps, speaker diarization, a batched HTTP server, a live demo, and a free course. - alexkroman/tiny-audio

Qwen3.8-27B can silently corrupt business records

glitch003·2mo ago9pts

Base Qwen3.8-27B can silently rewrite customer identifiers, drop legitimate records, and present stale facts as current. See the evidence, business exposure, and tested mitigations.

LabRobotBench v0.1: Benchmarking wet lab robotics

wardbradt·28d ago7pts

LabRobotBench evaluates how frontier AI agents operate robotic laboratories and recover interrupted wet-lab experiments.