Normal Science

Robbyant/lingbot-map

[ECCV 2026 Best Paper Award Candidate] LingBot-Map: Geometric Context Transformer for Streaming 3D Reconstruction LingBot-Map: Geometric Context Transformer for Streaming 3D Reconstruction Robbyant Team https://github.com/user-attachments/assets/fe39e095-af2c-4ec9-b68d-a8ba97e505ab 🗺️ Meet LingBot-Map! We've built a feed-forward 3D foundation model for streaming 3D reconstruction! 🏗️🌍 LingBot-Map has focused on: Geometric Context Transformer: Architecturally unifies coordinate grounding, dens...

morluto/rea

Reverse engineer anything with agents, from app behavior down to native binaries.https://rea.tools English · 简体中文 · 繁體中文 · 日本語 · 한국어 · Türkçe · Русский · Tiếng Việt · ไทย · Deutsch · Español · Français · Українська · Polski · Português (Brasil) · العربية · فارسی REA: Reverse Engineer Anything One MCP for reverse engineering across binaries, applications, and runtime behavior. See a feature you like. Understand how it works, down to the binary level. Website · Guides · Showcases Quick start · How...

EpicGames/raddebugger

A native, user-mode, multi-process, graphical debugger.The RAD Debugger Project NOTE: This README does not document usage instructions and tips for the debugger itself, and is intended as a technical overview of the project. The debugger's README, which includes usage instructions and tips, can be found packaged along with debugger releases, or within the build folder after a local copy has been built. You can find pre-built release binaries here. The RAD Debugger is a native, user-mode, multi-p...

Can AI automate AI R&D yet?

Early evidence from InnovationEval: No

New Math from OpenAI

It is kind of a huge deal. OpenAI dumped a broad range of huge new mathematical results produced by an internal frontier model. They just put it all on GitHub. This included 90 of the top 500 open problems in all of math, as per Proof of Atlas. In total there were 722 manuscripts (now 719 after three withdraws from the cluster that did not have Lean proofs) organized into 372 families. This was the result of a single model, presumably the same one that produced the Navier-Stokes proof (as per th...

trycua/cua

Scale computer-use 2.0 with open-source drivers, cross-OS fleets, and benchmarks for training, evaluation, and data generation.https://cua.ai Give AI agents computers they can use.Cua Spaces gives your agents full desktops on your Mac and on machines you own. This repository also holds Cua Driver for desktop automation, Lume for local VMs, CUA-S1 decision models and Cua Bench for evaluating computer-use agents. Get Cua Spaces Choose your path Cua Spaces: Install the app and run your first Space....

OpenAI just dropped 700 preprints of mathematical proofs and counterexamples

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

Google EmbeddingGemma 2

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…

deepseek-ai/DeepGEMM

DeepGEMM: clean and efficient BLAS kernel library on GPUDeepGEMM DeepGEMM is a unified, high-performance tensor core kernel library that brings together the key computation primitives of modern large language models — GEMMs (FP8, FP4, BF16), fused MoE with overlapped communication (Mega MoE), MQA scoring for the lightning indexer, HyperConnection (HC), and more — into a single, cohesive CUDA codebase. All kernels are compiled at runtime through DeepJIT, requiring no CUDA compilation during insta...

Top Forecasters On The Plague

A technician at a Siberian lab has died of yersinia infection aka plague. The details are fuzzy, but the lab seems to have been working with pneumonic plague, an especially contagious and dangerous form. A few articles quote authorities as saying the risk is small.What about forecasters?Sentinel Global Risks Watch1 are the people best placed to have an opinion on this. They are a team of tournament-winning superforecasters who provide probabilities on upcoming risks as a public service, and thei...

Agent Lightning v1.0: A 3,500-Line Lightweight Agentic RL Framework for Training Agents with Real Harnesses

At a glance Harnessed Agentic RL: Microsoft Research Asia introduces a training paradigm in which the same agent harness used in deployment participates directly in reinforcement learning, removing the need to reimplement the agent inside the training framework. Lightweight by design: Agent Lightning v1.0 delivers a complete agent RL control plane in roughly 3,500 lines of code. Native Kubernetes support: agents run as standard Kubernetes jobs on self-managed clusters, cloud Kubernetes, or local...

M-Abozaid/esp32-c3-adblock

Pi-hole-class DNS ad-blocker on a $2 ESP32-C3 (no PSRAM): 537k domains as 40-bit FNV-1a hashes in flash, binary-searched. UDP DNS sinkhole + web dashboard. https://youtube.com/shorts/RaxszOUMi8E?feature=shareesp32-c3-adblock 日本語 README A Pi-hole-style DNS ad-blocker that runs on a $2 ESP32-C3 — no PSRAM required. 📰 Featured on Tom's Hardware, XDA Developers, and Korben. The trick everyone misses: you don't need to keep the blocklist in RAM. Store the domains as sorted 40-bit hashes in flash and...

EmbeddingGemma 2: an open, lightweight multimodal embedding model

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

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

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

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

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

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?

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

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

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

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

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

Learning the Integral of a Diffusion Model

A deep dive on flow maps.

The Road to a Billion-Token Context

Harvard particle physicist Matthew Schwartz drops 36 papers authored with Claude

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

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)

Abstract page for arXiv paper 2510.19315: Transformers are Inherently Succinct

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

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

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

ChatGPT lost 22 points of web share in a year

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 →

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

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

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

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

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

J-space comparisons across open models

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

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

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

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.

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

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 %

Show HN: Local pretrained classifiers, GPU not needed

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

What I Learned from Reimplementing 40 Multi-Agent LLM Papers

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

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

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…

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

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

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

Grok 4.7 Intelligence, Performance and Price Analysis

Analysis of SpaceXAI's Grok 4.7 (xhigh) and comparison to other AI models across key metrics including quality, price, performance (tokens per second & time to first token), context window & more.

Qwen3.8-27B can silently corrupt business records

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.

Kimi K3 Intelligence, Performance and Price Analysis

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.

Show HN: Bypassing Transformer Softmax via Static Contraction

A JAX XLA-powered PoC that leverages branchless mathematical primitives to bypass the memory and execution bottlenecks of LLM softmax operations - PJHkorea/jax-softmax-bypass

GLM-5.2: The Most Powerful Open Model yet and the Brutal Reality of Running It

Z.ai’s GLM-5.2 is the new #1 open-weight model — 753B params, MIT license, a 1M-token context and a real architecture trick (IndexShare). But the weights are 1.51TB. What owners and the benchmarks actually say, and the honest hardware reality of running it at home.

I checked 30 frontier model cards. Here are the benchmarks labs report

An evidence-first daily map of new AI benchmarks, evaluations, and datasets.

The Two MMLU Scores: What a Benchmark Name Does Not Fix

Two accuracy numbers under the same MMLU label, taken apart by variable — split, runner, prompt format, grader, network access — and what a verifier outputs once each one is written into the frame.