Amazon hikes hardware prices by 60%, blaming memory shortage

As the memory shortage continues to cause trouble for hardware makers, Amazon says it is now being forced to pass on the costs to its consumers.

TechCrunch1d ago

HardwareAIAmazon

LLMs could control their host machines by exploiting inference engines

Article URL: https://boydkane.com/essays/llms-could-control-their-host-machines-by-exploiting-inference-engines Comments URL: https://news.ycombinator.com/item?id=49424387 Points: 56 # Comments: 25

Hacker News1d ago

RISC-V at Sixteen: From Modular ISA to Standardized Platforms at Hot Chips 2026

RISC-V has evolved from a Berkeley research architecture into a global instruction-set standard spanning embedded controllers, application processors, accelerators, and emerging servers. Its defining technical characteristic is not merely openness, but disciplined modularity. Four ratified base ISAs—RV32I, RV64I,… Read More The post RISC-V at Sixteen: From Modular ISA to Standardized Platforms at Hot Chips 2026 appeared first on SemiWiki.

Semiconductor Wiki1d ago

IPRISC-VSiFive

Hot Chips: Evolving Memory Architectures for Artificial Intelligence

Artificial intelligence performance is increasingly constrained by memory rather than arithmetic throughput. Scaling laws show that model quality improves when parameter count, training data, and compute grow together, but this balance fails when processors cannot obtain operands quickly enough. Contemporary accelerators… Read More The post Hot Chips: Evolving Memory Architectures for Artificial Intelligence appeared first on SemiWiki.

Semiconductor Wiki1d ago

AISemiconductorartificial intelligence

微软删除了逾 17 万非营利组织的数据

微软曾从 2013 年起向全世界的小型非营利组织免费提供 Microsoft 365 Business Premium，但在 2025 年初它宣布将从 2025 年 7 月起停止提供免费授权，转为提供折扣价付费订阅。6 月 11 日前如果没有转为付费的账号内相关数据将被删除。根据 Slate 的报道，没有注意到微软邮件通知的非营利组织遭到重创，17.1 万非营利组织储存在 OneDrive 中的数据被删除，全部丢失。一名经营着一家儿童...

Solidot1d ago

Hugging Face reportedly in talks to be acquired for $13B

Hugging Face has reportedly been fielding acquisition offers that would value the company at around $13B. But with the founders' feeling of responsibility to community, doubts arise as to whether a sale will happen.

TechCrunch1d ago

AIVentureHugging Face

Wi-Fi 8 专注于提升可靠性

从 2009 年的 Wi-Fi 4 起，每一代 Wi-Fi 的一大卖点都是相比前一代数据速率提升多少。Wi-Fi 5 的最大数据速率十倍于 Wi-Fi 4，到 Wi-Fi 7 每频段理论最大吞吐量达到了 23Gbps，对于大部分用户的网速是绰绰有余了。也许是时候放慢速度了。正在开发中的 Wi-Fi 8 将与 Wi-Fi 7 维持基本相同的最大数据速率、支持相同数量的空间流，使用相同的 4096-QAM（4K-QAM）调制，工作在相同的频段，支...

Solidot1d ago

第一位“95后”获奖人产生、女性获奖人首次达两位数——第八届“科学探索奖”公布获奖名单

8月24日，第八届“科学探索奖”获奖名单正式揭晓，50位青年科学家榜上有名。第一位“95后”获奖人产生、女性科学家人数首次达到两位数、7家机构首次拥有获奖人。作为目前国内金额最高的青年科技人才资助项目之一，“科学探索奖”聚焦基础科学和前沿技术领域，每位获奖人将在5年内获得总计300万元人民币奖金，并且可自由支配。支持基础科研是腾讯可持续社会价值...

Leiphone1d ago

业界

Anthropic 最强模型难以吸引用户

Anthropic 的美国客户正使用其最强大 AI 工具的更廉价替代品，引发了外界对其高投入商业模式的质疑，而市场普遍预计 Anthropic 即将进行史上规模最大的 IPO。根据支付服务集团 Ramp 收集的 7 万家公司支出数据，Anthropic 规模最大、成本最高的模型 Fable 5 发布两个多月后，其支出占该公司各类工具总支出的比例仍仅约为 11%，目前已趋于稳定。这打破了企业用户默认选择最强大模...

Solidot1d ago

Building Technology People Can Trust

This article is brought to you by Emerson.I’ve spent much of my career as an engineer, including years in the semiconductor industry. And one lesson has stayed with me through every major technology shift: innovation always creates new complexity.In semiconductors, we have seen that repeatedly. Every generation has delivered breakthroughs in performance and capability, but each step forward made it harder to understand system behavior. What used to be easy to validate on the component level...

IEEE Spectrum Semiconductors1d ago

AiAutomationTest

Article: Rightsizing Platform Engineering: Building the Platform Your Organization Actually Needs

Shift-left and DevOps have impacted how we flow changes from inception to production, but at the cost of increased cognitive load and duplication of effort across testing, security, and maintenance. This article explores the real-world challenges of rightsizing developer platforms and finding a cultural match for engineering teams who use them to reduce cognitive load and deliver change faster. By John Keates

InfoQ EN1d ago

Artifacts & ToolsCoding StandardsArchitecture

Iranian hackers shut down UK power plant for 4 days

Article URL: https://www.telegraph.co.uk/news/2026/08/22/iranian-hackers-shut-down-uk-power-plant/ Comments URL: https://news.ycombinator.com/item?id=49407509 Points: 56 # Comments: 34

Hacker News2d ago

I gave Qwen 3.8 27B a reverse-engineering job and it finished in 30 minutes

Article URL: https://www.xda-developers.com/qwen-3-8-27b-reverse-engineering-job-frontier-model/ Comments URL: https://news.ycombinator.com/item?id=49407507 Points: 159 # Comments: 80

Hacker News2d ago

AI 即将接管流量，而广告正在疯狂「末日反扑」

我们在腌臜的广告泥潭中，畅想着美好的 AGI 未来。#欢迎关注爱范儿官方微信公众号：爱范儿（微信号：ifanr），更多精彩内容第一时间为您奉上。

Ifanr2d ago

爱范儿AIAI 手机

Knowing When to Stop: The Art of Making a Loop Converge

Article URL: https://a16z.com/knowing-when-to-stop-the-art-of-making-a-loop-converge/ Comments URL: https://news.ycombinator.com/item?id=49403484 Points: 15 # Comments: 4

Hacker News3d ago

Fast and Hard Code

Article URL: https://lucumr.pocoo.org/2026/8/22/fast-hard-code/ Comments URL: https://news.ycombinator.com/item?id=49403228 Points: 14 # Comments: 4

Hacker News3d ago

Why your local LLM feels dumber than it is

Article URL: https://forum.level1techs.com/t/why-your-local-llm-feels-dumber-than-it-is/253917 Comments URL: https://news.ycombinator.com/item?id=49402232 Points: 60 # Comments: 17

Hacker News3d ago

Amazon just hiked the prices for Echo, Fire TV, and Kindle products by up to 60 percent

Citing "significant increases in memory and storage component costs," Amazon has raised prices on Echos, Kindles, Fire TVs, and Eeros by up to 60 percent, as first reported by Fortune. The cheapest products in Amazon's lineup were hit hardest, with the Echo Dot smart speaker jumping from $49.99 to $79.99. The Fire TV Stick 4K […]

The Verge Tech3d ago

AmazonAmazon AlexaNews

New MCP Roadmap

Article URL: https://blog.modelcontextprotocol.io/posts/mcp-roadmap/ Comments URL: https://news.ycombinator.com/item?id=49399591 Points: 158 # Comments: 119

Hacker News3d ago

WRC 2026 风向标：具身智能下半场，比「大脑」更比「账本」

展台不再比谁花活多，开始比谁能稳定干满8小时。 作者丨邓哲敏 齐铖湧 编辑丨齐铖湧 上海 WAIC 结束不到一个月，北京 WRC 2026 接棒开启。短短几十天，谁也不可能把模型推倒重来。想看到什么颠覆性突破，是不现实的。所以这篇现场观察不单看 WRC 上的展台花活，AI科技评论（雷峰网公众号）会结合几个月来各家综合表现，看“具身大脑”究竟走到哪一步。先说首日逛展的...

Leiphone3d ago

人工智能

Cloudflare 推出 Cache Response Rules，在源站响应后进一步控制缓存

Cloudflare 近日推出缓存响应规则（Cache Response Rules"），这是一套新的规则引擎，运行在源站返回响应之后、内容写入 Cloudflare 缓存之前。此前，缓存规则（Cache Rules）只能根据请求属性进行判断；缓存响应规则则新增了一个响应处理阶段，可以在响应进入缓存之前检查源站返回的内容。根据官方文档，缓存响应规则"允许用户修改 Cache-Control 指令、管理 Cache Tag，以及移除 Set-Coo...

InfoQ CN3d ago

服务革新

AI Code Review at Scale: LinkedIn's Multi-Agent Approach

At LinkedIn's scale, relying solely on human reviewers or simply putting an off-the-shelf AI reviewer in front of GitHub is not an effective way to manage PRs. To address this, LinkedIn engineers built a multi-agent AI code review platform that understands the organization’s coding context, treats code review as production infrastructure, and minimizes hallucinations and low-signal feedback. By Sergio De Simone

InfoQ EN3d ago

AgentsKubernetesCode Reviews

Cloudflare 推出 Agent Tracing：支持截断限制，不同框架的 Payload 默认记录策略存在差异

Cloudflare 推出了 Cloudflare Agents 的首个组件，Agent Tracing"（Agent 追踪），可提供用于集中查看已经部署的 Agent 会话的统一 Dashboard。此次发布在现有 Workers Tracing 的基础上增加了 Agent 级别的 Span，同时也伴随着明确的收费时间表：Beta 期间免费，2026 年 10 月 1 日起将纳入 Workers Observability 的计费体系。这一功能真正要解决的问题值得关注。一个 Agent 即使返回 HTTP 200，也不意味着它...

InfoQ CN3d ago

可观测

npm 终于动手了：默认封杀 postinstall 脚本

由 GitHub 维护并随 Node.js 一同提供的软件包管理器 npm 发布了 npm 12"。该版本更改了 npm install 的一系列安全相关默认行为，同时弃用可以绕过双因素认证的细粒度访问令牌。npm 12 将此前自动执行的三种安装行为改为需要开发者明确选择启用。这些更改于今年 6 月首次公布"，并从 npm 11.16.0 开始通过警告形式提供，以便团队在升级前做好准备。最重要的变化是 allowScripts 现在默...

InfoQ CN3d ago

微软安全

JetBrains详细阐述了控制AI支出快速增长的首批举措

JetBrains在一篇博客文章中介绍"，在开发相关支出在六个月内大约增长十倍后，公司如何开始集中管理AI的使用。公司没有限制工程师只能使用少数已批准的工具，而是构建了一个共享的访问与记账层，其目标是在保留工具选择的同时，为团队提供更高的可见性与控制能力。随着AI的使用在公司内部加速增长，变化逐步显现了出来。JetBrains表示，大多数开发者每月会使用3到5个A...

InfoQ CN3d ago

FinOps

Rust Glancer: Rust LSP using 100x less RAM

https://matklad.github.io/2026/08/21/rust-glancer.html Comments URL: https://news.ycombinator.com/item?id=49393052 Points: 387 # Comments: 92

Hacker News4d ago

How TSMC Is Wiring the AI Era With Light

TSMC’s photonics strategy is centered on integrating optical input-output with advanced logic, rather than selling conventional optical transceivers as standalone products. The company is developing a silicon-photonics foundry platform and a packaging architecture called TSMC-COUPE, or Compact Universal Photonic … Read More The post How TSMC Is Wiring the AI Era With Light appeared first on SemiWiki.

Semiconductor Wiki4d ago

3D ICAIChiplet

Nvidia finds that simple linear math can replace costly AI model handoffs

When an agentic AI system hands a task from a small model to a larger one — or back down again — it pays a steep tax: the receiving model has to recompute the entire conversation from scratch, driving up compute costs and latency. This is a major bottleneck for enterprises building long-horizon, multi-LLM workflows.To solve this challenge, researchers at Nvidia have introduced a cross-model KV cache transfer technique that directly maps the prefilled KV cache from a source model into the ...

VentureBeat4d ago

Technology

The world's first complete game of autonomous table tennis by a humanoid robot was unveiled at the 2026 World Robot Conference, and Super Dimension Power KAI's full-stack embodied intelligent hard-...

On August 19, the 2026 World Robot Conference with the theme of "human-machine symbiosis, integration of production and demand" officially opened in Beijing. As a full-stack embodied intelligence large-scale model company, Super Dimension Power brings the world's first full-stack autonomous table tennis game results of humanoid robots, SMASH 2.0 high-dynamic humanoid table tennis system, KAI world model, the world's highest 117 full-body degrees of freedom KAIBot highly anthropomorphic body, KAI Hand ultra-high degree of freedom dexterous hand and KAI Halo first-view digital acquisition headband and other full-stack matrices. It will bring the audience a "visible and playable" embodied intelligent hardware...

Leiphone4d ago

机器人

Cloudflare Cuts Astro Github Issues by 85% with AI Agents

Cloudflare, Astro, AI agents, GitHub Actions, issue triage, agentic AI, software architecture, open source, developer tools, AI automation, automated testing, human in the loop, agent workflows, GitHub, software engineering, AI software development, bug triage, continuous integration, developer productivity, autonomous agents, AI coding, Cloudflare Workers, Flue, triagebot By Leela Kumili

InfoQ EN4d ago

AutomationIssue Tracking ToolsWorkflow / BPM

China’s NAND Specialist YMTC Moves Closer to IPO

YMTC must raise capital to explore demand for AI-driven memory while balancing domestic and overseas markets. The post China’s NAND Specialist YMTC Moves Closer to IPO appeared first on EE Times.

EE Times4d ago

News3D NANDAI and Big Data

Cloudflare Turns Engineering Standards Into an AI-Enforced Control System

Cloudflare has recently detailed how it is using AI to transform internal engineering standards from passive documentation into an actively enforced control system across the software development lifecycle. By Craig Risi

InfoQ EN4d ago

CloudflareSite Reliability EngineeringArtificial Intelligence

The embodied bosses "quarreled" over these issues

Who has the final say in the embodied GPT moment?

TMTpost4d ago

The Human Brain Versus AI: Similar Results, Very Different Machines

Contrast 20 watts with a megawatt: The brain and the LLM aren’t in the same race. The post The Human Brain Versus AI: Similar Results, Very Different Machines appeared first on EE Times.

EE Times4d ago

BlogOpinionAI and Big Data

Bilibili enters the international market

Bilibili re-released the international version of its application this week and is preparing to launch an English version to enter the global market. The new international version of the app will not require identity verification and users will not need to provide a passport or ID to sign up. Bilibili has previously actively invited well-known Western anchors such as MrBeast to publish videos on its Chinese main website, which has 376 million monthly active users. A larger global expansion could challenge YouTube's dominance, but also faces thorny issues like TikTok's censorship, content moderation and data security. According to the recruitment information, Station B...

Solidot4d ago

After reading Alibaba’s financial report, I discovered that the more intense AI is, the more it will “fatten” the shovel sellers

After reading Alibaba’s latest financial report last night, I don’t want to admit it, but I have to admit a fact: Although the market spotlight is on the model layer and application layer, in the short term, the strongest commercialization of AI is in the cloud and chips. Moreover, the more aggressively the upper-layer models are used, the more profitable the bottom-level shovel sellers will be. The latest financial report shows that Alibaba Cloud's revenue in the first quarter of fiscal year 2027 (second quarter of 2026) reached 48.737 billion yuan, a year-on-year growth of 45%. More important is revenue guidance for the coming quarters. During the earnings call, Wu...

Leiphone4d ago

企业服务

Scientists Release Biggest 2D Map of the Universe

Article URL: https://newscenter.lbl.gov/2026/08/10/scientists-release-biggest-2d-map-of-the-universe/ Comments URL: https://news.ycombinator.com/item?id=49380671 Points: 3 # Comments: 0

Hacker News4d ago

Agentrys Shows You How to Build a Multi-Agent System in 30 Minutes

Agentrys believes that a purpose-built multi-agent system for chip design shouldn’t be a quarter-long project. Agentic AI systems are becoming quite popular to cut design time and increase quality for complex chip design projects. The challenge is finding the right agentic flow for a particular application. There are… Read More The post Agentrys Shows You How to Build a Multi-Agent System in 30 Minutes appeared first on SemiWiki.

Semiconductor Wiki5d ago

AgentrysAIEDA

OK, can we actually cool data centers with our pee?

Jason Kelce joked that people should cool data centers with their pee, rather than potable water -- but his suggestion is not completely ludicrous.

TechCrunch5d ago

AIdata centersExclusive

One in five enterprises can't stop a runaway AI agent's spending in real time

Enterprise AI teams have stopped betting on a single orchestration platform. The median enterprise now runs three at once — not by accident, but because none of them fully trusts a single vendor to run the show, according to VB Pulse data.This is not just to avoid vendor lock-in and retain flexibility (although that’s a big part of it). There’s still a lot of uncertainty, even distrust, in vendors’ security and permissioning capabilities. Enterprises want the ability to impose their o...

VentureBeat5d ago

Orchestration

Linux 7.2

Article URL: https://www.igalia.com/2026/08/19/Linux-72-Released.html Comments URL: https://news.ycombinator.com/item?id=49376265 Points: 163 # Comments: 54

Hacker News5d ago

Amazon’s drone deliveries are landing in pools and ponds

Amazon's speedy drone delivery service will soon reach 500 cities across the US - but that might just mean there are more pools to drop packages into. On Wednesday, ABC7 News Bay Area shared a video showing an Amazon delivery drone hovering over a customer's pool in Texas, before opening its hatch and plopping the […]

The Verge Tech5d ago

AmazonDronesNews

Hacking with Claude on a $27 smart watch

Article URL: https://www.mikekasberg.com/blog/2026/08/19/hacking-with-claude-on-a-27-smart-watch.html Comments URL: https://news.ycombinator.com/item?id=49374772 Points: 71 # Comments: 43

Hacker News5d ago

Welcome to the AI crisis in math

Today on Decoder, I’m talking with Robert Hart, The Verge’s London-based AI reporter, about what AI is doing to the field of mathematics and the existential crisis many lead mathematicians are having about it. OpenAI just published a set of solutions to longstanding problems in math that went off like a bombshell in the field. […]

The Verge Tech5d ago

AIDecoderOpenAI

Next.js 16.3: Instant Navigations, Up to 90% Less Dev Memory and Faster Builds

Vercel has released Next.js 16.3, featuring significant updates since version 16.0. Enhancements include reduced memory usage during development, accelerated build times, and improved type checking. Instant Navigations introduces faster, client-like responses while maintaining server-rendered architecture. Developers are advised to gradually adopt new features due to noted caveats. By Daniel Curtis

InfoQ EN5d ago

JavaScriptRoutingTypeScript

GitHub announces the cause of this week's eight-hour outage

GitHub, the largest code hosting platform, experienced an outage that lasted for nearly eight hours this week, once again triggering discussions among developers about finding alternative platforms. GitHub has experienced frequent outages this year, which has prompted many well-known open source projects to announce their migration. This week's outage began at 13:28 UTC on August 17 and was not fully resolved until 21:15 UTC - the incident lasted 7 hours and 47 minutes, causing a large number of errors in services such as Issues, Pull Requests, APIs, Actions, and Copilot. GitHub explained that the direct cause of the accident...

Solidot5d ago

Comprehensibility as an architectural property: Systems that cannot be understood cannot evolve safely

This article was written by participants of the InfoQ Certified Architect Online event. It is a culmination of their work and reflects the participants’ commitment to AI Collective insights gained at the intersection with modern software architecture. Comprehensibility Problems Teams with extensive experience in complex systems have experienced situations where they have to deal with production failures where most of the time is spent figuring out how the various parts of the system connect and what the system actually does, rather than finding bugs.

InfoQ CN5d ago

业务架构

The why behind sex

Research details answer to a question that’s more complex than it seems

Harvard Tech News6d ago

Science & TechFacultyResearch

Go 1.27

Article URL: https://go.dev/blog/go1.27 Comments URL: https://news.ycombinator.com/item?id=49365405 Points: 328 # Comments: 62

Hacker News6d ago

Running Local LLMs on the Arduino® UNO™ Q Board: a Practical Guide

Learn how to run local LLMs on Arduino UNO Q — from model selection and quantization to memory constraints and real-world edge AI use cases. The post Running Local LLMs on the Arduino® UNO™ Q Board: a Practical Guide appeared first on EE Times.

EE Times6d ago

Sponsored content

AI crawlers are pouring into e-commerce, and the security line of defense is shifting from "interception" to "judgment"

AI is entering e-commerce, but it not only brings new shopping guides, searches, and transactions, it is also changing the security issues faced by e-commerce websites. Recently, when interpreting Akamai's latest "Internet Security Situation Report", Akamai's Senior Solutions Manager for Greater China Ma Jun mentioned that in 2025, more than 17 trillion bots will visit e-commerce websites around the world, a year-on-year increase of 19%; of which the Asia-Pacific region will grow by more than 63%, making it the fastest growing region. At the same time, API attacks increased by 9%, surpassing traditional web application attacks for the first time and becoming the report statistics...

InfoQ CN6d ago

安全

Digital signal processing pioneer Liu Bizhi died at the age of 91

IEEE Spectrum published an article in memory of digital signal processing pioneer Liu Bizhi, who passed away in May this year. Liu is widely regarded as one of the founders of the modern field of digital signal processing, whose research drove the transition from analog to digital in sound, image and video processing. Although little known outside engineering circles, his work is part of technology used by billions of people around the world. The low-power digital signal processors that make cell phone calls, streaming video and Internet communications possible trace their roots back to his work in 19...

Solidot6d ago

Eight global mainstream agents tested on Wall Street: Qianwen Office ranked first in overall ranking, surpassing Claude Cowork and Codex

According to news on August 19, analysts from Wall Street investment bank Jefferies recently conducted actual tests on eight global mainstream AI Agents for real office tasks. The results showed that through the collaboration of models and Harness, Alibaba Qianwen Office ranked first in overall score, surpassing Agent tools such as Claude Cowork and Codex in the United States. This test set up a total of 5 real office tasks, including completing the company's annual report summary based on multiple documents, searching and comparing company operating data online, operating a real desktop browser to complete information retrieval and document generation, and producing English based on the data...

Leiphone6d ago

业界

The Future Of AI Compute Won’t Run On Just One Kind Of Chip

Power, token cost, interconnects, and software orchestration are pushing data centers toward heterogeneous clusters built from CPUs, GPUs, NPUs, optics, and custom accelerators. The post The Future Of AI Compute Won’t Run On Just One Kind Of Chip appeared first on Semiconductor Engineering.

Semiconductor Engineering6d ago

Low Power-High PerformanceRound TablesTop Stories

IJCAI-ECAI 2026 opens: Wu Jiajun won the "Computer and Thought Award", the SMOTE algorithm won the first Time Test Award

The Classic Award pays tribute to the AI ​​pioneers, 14 young scholars stand on the new focus stage, and 713 main conference papers provide a pulse on the future evolution. Author丨Cen Feng Ma Xiaoning On August 18, 2026 local time, the 35th International Joint Conference on Artificial Intelligence (IJCAI-ECAI 2026) kicked off in Bremen, Germany. As one of the oldest and most authoritative top academic conferences on artificial intelligence, IJCAI has witnessed the ups and downs and prosperity of AI for more than half a century. This year, the conference returns to Europe, not only bringing together the world's top AI scholars and industry leaders...

LeiphoneAug 18, 2026

顶会专区

Beyond Consensus: Fragmentation of AI Policy in the Linux Ecosystem

As Large Language Models (LLMs) redefine the boundaries of software development, the Linux ecosystem is facing a complex intersection of legal risks, technical integrity, and philosophical purity. Projects are developing policies that vary: from GCC's restrictive "hardline" stance due to legal security and compiler accuracy concerns, to pragmatism led by Linux kernel maintainers that places a greater emphasis on personal responsibility and in-depth understanding. Meanwhile, the Kubernetes community has adopted a pragmatic model based on disclosure...

InfoQ CNAug 18, 2026

开源

How 35 percent of US employees are left on the margins

In a new book, Paul Osterman details his research finding more than one-third of the country’s workers are “disposable,” with little security or chance of job advancement.

MIT News ResearchAug 18, 2026

Books and authorsFacultyResearch

Wearable Sensors and AI Could Monitor Blood Pressure in ICU

Wearable sensors coupled with AI can generate blood pressure readings nearly as well as the invasive and risky arterial lines relied on in intensive care units and operating rooms, new work by Johns Hopkins University researchers demonstrates. The post Wearable Sensors and AI Could Monitor Blood Pressure in ICU appeared first on Semiconductor Digest.

Semiconductor DigestAug 18, 2026

AIMEMS

UMass Amherst Engineers Make Edge AI More Efficient by Redesigning Both Algorithm and Hardware

System leverages hyperdimensional computing algorithms and analog in-memory computing hardware to achieve quick, accurate processing. The post UMass Amherst Engineers Make Edge AI More Efficient by Redesigning Both Algorithm and Hardware appeared first on Semiconductor Digest.

Semiconductor DigestAug 17, 2026

AI

DNA shaper steers nervous system development

A study of worms finds a protein that helps shape the structure of the genome is critical for establishing the identity of some neurons.

MIT News ResearchAug 17, 2026

ResearchGeneticsDNA

Amazon hikes hardware prices by 60%, blaming memory shortage

As the memory shortage continues to cause trouble for hardware makers, Amazon says it is now being forced to pass on the costs to its consumers.

TechCrunch1d ago

HardwareAIAmazon

RISC-V at Sixteen: From Modular ISA to Standardized Platforms at Hot Chips 2026

RISC-V has evolved from a Berkeley research architecture into a global instruction-set standard spanning embedded controllers, application processors, accelerators, and emerging servers. Its defining technical characteristic is not merely openness, but disciplined modularity. Four ratified base ISAs—RV32I, RV64I,… Read More The post RISC-V at Sixteen: From Modular ISA to Standardized Platforms at Hot Chips 2026 appeared first on SemiWiki.

Semiconductor Wiki1d ago

IPRISC-VSiFive

微软删除了逾 17 万非营利组织的数据

微软曾从 2013 年起向全世界的小型非营利组织免费提供 Microsoft 365 Business Premium，但在 2025 年初它宣布将从 2025 年 7 月起停止提供免费授权，转为提供折扣价付费订阅。6 月 11 日前如果没有转为付费的账号内相关数据将被删除。根据 Slate 的报道，没有注意到微软邮件通知的非营利组织遭到重创，17.1 万非营利组织储存在 OneDrive 中的数据被删除，全部丢失。一名经营着一家儿童...

Solidot1d ago

Wi-Fi 8 专注于提升可靠性

从 2009 年的 Wi-Fi 4 起，每一代 Wi-Fi 的一大卖点都是相比前一代数据速率提升多少。Wi-Fi 5 的最大数据速率十倍于 Wi-Fi 4，到 Wi-Fi 7 每频段理论最大吞吐量达到了 23Gbps，对于大部分用户的网速是绰绰有余了。也许是时候放慢速度了。正在开发中的 Wi-Fi 8 将与 Wi-Fi 7 维持基本相同的最大数据速率、支持相同数量的空间流，使用相同的 4096-QAM（4K-QAM）调制，工作在相同的频段，支...

Solidot1d ago

Anthropic 最强模型难以吸引用户

Anthropic 的美国客户正使用其最强大 AI 工具的更廉价替代品，引发了外界对其高投入商业模式的质疑，而市场普遍预计 Anthropic 即将进行史上规模最大的 IPO。根据支付服务集团 Ramp 收集的 7 万家公司支出数据，Anthropic 规模最大、成本最高的模型 Fable 5 发布两个多月后，其支出占该公司各类工具总支出的比例仍仅约为 11%，目前已趋于稳定。这打破了企业用户默认选择最强大模...

Solidot1d ago

Article: Rightsizing Platform Engineering: Building the Platform Your Organization Actually Needs

Shift-left and DevOps have impacted how we flow changes from inception to production, but at the cost of increased cognitive load and duplication of effort across testing, security, and maintenance. This article explores the real-world challenges of rightsizing developer platforms and finding a cultural match for engineering teams who use them to reduce cognitive load and deliver change faster. By John Keates

InfoQ EN1d ago

Artifacts & ToolsCoding StandardsArchitecture

I gave Qwen 3.8 27B a reverse-engineering job and it finished in 30 minutes

Article URL: https://www.xda-developers.com/qwen-3-8-27b-reverse-engineering-job-frontier-model/ Comments URL: https://news.ycombinator.com/item?id=49407507 Points: 159 # Comments: 80

Hacker News2d ago

Knowing When to Stop: The Art of Making a Loop Converge

Article URL: https://a16z.com/knowing-when-to-stop-the-art-of-making-a-loop-converge/ Comments URL: https://news.ycombinator.com/item?id=49403484 Points: 15 # Comments: 4

Hacker News3d ago

Why your local LLM feels dumber than it is

Article URL: https://forum.level1techs.com/t/why-your-local-llm-feels-dumber-than-it-is/253917 Comments URL: https://news.ycombinator.com/item?id=49402232 Points: 60 # Comments: 17

Hacker News3d ago

New MCP Roadmap

Article URL: https://blog.modelcontextprotocol.io/posts/mcp-roadmap/ Comments URL: https://news.ycombinator.com/item?id=49399591 Points: 158 # Comments: 119

Hacker News3d ago

Cloudflare 推出 Cache Response Rules，在源站响应后进一步控制缓存

Cloudflare 近日推出缓存响应规则（Cache Response Rules"），这是一套新的规则引擎，运行在源站返回响应之后、内容写入 Cloudflare 缓存之前。此前，缓存规则（Cache Rules）只能根据请求属性进行判断；缓存响应规则则新增了一个响应处理阶段，可以在响应进入缓存之前检查源站返回的内容。根据官方文档，缓存响应规则"允许用户修改 Cache-Control 指令、管理 Cache Tag，以及移除 Set-Coo...

InfoQ CN3d ago

服务革新

Cloudflare 推出 Agent Tracing：支持截断限制，不同框架的 Payload 默认记录策略存在差异

Cloudflare 推出了 Cloudflare Agents 的首个组件，Agent Tracing"（Agent 追踪），可提供用于集中查看已经部署的 Agent 会话的统一 Dashboard。此次发布在现有 Workers Tracing 的基础上增加了 Agent 级别的 Span，同时也伴随着明确的收费时间表：Beta 期间免费，2026 年 10 月 1 日起将纳入 Workers Observability 的计费体系。这一功能真正要解决的问题值得关注。一个 Agent 即使返回 HTTP 200，也不意味着它...

InfoQ CN3d ago

可观测

JetBrains详细阐述了控制AI支出快速增长的首批举措

JetBrains在一篇博客文章中介绍"，在开发相关支出在六个月内大约增长十倍后，公司如何开始集中管理AI的使用。公司没有限制工程师只能使用少数已批准的工具，而是构建了一个共享的访问与记账层，其目标是在保留工具选择的同时，为团队提供更高的可见性与控制能力。随着AI的使用在公司内部加速增长，变化逐步显现了出来。JetBrains表示，大多数开发者每月会使用3到5个A...

InfoQ CN3d ago

FinOps

How TSMC Is Wiring the AI Era With Light

TSMC’s photonics strategy is centered on integrating optical input-output with advanced logic, rather than selling conventional optical transceivers as standalone products. The company is developing a silicon-photonics foundry platform and a packaging architecture called TSMC-COUPE, or Compact Universal Photonic … Read More The post How TSMC Is Wiring the AI Era With Light appeared first on SemiWiki.

Semiconductor Wiki4d ago

3D ICAIChiplet

The world's first complete game of autonomous table tennis by a humanoid robot was unveiled at the 2026 World Robot Conference, and Super Dimension Power KAI's full-stack embodied intelligent hard-...

On August 19, the 2026 World Robot Conference with the theme of "human-machine symbiosis, integration of production and demand" officially opened in Beijing. As a full-stack embodied intelligence large-scale model company, Super Dimension Power brings the world's first full-stack autonomous table tennis game results of humanoid robots, SMASH 2.0 high-dynamic humanoid table tennis system, KAI world model, the world's highest 117 full-body degrees of freedom KAIBot highly anthropomorphic body, KAI Hand ultra-high degree of freedom dexterous hand and KAI Halo first-view digital acquisition headband and other full-stack matrices. It will bring the audience a "visible and playable" embodied intelligent hardware...

Leiphone4d ago

机器人

China’s NAND Specialist YMTC Moves Closer to IPO

YMTC must raise capital to explore demand for AI-driven memory while balancing domestic and overseas markets. The post China’s NAND Specialist YMTC Moves Closer to IPO appeared first on EE Times.

EE Times4d ago

News3D NANDAI and Big Data

The embodied bosses "quarreled" over these issues

Who has the final say in the embodied GPT moment?

TMTpost4d ago

Bilibili enters the international market

Bilibili re-released the international version of its application this week and is preparing to launch an English version to enter the global market. The new international version of the app will not require identity verification and users will not need to provide a passport or ID to sign up. Bilibili has previously actively invited well-known Western anchors such as MrBeast to publish videos on its Chinese main website, which has 376 million monthly active users. A larger global expansion could challenge YouTube's dominance, but also faces thorny issues like TikTok's censorship, content moderation and data security. According to the recruitment information, Station B...

Solidot4d ago

Scientists Release Biggest 2D Map of the Universe

Article URL: https://newscenter.lbl.gov/2026/08/10/scientists-release-biggest-2d-map-of-the-universe/ Comments URL: https://news.ycombinator.com/item?id=49380671 Points: 3 # Comments: 0

Hacker News4d ago

OK, can we actually cool data centers with our pee?

Jason Kelce joked that people should cool data centers with their pee, rather than potable water -- but his suggestion is not completely ludicrous.

TechCrunch5d ago

AIdata centersExclusive

Linux 7.2

Article URL: https://www.igalia.com/2026/08/19/Linux-72-Released.html Comments URL: https://news.ycombinator.com/item?id=49376265 Points: 163 # Comments: 54

Hacker News5d ago

Hacking with Claude on a $27 smart watch

Article URL: https://www.mikekasberg.com/blog/2026/08/19/hacking-with-claude-on-a-27-smart-watch.html Comments URL: https://news.ycombinator.com/item?id=49374772 Points: 71 # Comments: 43

Hacker News5d ago

Next.js 16.3: Instant Navigations, Up to 90% Less Dev Memory and Faster Builds

Vercel has released Next.js 16.3, featuring significant updates since version 16.0. Enhancements include reduced memory usage during development, accelerated build times, and improved type checking. Instant Navigations introduces faster, client-like responses while maintaining server-rendered architecture. Developers are advised to gradually adopt new features due to noted caveats. By Daniel Curtis

InfoQ EN5d ago

JavaScriptRoutingTypeScript

Comprehensibility as an architectural property: Systems that cannot be understood cannot evolve safely

This article was written by participants of the InfoQ Certified Architect Online event. It is a culmination of their work and reflects the participants’ commitment to AI Collective insights gained at the intersection with modern software architecture. Comprehensibility Problems Teams with extensive experience in complex systems have experienced situations where they have to deal with production failures where most of the time is spent figuring out how the various parts of the system connect and what the system actually does, rather than finding bugs.

InfoQ CN5d ago

业务架构

Go 1.27

Article URL: https://go.dev/blog/go1.27 Comments URL: https://news.ycombinator.com/item?id=49365405 Points: 328 # Comments: 62

Hacker News6d ago

AI crawlers are pouring into e-commerce, and the security line of defense is shifting from "interception" to "judgment"

AI is entering e-commerce, but it not only brings new shopping guides, searches, and transactions, it is also changing the security issues faced by e-commerce websites. Recently, when interpreting Akamai's latest "Internet Security Situation Report", Akamai's Senior Solutions Manager for Greater China Ma Jun mentioned that in 2025, more than 17 trillion bots will visit e-commerce websites around the world, a year-on-year increase of 19%; of which the Asia-Pacific region will grow by more than 63%, making it the fastest growing region. At the same time, API attacks increased by 9%, surpassing traditional web application attacks for the first time and becoming the report statistics...

InfoQ CN6d ago

安全

Eight global mainstream agents tested on Wall Street: Qianwen Office ranked first in overall ranking, surpassing Claude Cowork and Codex

According to news on August 19, analysts from Wall Street investment bank Jefferies recently conducted actual tests on eight global mainstream AI Agents for real office tasks. The results showed that through the collaboration of models and Harness, Alibaba Qianwen Office ranked first in overall score, surpassing Agent tools such as Claude Cowork and Codex in the United States. This test set up a total of 5 real office tasks, including completing the company's annual report summary based on multiple documents, searching and comparing company operating data online, operating a real desktop browser to complete information retrieval and document generation, and producing English based on the data...

Leiphone6d ago

业界

IJCAI-ECAI 2026 opens: Wu Jiajun won the "Computer and Thought Award", the SMOTE algorithm won the first Time Test Award

The Classic Award pays tribute to the AI ​​pioneers, 14 young scholars stand on the new focus stage, and 713 main conference papers provide a pulse on the future evolution. Author丨Cen Feng Ma Xiaoning On August 18, 2026 local time, the 35th International Joint Conference on Artificial Intelligence (IJCAI-ECAI 2026) kicked off in Bremen, Germany. As one of the oldest and most authoritative top academic conferences on artificial intelligence, IJCAI has witnessed the ups and downs and prosperity of AI for more than half a century. This year, the conference returns to Europe, not only bringing together the world's top AI scholars and industry leaders...

LeiphoneAug 18, 2026

顶会专区

How 35 percent of US employees are left on the margins

In a new book, Paul Osterman details his research finding more than one-third of the country’s workers are “disposable,” with little security or chance of job advancement.

MIT News ResearchAug 18, 2026

Books and authorsFacultyResearch

UMass Amherst Engineers Make Edge AI More Efficient by Redesigning Both Algorithm and Hardware

System leverages hyperdimensional computing algorithms and analog in-memory computing hardware to achieve quick, accurate processing. The post UMass Amherst Engineers Make Edge AI More Efficient by Redesigning Both Algorithm and Hardware appeared first on Semiconductor Digest.

Semiconductor DigestAug 17, 2026

AI

LLMs could control their host machines by exploiting inference engines

Article URL: https://boydkane.com/essays/llms-could-control-their-host-machines-by-exploiting-inference-engines Comments URL: https://news.ycombinator.com/item?id=49424387 Points: 56 # Comments: 25

Hacker News1d ago

Hot Chips: Evolving Memory Architectures for Artificial Intelligence

Artificial intelligence performance is increasingly constrained by memory rather than arithmetic throughput. Scaling laws show that model quality improves when parameter count, training data, and compute grow together, but this balance fails when processors cannot obtain operands quickly enough. Contemporary accelerators… Read More The post Hot Chips: Evolving Memory Architectures for Artificial Intelligence appeared first on SemiWiki.

Semiconductor Wiki1d ago

AISemiconductorartificial intelligence

Hugging Face reportedly in talks to be acquired for $13B

Hugging Face has reportedly been fielding acquisition offers that would value the company at around $13B. But with the founders' feeling of responsibility to community, doubts arise as to whether a sale will happen.

TechCrunch1d ago

AIVentureHugging Face

第一位“95后”获奖人产生、女性获奖人首次达两位数——第八届“科学探索奖”公布获奖名单

8月24日，第八届“科学探索奖”获奖名单正式揭晓，50位青年科学家榜上有名。第一位“95后”获奖人产生、女性科学家人数首次达到两位数、7家机构首次拥有获奖人。作为目前国内金额最高的青年科技人才资助项目之一，“科学探索奖”聚焦基础科学和前沿技术领域，每位获奖人将在5年内获得总计300万元人民币奖金，并且可自由支配。支持基础科研是腾讯可持续社会价值...

Leiphone1d ago

业界

Building Technology People Can Trust

This article is brought to you by Emerson.I’ve spent much of my career as an engineer, including years in the semiconductor industry. And one lesson has stayed with me through every major technology shift: innovation always creates new complexity.In semiconductors, we have seen that repeatedly. Every generation has delivered breakthroughs in performance and capability, but each step forward made it harder to understand system behavior. What used to be easy to validate on the component level...

IEEE Spectrum Semiconductors1d ago

AiAutomationTest

Iranian hackers shut down UK power plant for 4 days

Article URL: https://www.telegraph.co.uk/news/2026/08/22/iranian-hackers-shut-down-uk-power-plant/ Comments URL: https://news.ycombinator.com/item?id=49407509 Points: 56 # Comments: 34

Hacker News2d ago

AI 即将接管流量，而广告正在疯狂「末日反扑」

我们在腌臜的广告泥潭中，畅想着美好的 AGI 未来。#欢迎关注爱范儿官方微信公众号：爱范儿（微信号：ifanr），更多精彩内容第一时间为您奉上。

Ifanr2d ago

爱范儿AIAI 手机

Fast and Hard Code

Article URL: https://lucumr.pocoo.org/2026/8/22/fast-hard-code/ Comments URL: https://news.ycombinator.com/item?id=49403228 Points: 14 # Comments: 4

Hacker News3d ago

Amazon just hiked the prices for Echo, Fire TV, and Kindle products by up to 60 percent

Citing "significant increases in memory and storage component costs," Amazon has raised prices on Echos, Kindles, Fire TVs, and Eeros by up to 60 percent, as first reported by Fortune. The cheapest products in Amazon's lineup were hit hardest, with the Echo Dot smart speaker jumping from $49.99 to $79.99. The Fire TV Stick 4K […]

The Verge Tech3d ago

AmazonAmazon AlexaNews

WRC 2026 风向标：具身智能下半场，比「大脑」更比「账本」

展台不再比谁花活多，开始比谁能稳定干满8小时。 作者丨邓哲敏 齐铖湧 编辑丨齐铖湧 上海 WAIC 结束不到一个月，北京 WRC 2026 接棒开启。短短几十天，谁也不可能把模型推倒重来。想看到什么颠覆性突破，是不现实的。所以这篇现场观察不单看 WRC 上的展台花活，AI科技评论（雷峰网公众号）会结合几个月来各家综合表现，看“具身大脑”究竟走到哪一步。先说首日逛展的...

Leiphone3d ago

人工智能

AI Code Review at Scale: LinkedIn's Multi-Agent Approach

At LinkedIn's scale, relying solely on human reviewers or simply putting an off-the-shelf AI reviewer in front of GitHub is not an effective way to manage PRs. To address this, LinkedIn engineers built a multi-agent AI code review platform that understands the organization’s coding context, treats code review as production infrastructure, and minimizes hallucinations and low-signal feedback. By Sergio De Simone

InfoQ EN3d ago

AgentsKubernetesCode Reviews

npm 终于动手了：默认封杀 postinstall 脚本

由 GitHub 维护并随 Node.js 一同提供的软件包管理器 npm 发布了 npm 12"。该版本更改了 npm install 的一系列安全相关默认行为，同时弃用可以绕过双因素认证的细粒度访问令牌。npm 12 将此前自动执行的三种安装行为改为需要开发者明确选择启用。这些更改于今年 6 月首次公布"，并从 npm 11.16.0 开始通过警告形式提供，以便团队在升级前做好准备。最重要的变化是 allowScripts 现在默...

InfoQ CN3d ago

微软安全

Rust Glancer: Rust LSP using 100x less RAM

https://matklad.github.io/2026/08/21/rust-glancer.html Comments URL: https://news.ycombinator.com/item?id=49393052 Points: 387 # Comments: 92

Hacker News4d ago

Nvidia finds that simple linear math can replace costly AI model handoffs

When an agentic AI system hands a task from a small model to a larger one — or back down again — it pays a steep tax: the receiving model has to recompute the entire conversation from scratch, driving up compute costs and latency. This is a major bottleneck for enterprises building long-horizon, multi-LLM workflows.To solve this challenge, researchers at Nvidia have introduced a cross-model KV cache transfer technique that directly maps the prefilled KV cache from a source model into the ...

VentureBeat4d ago

Technology

Cloudflare Cuts Astro Github Issues by 85% with AI Agents

Cloudflare, Astro, AI agents, GitHub Actions, issue triage, agentic AI, software architecture, open source, developer tools, AI automation, automated testing, human in the loop, agent workflows, GitHub, software engineering, AI software development, bug triage, continuous integration, developer productivity, autonomous agents, AI coding, Cloudflare Workers, Flue, triagebot By Leela Kumili

InfoQ EN4d ago

AutomationIssue Tracking ToolsWorkflow / BPM

Cloudflare Turns Engineering Standards Into an AI-Enforced Control System

Cloudflare has recently detailed how it is using AI to transform internal engineering standards from passive documentation into an actively enforced control system across the software development lifecycle. By Craig Risi

InfoQ EN4d ago

CloudflareSite Reliability EngineeringArtificial Intelligence

The Human Brain Versus AI: Similar Results, Very Different Machines

Contrast 20 watts with a megawatt: The brain and the LLM aren’t in the same race. The post The Human Brain Versus AI: Similar Results, Very Different Machines appeared first on EE Times.

EE Times4d ago

BlogOpinionAI and Big Data

After reading Alibaba’s financial report, I discovered that the more intense AI is, the more it will “fatten” the shovel sellers

After reading Alibaba’s latest financial report last night, I don’t want to admit it, but I have to admit a fact: Although the market spotlight is on the model layer and application layer, in the short term, the strongest commercialization of AI is in the cloud and chips. Moreover, the more aggressively the upper-layer models are used, the more profitable the bottom-level shovel sellers will be. The latest financial report shows that Alibaba Cloud's revenue in the first quarter of fiscal year 2027 (second quarter of 2026) reached 48.737 billion yuan, a year-on-year growth of 45%. More important is revenue guidance for the coming quarters. During the earnings call, Wu...

Leiphone4d ago

企业服务

Agentrys Shows You How to Build a Multi-Agent System in 30 Minutes

Agentrys believes that a purpose-built multi-agent system for chip design shouldn’t be a quarter-long project. Agentic AI systems are becoming quite popular to cut design time and increase quality for complex chip design projects. The challenge is finding the right agentic flow for a particular application. There are… Read More The post Agentrys Shows You How to Build a Multi-Agent System in 30 Minutes appeared first on SemiWiki.

Semiconductor Wiki5d ago

AgentrysAIEDA

One in five enterprises can't stop a runaway AI agent's spending in real time

Enterprise AI teams have stopped betting on a single orchestration platform. The median enterprise now runs three at once — not by accident, but because none of them fully trusts a single vendor to run the show, according to VB Pulse data.This is not just to avoid vendor lock-in and retain flexibility (although that’s a big part of it). There’s still a lot of uncertainty, even distrust, in vendors’ security and permissioning capabilities. Enterprises want the ability to impose their o...

VentureBeat5d ago

Orchestration

Amazon’s drone deliveries are landing in pools and ponds

Amazon's speedy drone delivery service will soon reach 500 cities across the US - but that might just mean there are more pools to drop packages into. On Wednesday, ABC7 News Bay Area shared a video showing an Amazon delivery drone hovering over a customer's pool in Texas, before opening its hatch and plopping the […]

The Verge Tech5d ago

AmazonDronesNews

Welcome to the AI crisis in math

Today on Decoder, I’m talking with Robert Hart, The Verge’s London-based AI reporter, about what AI is doing to the field of mathematics and the existential crisis many lead mathematicians are having about it. OpenAI just published a set of solutions to longstanding problems in math that went off like a bombshell in the field. […]

The Verge Tech5d ago

AIDecoderOpenAI

GitHub announces the cause of this week's eight-hour outage

GitHub, the largest code hosting platform, experienced an outage that lasted for nearly eight hours this week, once again triggering discussions among developers about finding alternative platforms. GitHub has experienced frequent outages this year, which has prompted many well-known open source projects to announce their migration. This week's outage began at 13:28 UTC on August 17 and was not fully resolved until 21:15 UTC - the incident lasted 7 hours and 47 minutes, causing a large number of errors in services such as Issues, Pull Requests, APIs, Actions, and Copilot. GitHub explained that the direct cause of the accident...

Solidot5d ago

The why behind sex

Research details answer to a question that’s more complex than it seems

Harvard Tech News6d ago

Science & TechFacultyResearch

Running Local LLMs on the Arduino® UNO™ Q Board: a Practical Guide

Learn how to run local LLMs on Arduino UNO Q — from model selection and quantization to memory constraints and real-world edge AI use cases. The post Running Local LLMs on the Arduino® UNO™ Q Board: a Practical Guide appeared first on EE Times.

EE Times6d ago

Sponsored content

Digital signal processing pioneer Liu Bizhi died at the age of 91

IEEE Spectrum published an article in memory of digital signal processing pioneer Liu Bizhi, who passed away in May this year. Liu is widely regarded as one of the founders of the modern field of digital signal processing, whose research drove the transition from analog to digital in sound, image and video processing. Although little known outside engineering circles, his work is part of technology used by billions of people around the world. The low-power digital signal processors that make cell phone calls, streaming video and Internet communications possible trace their roots back to his work in 19...

Solidot6d ago

The Future Of AI Compute Won’t Run On Just One Kind Of Chip

Power, token cost, interconnects, and software orchestration are pushing data centers toward heterogeneous clusters built from CPUs, GPUs, NPUs, optics, and custom accelerators. The post The Future Of AI Compute Won’t Run On Just One Kind Of Chip appeared first on Semiconductor Engineering.

Semiconductor Engineering6d ago

Low Power-High PerformanceRound TablesTop Stories

Beyond Consensus: Fragmentation of AI Policy in the Linux Ecosystem

As Large Language Models (LLMs) redefine the boundaries of software development, the Linux ecosystem is facing a complex intersection of legal risks, technical integrity, and philosophical purity. Projects are developing policies that vary: from GCC's restrictive "hardline" stance due to legal security and compiler accuracy concerns, to pragmatism led by Linux kernel maintainers that places a greater emphasis on personal responsibility and in-depth understanding. Meanwhile, the Kubernetes community has adopted a pragmatic model based on disclosure...

InfoQ CNAug 18, 2026

开源

Wearable Sensors and AI Could Monitor Blood Pressure in ICU

Wearable sensors coupled with AI can generate blood pressure readings nearly as well as the invasive and risky arterial lines relied on in intensive care units and operating rooms, new work by Johns Hopkins University researchers demonstrates. The post Wearable Sensors and AI Could Monitor Blood Pressure in ICU appeared first on Semiconductor Digest.

Semiconductor DigestAug 18, 2026

AIMEMS

DNA shaper steers nervous system development

A study of worms finds a protein that helps shape the structure of the genome is critical for establishing the identity of some neurons.

MIT News ResearchAug 17, 2026

ResearchGeneticsDNA

Amazon hikes hardware prices by 60%, blaming memory shortage

As the memory shortage continues to cause trouble for hardware makers, Amazon says it is now being forced to pass on the costs to its consumers.

TechCrunch1d ago

HardwareAIAmazon

Hot Chips: Evolving Memory Architectures for Artificial Intelligence

Artificial intelligence performance is increasingly constrained by memory rather than arithmetic throughput. Scaling laws show that model quality improves when parameter count, training data, and compute grow together, but this balance fails when processors cannot obtain operands quickly enough. Contemporary accelerators… Read More The post Hot Chips: Evolving Memory Architectures for Artificial Intelligence appeared first on SemiWiki.

Semiconductor Wiki1d ago

AISemiconductorartificial intelligence

Wi-Fi 8 专注于提升可靠性

从 2009 年的 Wi-Fi 4 起，每一代 Wi-Fi 的一大卖点都是相比前一代数据速率提升多少。Wi-Fi 5 的最大数据速率十倍于 Wi-Fi 4，到 Wi-Fi 7 每频段理论最大吞吐量达到了 23Gbps，对于大部分用户的网速是绰绰有余了。也许是时候放慢速度了。正在开发中的 Wi-Fi 8 将与 Wi-Fi 7 维持基本相同的最大数据速率、支持相同数量的空间流，使用相同的 4096-QAM（4K-QAM）调制，工作在相同的频段，支...

Solidot1d ago

Building Technology People Can Trust

This article is brought to you by Emerson.I’ve spent much of my career as an engineer, including years in the semiconductor industry. And one lesson has stayed with me through every major technology shift: innovation always creates new complexity.In semiconductors, we have seen that repeatedly. Every generation has delivered breakthroughs in performance and capability, but each step forward made it harder to understand system behavior. What used to be easy to validate on the component level...

IEEE Spectrum Semiconductors1d ago

AiAutomationTest

I gave Qwen 3.8 27B a reverse-engineering job and it finished in 30 minutes

Article URL: https://www.xda-developers.com/qwen-3-8-27b-reverse-engineering-job-frontier-model/ Comments URL: https://news.ycombinator.com/item?id=49407507 Points: 159 # Comments: 80

Hacker News2d ago

Fast and Hard Code

Article URL: https://lucumr.pocoo.org/2026/8/22/fast-hard-code/ Comments URL: https://news.ycombinator.com/item?id=49403228 Points: 14 # Comments: 4

Hacker News3d ago

New MCP Roadmap

Article URL: https://blog.modelcontextprotocol.io/posts/mcp-roadmap/ Comments URL: https://news.ycombinator.com/item?id=49399591 Points: 158 # Comments: 119

Hacker News3d ago

AI Code Review at Scale: LinkedIn's Multi-Agent Approach

At LinkedIn's scale, relying solely on human reviewers or simply putting an off-the-shelf AI reviewer in front of GitHub is not an effective way to manage PRs. To address this, LinkedIn engineers built a multi-agent AI code review platform that understands the organization’s coding context, treats code review as production infrastructure, and minimizes hallucinations and low-signal feedback. By Sergio De Simone

InfoQ EN3d ago

AgentsKubernetesCode Reviews

JetBrains详细阐述了控制AI支出快速增长的首批举措

JetBrains在一篇博客文章中介绍"，在开发相关支出在六个月内大约增长十倍后，公司如何开始集中管理AI的使用。公司没有限制工程师只能使用少数已批准的工具，而是构建了一个共享的访问与记账层，其目标是在保留工具选择的同时，为团队提供更高的可见性与控制能力。随着AI的使用在公司内部加速增长，变化逐步显现了出来。JetBrains表示，大多数开发者每月会使用3到5个A...

InfoQ CN3d ago

FinOps

Nvidia finds that simple linear math can replace costly AI model handoffs

When an agentic AI system hands a task from a small model to a larger one — or back down again — it pays a steep tax: the receiving model has to recompute the entire conversation from scratch, driving up compute costs and latency. This is a major bottleneck for enterprises building long-horizon, multi-LLM workflows.To solve this challenge, researchers at Nvidia have introduced a cross-model KV cache transfer technique that directly maps the prefilled KV cache from a source model into the ...

VentureBeat4d ago

Technology

China’s NAND Specialist YMTC Moves Closer to IPO

YMTC must raise capital to explore demand for AI-driven memory while balancing domestic and overseas markets. The post China’s NAND Specialist YMTC Moves Closer to IPO appeared first on EE Times.

EE Times4d ago

News3D NANDAI and Big Data

The Human Brain Versus AI: Similar Results, Very Different Machines

Contrast 20 watts with a megawatt: The brain and the LLM aren’t in the same race. The post The Human Brain Versus AI: Similar Results, Very Different Machines appeared first on EE Times.

EE Times4d ago

BlogOpinionAI and Big Data

Scientists Release Biggest 2D Map of the Universe

Article URL: https://newscenter.lbl.gov/2026/08/10/scientists-release-biggest-2d-map-of-the-universe/ Comments URL: https://news.ycombinator.com/item?id=49380671 Points: 3 # Comments: 0

Hacker News4d ago

One in five enterprises can't stop a runaway AI agent's spending in real time

Enterprise AI teams have stopped betting on a single orchestration platform. The median enterprise now runs three at once — not by accident, but because none of them fully trusts a single vendor to run the show, according to VB Pulse data.This is not just to avoid vendor lock-in and retain flexibility (although that’s a big part of it). There’s still a lot of uncertainty, even distrust, in vendors’ security and permissioning capabilities. Enterprises want the ability to impose their o...

VentureBeat5d ago

Orchestration

Hacking with Claude on a $27 smart watch

Article URL: https://www.mikekasberg.com/blog/2026/08/19/hacking-with-claude-on-a-27-smart-watch.html Comments URL: https://news.ycombinator.com/item?id=49374772 Points: 71 # Comments: 43

Hacker News5d ago

GitHub announces the cause of this week's eight-hour outage

GitHub, the largest code hosting platform, experienced an outage that lasted for nearly eight hours this week, once again triggering discussions among developers about finding alternative platforms. GitHub has experienced frequent outages this year, which has prompted many well-known open source projects to announce their migration. This week's outage began at 13:28 UTC on August 17 and was not fully resolved until 21:15 UTC - the incident lasted 7 hours and 47 minutes, causing a large number of errors in services such as Issues, Pull Requests, APIs, Actions, and Copilot. GitHub explained that the direct cause of the accident...

Solidot5d ago

Go 1.27

Article URL: https://go.dev/blog/go1.27 Comments URL: https://news.ycombinator.com/item?id=49365405 Points: 328 # Comments: 62

Hacker News6d ago

Digital signal processing pioneer Liu Bizhi died at the age of 91

IEEE Spectrum published an article in memory of digital signal processing pioneer Liu Bizhi, who passed away in May this year. Liu is widely regarded as one of the founders of the modern field of digital signal processing, whose research drove the transition from analog to digital in sound, image and video processing. Although little known outside engineering circles, his work is part of technology used by billions of people around the world. The low-power digital signal processors that make cell phone calls, streaming video and Internet communications possible trace their roots back to his work in 19...

Solidot6d ago

IJCAI-ECAI 2026 opens: Wu Jiajun won the "Computer and Thought Award", the SMOTE algorithm won the first Time Test Award

The Classic Award pays tribute to the AI ​​pioneers, 14 young scholars stand on the new focus stage, and 713 main conference papers provide a pulse on the future evolution. Author丨Cen Feng Ma Xiaoning On August 18, 2026 local time, the 35th International Joint Conference on Artificial Intelligence (IJCAI-ECAI 2026) kicked off in Bremen, Germany. As one of the oldest and most authoritative top academic conferences on artificial intelligence, IJCAI has witnessed the ups and downs and prosperity of AI for more than half a century. This year, the conference returns to Europe, not only bringing together the world's top AI scholars and industry leaders...

LeiphoneAug 18, 2026

顶会专区

Wearable Sensors and AI Could Monitor Blood Pressure in ICU

Wearable sensors coupled with AI can generate blood pressure readings nearly as well as the invasive and risky arterial lines relied on in intensive care units and operating rooms, new work by Johns Hopkins University researchers demonstrates. The post Wearable Sensors and AI Could Monitor Blood Pressure in ICU appeared first on Semiconductor Digest.

Semiconductor DigestAug 18, 2026

AIMEMS

LLMs could control their host machines by exploiting inference engines

Article URL: https://boydkane.com/essays/llms-could-control-their-host-machines-by-exploiting-inference-engines Comments URL: https://news.ycombinator.com/item?id=49424387 Points: 56 # Comments: 25

Hacker News1d ago

微软删除了逾 17 万非营利组织的数据

微软曾从 2013 年起向全世界的小型非营利组织免费提供 Microsoft 365 Business Premium，但在 2025 年初它宣布将从 2025 年 7 月起停止提供免费授权，转为提供折扣价付费订阅。6 月 11 日前如果没有转为付费的账号内相关数据将被删除。根据 Slate 的报道，没有注意到微软邮件通知的非营利组织遭到重创，17.1 万非营利组织储存在 OneDrive 中的数据被删除，全部丢失。一名经营着一家儿童...

Solidot1d ago

第一位“95后”获奖人产生、女性获奖人首次达两位数——第八届“科学探索奖”公布获奖名单

8月24日，第八届“科学探索奖”获奖名单正式揭晓，50位青年科学家榜上有名。第一位“95后”获奖人产生、女性科学家人数首次达到两位数、7家机构首次拥有获奖人。作为目前国内金额最高的青年科技人才资助项目之一，“科学探索奖”聚焦基础科学和前沿技术领域，每位获奖人将在5年内获得总计300万元人民币奖金，并且可自由支配。支持基础科研是腾讯可持续社会价值...

Leiphone1d ago

业界

Article: Rightsizing Platform Engineering: Building the Platform Your Organization Actually Needs

Shift-left and DevOps have impacted how we flow changes from inception to production, but at the cost of increased cognitive load and duplication of effort across testing, security, and maintenance. This article explores the real-world challenges of rightsizing developer platforms and finding a cultural match for engineering teams who use them to reduce cognitive load and deliver change faster. By John Keates

InfoQ EN1d ago

Artifacts & ToolsCoding StandardsArchitecture

AI 即将接管流量，而广告正在疯狂「末日反扑」

我们在腌臜的广告泥潭中，畅想着美好的 AGI 未来。#欢迎关注爱范儿官方微信公众号：爱范儿（微信号：ifanr），更多精彩内容第一时间为您奉上。

Ifanr2d ago

爱范儿AIAI 手机

Why your local LLM feels dumber than it is

Article URL: https://forum.level1techs.com/t/why-your-local-llm-feels-dumber-than-it-is/253917 Comments URL: https://news.ycombinator.com/item?id=49402232 Points: 60 # Comments: 17

Hacker News3d ago

WRC 2026 风向标：具身智能下半场，比「大脑」更比「账本」

展台不再比谁花活多，开始比谁能稳定干满8小时。 作者丨邓哲敏 齐铖湧 编辑丨齐铖湧 上海 WAIC 结束不到一个月，北京 WRC 2026 接棒开启。短短几十天，谁也不可能把模型推倒重来。想看到什么颠覆性突破，是不现实的。所以这篇现场观察不单看 WRC 上的展台花活，AI科技评论（雷峰网公众号）会结合几个月来各家综合表现，看“具身大脑”究竟走到哪一步。先说首日逛展的...

Leiphone3d ago

人工智能

Cloudflare 推出 Agent Tracing：支持截断限制，不同框架的 Payload 默认记录策略存在差异

Cloudflare 推出了 Cloudflare Agents 的首个组件，Agent Tracing"（Agent 追踪），可提供用于集中查看已经部署的 Agent 会话的统一 Dashboard。此次发布在现有 Workers Tracing 的基础上增加了 Agent 级别的 Span，同时也伴随着明确的收费时间表：Beta 期间免费，2026 年 10 月 1 日起将纳入 Workers Observability 的计费体系。这一功能真正要解决的问题值得关注。一个 Agent 即使返回 HTTP 200，也不意味着它...

InfoQ CN3d ago

可观测

Rust Glancer: Rust LSP using 100x less RAM

https://matklad.github.io/2026/08/21/rust-glancer.html Comments URL: https://news.ycombinator.com/item?id=49393052 Points: 387 # Comments: 92

Hacker News4d ago

The world's first complete game of autonomous table tennis by a humanoid robot was unveiled at the 2026 World Robot Conference, and Super Dimension Power KAI's full-stack embodied intelligent hard-...

On August 19, the 2026 World Robot Conference with the theme of "human-machine symbiosis, integration of production and demand" officially opened in Beijing. As a full-stack embodied intelligence large-scale model company, Super Dimension Power brings the world's first full-stack autonomous table tennis game results of humanoid robots, SMASH 2.0 high-dynamic humanoid table tennis system, KAI world model, the world's highest 117 full-body degrees of freedom KAIBot highly anthropomorphic body, KAI Hand ultra-high degree of freedom dexterous hand and KAI Halo first-view digital acquisition headband and other full-stack matrices. It will bring the audience a "visible and playable" embodied intelligent hardware...

Leiphone4d ago

机器人

Cloudflare Turns Engineering Standards Into an AI-Enforced Control System

Cloudflare has recently detailed how it is using AI to transform internal engineering standards from passive documentation into an actively enforced control system across the software development lifecycle. By Craig Risi

InfoQ EN4d ago

CloudflareSite Reliability EngineeringArtificial Intelligence

Bilibili enters the international market

Bilibili re-released the international version of its application this week and is preparing to launch an English version to enter the global market. The new international version of the app will not require identity verification and users will not need to provide a passport or ID to sign up. Bilibili has previously actively invited well-known Western anchors such as MrBeast to publish videos on its Chinese main website, which has 376 million monthly active users. A larger global expansion could challenge YouTube's dominance, but also faces thorny issues like TikTok's censorship, content moderation and data security. According to the recruitment information, Station B...

Solidot4d ago

Agentrys Shows You How to Build a Multi-Agent System in 30 Minutes

Agentrys believes that a purpose-built multi-agent system for chip design shouldn’t be a quarter-long project. Agentic AI systems are becoming quite popular to cut design time and increase quality for complex chip design projects. The challenge is finding the right agentic flow for a particular application. There are… Read More The post Agentrys Shows You How to Build a Multi-Agent System in 30 Minutes appeared first on SemiWiki.

Semiconductor Wiki5d ago

AgentrysAIEDA

Linux 7.2

Article URL: https://www.igalia.com/2026/08/19/Linux-72-Released.html Comments URL: https://news.ycombinator.com/item?id=49376265 Points: 163 # Comments: 54

Hacker News5d ago

Welcome to the AI crisis in math

Today on Decoder, I’m talking with Robert Hart, The Verge’s London-based AI reporter, about what AI is doing to the field of mathematics and the existential crisis many lead mathematicians are having about it. OpenAI just published a set of solutions to longstanding problems in math that went off like a bombshell in the field. […]

The Verge Tech5d ago

AIDecoderOpenAI

Comprehensibility as an architectural property: Systems that cannot be understood cannot evolve safely

This article was written by participants of the InfoQ Certified Architect Online event. It is a culmination of their work and reflects the participants’ commitment to AI Collective insights gained at the intersection with modern software architecture. Comprehensibility Problems Teams with extensive experience in complex systems have experienced situations where they have to deal with production failures where most of the time is spent figuring out how the various parts of the system connect and what the system actually does, rather than finding bugs.

InfoQ CN5d ago

业务架构

Running Local LLMs on the Arduino® UNO™ Q Board: a Practical Guide

Learn how to run local LLMs on Arduino UNO Q — from model selection and quantization to memory constraints and real-world edge AI use cases. The post Running Local LLMs on the Arduino® UNO™ Q Board: a Practical Guide appeared first on EE Times.

EE Times6d ago

Sponsored content

Eight global mainstream agents tested on Wall Street: Qianwen Office ranked first in overall ranking, surpassing Claude Cowork and Codex

According to news on August 19, analysts from Wall Street investment bank Jefferies recently conducted actual tests on eight global mainstream AI Agents for real office tasks. The results showed that through the collaboration of models and Harness, Alibaba Qianwen Office ranked first in overall score, surpassing Agent tools such as Claude Cowork and Codex in the United States. This test set up a total of 5 real office tasks, including completing the company's annual report summary based on multiple documents, searching and comparing company operating data online, operating a real desktop browser to complete information retrieval and document generation, and producing English based on the data...

Leiphone6d ago

业界

Beyond Consensus: Fragmentation of AI Policy in the Linux Ecosystem

As Large Language Models (LLMs) redefine the boundaries of software development, the Linux ecosystem is facing a complex intersection of legal risks, technical integrity, and philosophical purity. Projects are developing policies that vary: from GCC's restrictive "hardline" stance due to legal security and compiler accuracy concerns, to pragmatism led by Linux kernel maintainers that places a greater emphasis on personal responsibility and in-depth understanding. Meanwhile, the Kubernetes community has adopted a pragmatic model based on disclosure...

InfoQ CNAug 18, 2026

开源

UMass Amherst Engineers Make Edge AI More Efficient by Redesigning Both Algorithm and Hardware

System leverages hyperdimensional computing algorithms and analog in-memory computing hardware to achieve quick, accurate processing. The post UMass Amherst Engineers Make Edge AI More Efficient by Redesigning Both Algorithm and Hardware appeared first on Semiconductor Digest.

Semiconductor DigestAug 17, 2026

AI

RISC-V at Sixteen: From Modular ISA to Standardized Platforms at Hot Chips 2026

RISC-V has evolved from a Berkeley research architecture into a global instruction-set standard spanning embedded controllers, application processors, accelerators, and emerging servers. Its defining technical characteristic is not merely openness, but disciplined modularity. Four ratified base ISAs—RV32I, RV64I,… Read More The post RISC-V at Sixteen: From Modular ISA to Standardized Platforms at Hot Chips 2026 appeared first on SemiWiki.

Semiconductor Wiki1d ago

IPRISC-VSiFive

Hugging Face reportedly in talks to be acquired for $13B

Hugging Face has reportedly been fielding acquisition offers that would value the company at around $13B. But with the founders' feeling of responsibility to community, doubts arise as to whether a sale will happen.

TechCrunch1d ago

AIVentureHugging Face

Anthropic 最强模型难以吸引用户

Anthropic 的美国客户正使用其最强大 AI 工具的更廉价替代品，引发了外界对其高投入商业模式的质疑，而市场普遍预计 Anthropic 即将进行史上规模最大的 IPO。根据支付服务集团 Ramp 收集的 7 万家公司支出数据，Anthropic 规模最大、成本最高的模型 Fable 5 发布两个多月后，其支出占该公司各类工具总支出的比例仍仅约为 11%，目前已趋于稳定。这打破了企业用户默认选择最强大模...

Solidot1d ago

Iranian hackers shut down UK power plant for 4 days

Article URL: https://www.telegraph.co.uk/news/2026/08/22/iranian-hackers-shut-down-uk-power-plant/ Comments URL: https://news.ycombinator.com/item?id=49407509 Points: 56 # Comments: 34

Hacker News2d ago

Knowing When to Stop: The Art of Making a Loop Converge

Article URL: https://a16z.com/knowing-when-to-stop-the-art-of-making-a-loop-converge/ Comments URL: https://news.ycombinator.com/item?id=49403484 Points: 15 # Comments: 4

Hacker News3d ago

Amazon just hiked the prices for Echo, Fire TV, and Kindle products by up to 60 percent

Citing "significant increases in memory and storage component costs," Amazon has raised prices on Echos, Kindles, Fire TVs, and Eeros by up to 60 percent, as first reported by Fortune. The cheapest products in Amazon's lineup were hit hardest, with the Echo Dot smart speaker jumping from $49.99 to $79.99. The Fire TV Stick 4K […]

The Verge Tech3d ago

AmazonAmazon AlexaNews

Cloudflare 推出 Cache Response Rules，在源站响应后进一步控制缓存

Cloudflare 近日推出缓存响应规则（Cache Response Rules"），这是一套新的规则引擎，运行在源站返回响应之后、内容写入 Cloudflare 缓存之前。此前，缓存规则（Cache Rules）只能根据请求属性进行判断；缓存响应规则则新增了一个响应处理阶段，可以在响应进入缓存之前检查源站返回的内容。根据官方文档，缓存响应规则"允许用户修改 Cache-Control 指令、管理 Cache Tag，以及移除 Set-Coo...

InfoQ CN3d ago

服务革新

npm 终于动手了：默认封杀 postinstall 脚本

由 GitHub 维护并随 Node.js 一同提供的软件包管理器 npm 发布了 npm 12"。该版本更改了 npm install 的一系列安全相关默认行为，同时弃用可以绕过双因素认证的细粒度访问令牌。npm 12 将此前自动执行的三种安装行为改为需要开发者明确选择启用。这些更改于今年 6 月首次公布"，并从 npm 11.16.0 开始通过警告形式提供，以便团队在升级前做好准备。最重要的变化是 allowScripts 现在默...

InfoQ CN3d ago

微软安全

How TSMC Is Wiring the AI Era With Light

TSMC’s photonics strategy is centered on integrating optical input-output with advanced logic, rather than selling conventional optical transceivers as standalone products. The company is developing a silicon-photonics foundry platform and a packaging architecture called TSMC-COUPE, or Compact Universal Photonic … Read More The post How TSMC Is Wiring the AI Era With Light appeared first on SemiWiki.

Semiconductor Wiki4d ago

3D ICAIChiplet

Cloudflare Cuts Astro Github Issues by 85% with AI Agents

Cloudflare, Astro, AI agents, GitHub Actions, issue triage, agentic AI, software architecture, open source, developer tools, AI automation, automated testing, human in the loop, agent workflows, GitHub, software engineering, AI software development, bug triage, continuous integration, developer productivity, autonomous agents, AI coding, Cloudflare Workers, Flue, triagebot By Leela Kumili

InfoQ EN4d ago

AutomationIssue Tracking ToolsWorkflow / BPM

The embodied bosses "quarreled" over these issues

Who has the final say in the embodied GPT moment?

TMTpost4d ago

After reading Alibaba’s financial report, I discovered that the more intense AI is, the more it will “fatten” the shovel sellers

After reading Alibaba’s latest financial report last night, I don’t want to admit it, but I have to admit a fact: Although the market spotlight is on the model layer and application layer, in the short term, the strongest commercialization of AI is in the cloud and chips. Moreover, the more aggressively the upper-layer models are used, the more profitable the bottom-level shovel sellers will be. The latest financial report shows that Alibaba Cloud's revenue in the first quarter of fiscal year 2027 (second quarter of 2026) reached 48.737 billion yuan, a year-on-year growth of 45%. More important is revenue guidance for the coming quarters. During the earnings call, Wu...

Leiphone4d ago

企业服务

OK, can we actually cool data centers with our pee?

Jason Kelce joked that people should cool data centers with their pee, rather than potable water -- but his suggestion is not completely ludicrous.

TechCrunch5d ago

AIdata centersExclusive

Amazon’s drone deliveries are landing in pools and ponds

Amazon's speedy drone delivery service will soon reach 500 cities across the US - but that might just mean there are more pools to drop packages into. On Wednesday, ABC7 News Bay Area shared a video showing an Amazon delivery drone hovering over a customer's pool in Texas, before opening its hatch and plopping the […]

The Verge Tech5d ago

AmazonDronesNews

Next.js 16.3: Instant Navigations, Up to 90% Less Dev Memory and Faster Builds

Vercel has released Next.js 16.3, featuring significant updates since version 16.0. Enhancements include reduced memory usage during development, accelerated build times, and improved type checking. Instant Navigations introduces faster, client-like responses while maintaining server-rendered architecture. Developers are advised to gradually adopt new features due to noted caveats. By Daniel Curtis

InfoQ EN5d ago

JavaScriptRoutingTypeScript

The why behind sex

Research details answer to a question that’s more complex than it seems

Harvard Tech News6d ago

Science & TechFacultyResearch

AI crawlers are pouring into e-commerce, and the security line of defense is shifting from "interception" to "judgment"

AI is entering e-commerce, but it not only brings new shopping guides, searches, and transactions, it is also changing the security issues faced by e-commerce websites. Recently, when interpreting Akamai's latest "Internet Security Situation Report", Akamai's Senior Solutions Manager for Greater China Ma Jun mentioned that in 2025, more than 17 trillion bots will visit e-commerce websites around the world, a year-on-year increase of 19%; of which the Asia-Pacific region will grow by more than 63%, making it the fastest growing region. At the same time, API attacks increased by 9%, surpassing traditional web application attacks for the first time and becoming the report statistics...

InfoQ CN6d ago

安全

The Future Of AI Compute Won’t Run On Just One Kind Of Chip

Power, token cost, interconnects, and software orchestration are pushing data centers toward heterogeneous clusters built from CPUs, GPUs, NPUs, optics, and custom accelerators. The post The Future Of AI Compute Won’t Run On Just One Kind Of Chip appeared first on Semiconductor Engineering.

Semiconductor Engineering6d ago

Low Power-High PerformanceRound TablesTop Stories

How 35 percent of US employees are left on the margins

In a new book, Paul Osterman details his research finding more than one-third of the country’s workers are “disposable,” with little security or chance of job advancement.

MIT News ResearchAug 18, 2026

Books and authorsFacultyResearch

DNA shaper steers nervous system development

A study of worms finds a protein that helps shape the structure of the genome is critical for establishing the identity of some neurons.

MIT News ResearchAug 17, 2026

ResearchGeneticsDNA