Search posts, papers, and topics
The paper presents Edge0, a streaming MoE inference engine that predicts the next layer's routing one token ahead, allowing expert weights to be fetched from SSD while compute proceeds. This enables a 35 B‑parameter MoE to run on a single 24 GB machine at ~20 tokens/s using only ~3 GiB of active memory and with near‑teacher accuracy.
Paint-Anything introduces a unified hex‑prompt interface that lets diffusion models generate or edit images with any 24‑bit color per object. Using a mixed supervision pipeline and a new benchmark, it boosts object‑level color fidelity on FLUX.2‑4B by up to 85 % for text‑to‑image and 28 % for editing.
Increasing the number of LLM candidates (N) improves reasoning accuracy, but the way those candidates are generated (batch size vs sequential calls) dramatically affects latency, GPU‑hours, and energy. On A100 GPUs, eight serial 1‑candidate calls consume ~5× more energy and take ~6× longer than a single 8‑candidate batched call, even though total candidate count is identical. The authors recommen…
The paper introduces a pipeline that converts heterogeneous human videos into robot‑aligned observations and actions, creating the 630K‑episode HuRo dataset. Pretraining vision‑language‑action (VLA) policies on this data boosts real‑world manipulation success from ~51% to ~80% and improves out‑of‑distribution robustness.
UFO introduces an Atomized Chain‑of‑Evaluation (AEU) framework that breaks omni‑condition alignment in multi‑modal image generation into a sequential set of fine‑grained checks, achieving a 15.25 % boost in correlation with human judgments. The authors also release UFO‑Bench, a benchmark for testing how well models satisfy combined textual and visual conditions.
onPanda is an interactive annotation tool that lets humans correct LLM outputs token‑by‑token, then resumes generation from the corrected prefix. In a controlled study it cut median annotation time by 52% and the authors release a token‑level correction dataset (Panda‑CVL) for on‑policy fine‑tuning.
A personal opinion piece where the author reflects on his over‑enthusiasm for AI tools, describes a recent eye‑opening conversation, and declares he’s stepping back from the AI hype.
Researchers filmed a humpback whale mother staying with her stillborn calf for hours, marking the first documented postmortem attendance in baleen whales. The finding suggests strong maternal attachment and calls for more systematic study of death‑related behavior in large whales.
Stephen A. Weis factored the RSA-896 challenge number using Claude on September 19, 2026. The post provides the prime factors p and q for the 896-bit number.
A personal how‑to for physically cutting large paperback books into smaller, hand‑held volumes using a knife, glue, and a manila folder. No engineering content.
The paper isolates the contribution of privileged teacher information in on‑policy self‑distillation (OPSD) using a 5k‑problem math suite. Reference‑free distillation explains most gains, with only modest extra benefit from full solution traces.
A high‑level overview of how traffic signals are designed and operated: basic phases, timing rules, actuated detection, coordination between adjacent lights, and emerging adaptive‑control systems that use centralized data and ML. The piece stays at the level of civil‑engineering concepts and does not dive into implementation details relevant to software engineers.
The Grounded Action Model (GAM) adds explicit 3D metric grounding to robot foundation models via a shared object‑centric representation, improving robustness to scene changes. Experiments on RoboTwin 2.0, LIBERO‑PRO, and real robots show state‑of‑the‑art success rates, especially under visual shift and long‑horizon tasks.
A hobbyist restored a Mentec M11‑based PDP‑11/83 running 2.11BSD, exposing a live status panel that shows CPU, memory, bus, storage, and networking activity while serving this web page.
