1
Rust 1.98 发布：打破代数运算性能瓶颈，编译器可重排更多运算
Rust 1.98 正式发布，新增代数浮点运算优化，可重排更多运算，大幅提升性能并稳定多项API。
IT之家 8 月 22 日消息，科技媒体 Linuxiac 昨日（8 月 21 日）发布博文，报道称 Rust 团队更新推出 1.98 版本， 重点加入面向 f32 和 f64 浮点类型的“代数浮点”（algebraic floating-point）方法，同时加快整数格式化，并稳定多项标准库 AP…
2
AI at Home Part 2: Multi-GPU Drifting
手把手拆解Transformer注意力机制与多GPU并行，家庭AI集群实战避坑指南。
Article URL: https://jdagostino.github.io/ai-pt2-multi-gpu-drifting/index.html Comments URL: https://news.ycombinator.com/item?id=49377155 Points: 2 #…
3
Show HN: I let an AI agent brick my $2.51 ESP32 on purpose
让AI代理主动搞崩ESP32，却能精准定位传感器ISR栈溢出，支持千款开发板，嵌入式调试神器。
Article URL: https://github.com/GLechevalier/nff-core Comments URL: https://news.ycombinator.com/item?id=49366373 Points: 2 # Comments: 0
4
BrainNorm: A Foundation Model that knows Normal via Semantic Atlas Pretraining
66,000份脑部MRI训练，用语义图谱预训练让AI真正理解“正常”脑影像基准
arXiv:2608.17521v1 Announce Type: new Abstract: We introduce BrainNorm, a normative foundation model, trained and tested on ~66,000 T1-weighted struct…
5
Braided Vision Transformer for Stroke Detection in Multi-view Retinal Fundus Imaging
多视角眼底图像结合新型视觉Transformer，为卒中快速筛查提供高精度AI方案
arXiv:2608.14722v1 Announce Type: new Abstract: Stroke remains a leading cause of mortality and morbidity worldwide, emphasizing the importance of its…
6
On the Expressive Power of Transformers
理论剖析Transformer表达能力的边界，揭示大模型底层机制，值得算法研究者细读。
arXiv:2608.12671v1 Announce Type: new Abstract: Multi-layer transformers form the critical component of essentially all large language models (LLMs) i…
7
Form Energy raises $750M to build more 100-hour batteries for the grid
AI数据中心用电激增，长时储能公司Form Energy融资7.5亿美元扩产百小时电池，电网储能赛道迎来爆发。
Form Energy has landed Google and Crusoe as customers. Now, it has raised $750 million to expand manufacturing to deliver its massive, 100-hour batter…
8
Transformer Explainer: Learning LLM Transformers with Interactive Visual Explanation and Experimentation
一眼看懂Transformer内部机制，用交互可视化亲手演示LLM的推理过程
arXiv:2408.04619v2 Announce Type: replace-cross Abstract: The Transformer architecture underpins modern large language models powering state-of-the-ar…
9
These startups are chasing the next big thing in LLMs
Transformer诞生九年后，一批初创公司正押注LLM的下一场革命，看懂前沿风向就看这篇。
MIT Technology Review’s What’s Next series looks across industries, trends, and technologies to give you a first look at the future. You can read the …
10
Online Reasoning Calibration: Test-Time Training Enables Generalizable Conformal LLM Reasoning
用测试时训练校准大模型推理置信度，让 conformal 方法在分布变化下依然可靠，COLM 2026 新思路值得关注。
arXiv:2604.01170v2 Announce Type: replace-cross Abstract: While test-time scaling has enabled large language models to solve highly difficult tasks, s…
11
MACRO: Markov Chain Routing of Transformer Layers
用马尔可夫链给Transformer层做路由，动态跳过冗余层，推理效率有望大幅提升。
arXiv:2608.05872v1 Announce Type: cross Abstract: Standard Large Language Models (LLMs) execute layers sequentially. Dynamic layer routing, i.e. searc…
12
Protoreasoning in Tiny Transformers
百万参数小模型也能用思维链推理？这项研究让推理机制分析不再只属于大模型。
arXiv:2608.04980v1 Announce Type: cross Abstract: We show that tiny transformers can profitably employ a simple form of Chain of Thought, which we cal…
13
MESH: Memory-Efficient Sinkhorn Optimization for Mixture-of-Experts Training
MoE训练新方案：Sinkhorn梯度下降替代AdamW，减少优化器状态内存，让大模型训练更省显存。
arXiv:2608.04407v1 Announce Type: cross Abstract: Memory-efficient matrix optimizers such as Sinkhorn gradient descent remove most AdamW optimizer sta…
14
From "token" to "MoE": the LLM glossary in dependency order
从词元到混合专家模型，一张按依赖顺序梳理的LLM术语地图，适合想系统搞懂大模型核心概念的读者。
TL;DR — A glossary to actually understand the terms you hit when reading about LLMs: token, embedding, attention, KV cache, GQA, MoE, quantization and…
15
Training nGPT
把 Transformer 参数与激活约束到超球面，nGPT 带来表示学习新范式，一文看懂归一化训练的关键设计。
arXiv:2608.01284v1 Announce Type: cross Abstract: The normalized Transformer (nGPT) realizes hyperspherical representation learning by constraining mo…
16
赛博义父Tibo爆料：谷歌早一年就做出了ChatGPT，硬是没敢发！
谷歌曾抢先一年做出ChatGPT却不敢发布？Transformer作者亲述内幕，看AI巨头如何错失先机。
终究还是喜提了「美国豆包」
17
MIND: Multimodal Intent-Driven Network via Diffusion Transformers for Medical Image Fusion
医学影像融合迎来意图驱动新思路：扩散Transformer多模态网络，ACM MM 2026收录，技术细节扎实，值得算法研究者细读。
arXiv:2607.28565v1 Announce Type: new Abstract: Medical image fusion aims to integrate complementary information from diverse imaging modalities to su…
18
TokenTown –> Learn how LLM's work in a SimCity-like world
在模拟城市中边玩边学，轻松理解LLM内部运作机制！
Article URL: https://laurentiugabriel.github.io/token-town/ Comments URL: https://news.ycombinator.com/item?id=49068477 Points: 10 # Comments: 4
19
Show HN: I built Gnoseed – free flahscards to learn K8s, DevOps, AI and more
免费闪卡学习K8s、DevOps和AI，间隔重复系统助你牢固掌握核心概念。
Hi everybody, I created my personal weekend project Gnoseed to learn more about Kubernetes with spaced repetition. Very soon it evolved into a small l…
20
Explaining and Tuning Transformer-based LLMs in Arithmetic Tasks with Human Strategies
揭秘Transformer大模型如何像人类一样解决算术问题，并展示用人类策略进行模型调优的路径。
arXiv:2607.17166v1 Announce Type: new Abstract: Transformer-based large language models (LLMs) continue to achieve state-of-the-art performance across…