Tech News

Google DeepMind

Google Research

Các nhà khoa học đang giúp siêu máy tính lượng tử tự học cách sửa các lỗi sai của chính nó để tính toán chính xác hơn.

Hugging Face

IEEE Spectrum

I’ve seen two total solar eclipses and have been duly impressed by what happens as the moon casts its shadow on Earth. But recently I’ve become even more intrigued by a similar phenomenon that doesn’t involve the sun or the moon—something called an asteroid occultation.That’s wha

IEEE Spectrum

This white paper gives distribution engineers and utility teams a practical overview of the substation exit, the short and high-consequence section where station capacity divides into individual feeders. It explains how covered, spacer-supported overhead construction can reduce c

OpenAI

GPT-6 Astra produces more structured, context-aware legal documents, freeing lawyers to focus on strategy.

OpenAI

With GPT‑6 Astra, invideo plans edits with greater precision, improves color correction and grading threefold, and produces 50 custom effects in one day.

OpenAI

OpenAI CEO Sam Altman discusses AI safety, human control, and international cooperation in remarks to the United Nations Security Council.

Hindsight: Agent Memory That Learns

1.6k stars · 27.3k total

Learn it. Build it. Ship it for others.

310 stars · 56.2k total

Fix X/Twitter and Bluesky embeds! Use multiple images, videos, polls, translations and more on Discord, Telegram and others

165 stars · 5.3k total

An offline-first photo manager for large local libraries

151 stars · 2.8k total

Diffusion model(SD,Flux,Wan,Qwen Image,Z-Image,...) inference in pure C/C++

69 stars · 7.2k total

Educational technology (EdTech) platforms collect highly sensitive student data, including behavioral logs, disability records, and academic histories. However, privacy considerations are often postponed rather than treated as a foundational design requirement. We present a mixed

Cricket is one of the most celebrated sports world-wide, and technological advancement has become deeply embedded in how the modern game is analyzed and coached. Cricket shot classification and automated performance analysis add a further dimension to this trend. Traditional appr

Knowledge graph completion is evaluated by ranking observed triples above randomly corrupted ones, which treats every unobserved fact as false. When the object being completed is a description logic knowledge base rather than a plain graph, the open world assumption and deductive

Surgical procedures follow a phase-to-step hierarchy, yet the video-language models used to recognize them are evaluated with flat per-level metrics that ignore cross-level coherence and error structure. In this paper we make two contributions to address this problem, (i) we intr

Postoperative acute kidney injury (AKI) after major non-cardiac surgery carries substantial morbidity, yet early intraoperative risk stratification remains difficult. In this retrospective cohort study, we propose SynerT, a waveform-only hybrid temporal backbone that combines a c

The conventional hearing aid is both costly and lim- ited in usage, as it is not intended to detect non-speech audio. Our objective is to develop a machine learning solution to provide a more accurate and affordable mechanism to identify surrounding sounds to improve the safety o

Existing texture datasets for tactile sensing primarily consist of sensor readings from a specific sensor interacting with available surfaces/objects rather than describing the textures themselves, limiting fair comparison between tactile sensors and hindering reproducible resear

Given a dataset of finitely many elements $\mathcal{T} = \{\mathbf{x}_i\}_{i = 1}^N$, the goal of dataset condensation (DC) is to construct a synthetic dataset $\mathcal{S} = \{\tilde{\mathbf{x}}_j\}_{j = 1}^M$ which is significantly smaller ($M \ll N$) such that a model trained

In-context learning (ICL) enables task adaptation at inference time by conditioning on demonstrations rather than updating model parameters. Although recent time-series foundation models incorporate contextual conditioning, retrieval, or example-based prompting, they typically re

In this work, we propose a hybrid iterative deep Ritz method (H-IDRM) for a class of interface problems for second-order elliptic operators. It is based on a new mixed formulation of the problem and involves solving a sequence of convex minimization problems. We employ a level-se

Generative AI is rapidly entering adolescents' everyday lives during a critical period of cognitive, social and emotional development. Yet its adoption is outpacing evidence on how adolescents themselves experience, understand and negotiate its expanding role in their lives. We e

Online platforms have become arenas for the public contestation of climate change, shaping how scientific knowledge, denial, and uncertainty are expressed and disputed. Yet longitudinal evidence remains limited for YouTube, especially for Portuguese-language discourse. Addressing

Self-supervised speech models encode rich phonetic information, but it remains unclear how to transform this information into interpretable metrics for second-language (L2) pronunciation assessment in spontaneous speech. We propose a native-reference coordinate geometry in which

Recent advances in large language models (LLMs) have enabled agents to tackle long-horizon tasks across diverse environments. To further improve agent performance, existing language world models typically predict environment observations, yet reconstructing high-entropy, executio

Claim denial management costs U.S. healthcare approximately $260 billion annually in administrative overhead. Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) can produce fluent clinical text, but single-agent architectures fail in high-stakes healthcare: the

Automatic failure recovery is another difficult area of DevOps deployments due to the complexity of the system, high workload and the limitations of the traditional rules-based or manual approach. This study proposes a hybrid model integrating deterministic, rule-based recovery w

Ground-based astronomical observations frequently contain streaks produced by artificial satellites, space debris, and potentially Near-Earth Objects (NEOs). While machine-learning models can reliably detect these features, their practical adoption in observatory operations is of

A language model agent acts through the tools it is given. The data it reads while working on a task can redirect what it does with those tools. A growing set of techniques for safe and secure agent execution therefore sits between the agent and its tools, aiming to enforce acces

This work presents the development of an automatic refactoring tool for Java code built on top of the Eclipse JDT API. The proposed approach transforms control structures containing break and continue statements within different types of loops into semantically equivalent constru

In mainstream agent frameworks, a step ends when the agent's own output says it has finished. Durable-execution platforms bound retries and time, but their checker conventionally lives in the same codebase as the work: a discipline the deployment is trusted to keep, not a propert

This article is crossposted from IEEE Spectrum’s careers newsletter. Sign up now to get insider tips, expert advice, and practical strategies, written in partnership with tech career development company Parsity and delivered to your inbox for free!I lost a cushy engineering manag

I’m sitting in a Rivian R1S SUV as it drives itself down the leafy streets of Palo Alto, Calif., through areas crowded with touchstones of tech history. We cruise near the landmark HP Garage, the one-car workshop where Hewlett-Packard, and, arguably, Silicon Valley, was founded i

Các nhà khoa học chế tạo ra những chiếc thuyền robot nhỏ tự động gắn kết lại với nhau như những miếng ghép hình để tạo thành cầu hoặc bến phao nổi trên mặt nước theo ý muốn.

Các nhà nghiên cứu dạy robot hiểu những hướng dẫn mơ hồ của con người nhờ sự trợ giúp của trí tuệ nhân tạo ngôn ngữ, giúp chúng làm việc nhà mà không cần con người phải chỉ dẫn từng ly từng tí.

Một hệ thống thông minh mới giúp các hệ thống trí tuệ nhân tạo phức tạp chạy mượt mà hơn, tiết kiệm điện năng và giảm chi phí máy chủ.

Gemini Robotics ER 2 helps robots reason, collaborate, and solve real-world tasks. It represents a step change in video understanding, tool orchestration, and multi-robot collaboration for robotic applications.

🔗 Cũng đưa tin: Google DeepMind

Các nhà nghiên cứu đang tạo ra một trợ lý ảo thông minh biết hỏi thăm các triệu chứng bệnh của bạn giống như một bác sĩ thực thụ.

Bài nghiên cứu giúp chúng ta hiểu rõ hơn về cách mà trí tuệ nhân tạo có thể tự tưởng tượng ra những bức tranh đẹp từ văn bản.

System-wide Dictation on Apple devices runs entirely on-device, and the speech it transcribes reaches the foundation model through a tokenizer: an encoder that maps short windows of waveform onto the representation the language model reads. Because that model is sparsely activate

We introduce a new method to guide flow matching models. Our approach, which we call probe guidance, uses the frozen internal states of an existing diffusion model to construct a guidance signal. This works using a similar principle as autoguidance, but eliminates the need for an

At a glance Challenges a core assumption in robotics AI: Our research shows that running physical AI inference exclusively on onboard GPUs can limit robot performance, battery life, and scalability, and that offloading inference to edge or cloud GPUs can offer significant adv

Research Note: CARE-X is a research model and not a Microsoft product offering or medical device. It has not been cleared or approved by any regulatory authority and is not intended for clinical diagnosis, screening, or patient care. The results described below are retrospective

At a glance Orchard is an open-source framework for scalable and cost-effective agentic AI research, built around Orchard Env, a reusable environment service for training and evaluating agents across task domains. The same Orchard infrastructure supports software-engineeri

When you get a good night’s sleep, you aren’t just giving your brain a chance to rest. Sleep activates a system only discovered in 2012 that washes out brain waste. Called the glymphatic system, it’s comparable to the better-known lymphatic system that moves and filters fluids th

Discover how the ZEISS Crossbeam 750 FIBSEM sets a new benchmark for precise TEM lamella prep, tomography, and advanced nanofabrication. This delivers better resolution, better SNR, larger usable FOV, and shorter acquisition times. Learn how uninterrupted FIB milling will reduce

In the chill of a London spring night, under overcast skies, iconic Trafalgar Square opens around me. Admiral Nelson rises on his pedestal, the National Gallery rests behind, the church of St Martin-in-the-Fields sits nearby. From the 13th century, the site served as the Royal Me

Giống như việc dùng một đường cao tốc rộng lớn để chuyển thật nhanh đồ chơi từ nhà máy này sang nhà máy khác, công nghệ này giúp các phần nhỏ của máy tính nói chuyện với nhau mượt mà hơn.

Giống như một nhóm thợ thủ công chuyên môn khác nhau cùng bàn cách chế tạo một cỗ máy, các trợ lý trí tuệ nhân tạo nhỏ hơn đang phối hợp với nhau để thiết kế chip thông minh hơn.

Giống như một người quản lý kho thông minh luôn kiểm tra kỹ mọi món hàng trước khi xuất xưởng, phần mềm tự động hóa mới giúp các kỹ sư thiết kế chip nhanh hơn mà vẫn đảm bảo không có lỗi.

TSMC shares updates about the Open Innovation Platform Ecosystem and vision for enabling AI-driven agentic workflows using TSMC AI Design Kit The post Inside TSMC’s Evolving Design Ecosystem: Shaping the Future of AI, with AI appeared first on EE Times.

As SkyWater scales its 200-mm and 300-mm manufacturing platforms to support diverse quantum modalities, leadership emphasizes that protecting customer IP remains its top priority. The post After IonQ Buyout, SkyWater Reiterates Role as Quantum Foundry appeared first on EE Times.

acing TI's October price hike? SKYEAST offers $30M+ ADI/TI spot stock, 2-hour quotes, and 7×24 support across time zones. Lock in supply now. The post TI’s October Price Hike: How to Secure ADI/TI Stock Without the Panic appeared first on EE Times.

IC3 is a state-of-the-art algorithm for hardware model checking that proves safety properties by incrementally constructing an inductive invariant consisting of a set of lemmas. Its effectiveness depends on generalization heuristics that identify useful lemmas and guide proof sea

Rapidly growing AI workloads are driving large investments in AI datacenters. This survey classifies industrial AI accelerators into four architectural categories and compares their compute and memory organizations. It examines how node-, rack-, and pod-scale interconnects suppor

NAND-backed storage offers the capacity needed to scale LLM prefix caching, but its block I/O path incurs CPU cache contention and host-DRAM staging in addition to NAND latency. Our characterization shows that these interface costs persist even with DRAM as the storage medium, mo

Providing students with experience from chip specification to fabricated silicon can strengthen chip-design education, but integrating tapeout into regular teaching is difficult to scale, especially across universities with different curricula, schedules, regulations, and technic

Một hệ thống máy thông minh mới biết tự phán đoán cách thức một món đồ bị hỏng để tháo rời nó ra và lấy lại những linh kiện còn dùng được.

Mỹ vừa đưa các loại robot thông minh từ nước ngoài vào danh sách hạn chế nhập khẩu vì lo ngại vấn đề an ninh mạng.

Công ty Walden Robotics ra mắt mẫu robot có bánh xe giúp làm việc nhà thực tế hơn là tạo hình giống hệt con người.

Industrial control electronics such as programmable logic controllers, servo drives and operator panels are routinely repaired in plant maintenance, but were never designed for automated disassembly. This paper presents a modular dual-arm robotic cell for repair-oriented disassem

In this paper, we present the Berkeley QUAD (Quasi-direct-drive, Underactuated, Asymmetric Design) Hand, a four-finger anthropomorphic robotic hand with 11 degrees of freedom and 8 degrees of actuation. The design utilizes QDD actuation at the base of each finger, enabling high f

We develop a sample-based framework for constructing information-driven hierarchical multi-resolution representations of probabilistic occupancy grids. Recent methods compute information-optimal abstractions via dynamic-programming-based exhaustive recursions, which become comput

We develop a decentralized Perception-Action-Communication (PAC) system for multi-robot teams that enables them to collaborate in large scale, outdoor environments. Our system natively supports deployments at any scale by leveraging a graph neural network (GNN) to diffuse informa

The 2026 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2026) will be held from 27 September – 1 October in Pittsburgh, USA. The programme includes plenary and keynote talks, workshops, tutorials, forums, competitions, and a debate. Keynote talks The k

On September 4, 2026, Oren Yomtov, a security researcher from Accomplish, responsibly reported a vulnerability affecting Cloudflare Containers and Cloudflare Sandboxes (which is built on Containers), through Cloudflare’s bug bounty program. Cloudflare has fully remediated the vul

We’ve rewritten the module registry in workerd, the core open-source component of the Workers runtime, to be faster, more standards-compliant, and more closely aligned with Node.js' module registry.Over the past few years, we’ve been adding support for more and more Node.js runti

Tool: Gemini 3.8 TTS Playground Google released two new Gemini text-to-speech models today - gemini-3.8-flash-tts and gemini-3.8-flash-lite-tts. They come with a library of over 2,000 voices, plus the ability to create a custom voice with "just a 30-second audio sample of

Tool: Shadow roots, explained with live examples Prompt to Fable 5.1 Medium: Build an artifact to explain shadow roots in CSS with interactive examples Tags: css

Hey, you know it's like super obvious if you're using AI to write your scripts for TikTok and YouTube, right? [...] It's not just the general AI-isms of "it's not X, it's Y", or the rule of three, or the really weird broken staccato-like way of writing where you just say a lot of

Hallucination in large language models usually refers to the model generating unfaithful, fabricated, inconsistent, or nonsensical content. As a term, hallucination has been somewhat generalized to cases when the model makes mistakes. Here, I would like to narrow down the problem

Diffusion models have demonstrated strong results on image synthesis in past years. Now the research community has started working on a harder task—using it for video generation. The task itself is a superset of the image case, since an image is a video of 1 frame, and it is much

852 points

744 points

566 points

552 points

464 points

351 points

325 points

315 points

292 points

276 points

133 upvotes

77 upvotes

53 upvotes

48 upvotes

35 upvotes

32 upvotes

21 upvotes

17 upvotes

One of these claim to move from pinecone to Qdrant seeing p99 retrieval latency drop from 180ms to 32ms. Even Confident Moved to pgvector from Pinecone

Made by replit and me

Hey everyone, With the heavy shift toward AI codegen tools, I realized a lot of us (myself included) are starting to lose touch with core, low-level engineering fundamentals. If an API or LLM service drops, are we still confident in our core mechanics? To fix this for myself, I c

TL;DR DEV has so many interesting series, but they can be easy to miss when we don't follow the...

67 reactions

If you checked my profile recently, you probably noticed my page has been completely quiet since...

56 reactions

A few weeks ago, I wrote about open-source projects moving toward paid or dual-licensing models, and...

42 reactions

TLDR - I finished a small project that I am sharing for my friends on DEV, that allows you...

38 reactions

Feels like it's been more than a minute since my last post. A lot has happened in my life over the...

30 reactions

Le Grand-Espace enhances the cultural offering of the Eastern Townships tourist region through a flexible performance venue capable of hosting a wide range of artistic disciplines, from theatre and dance to circus arts and productions for young audiences.

The house is built in a semi-dense part of the city of Stavanger and is mostly just a large outcropping piece of bedrock, with a very steep slope toward the street. The program is ordinary, a house of very modest size planned for a young couple with a small child.

A fire has severely damaged La Querola d'Ordino, a residential development in Ordino, Andorra, designed by Jean Nouvel in collaboration with Ribas & Ribas Arquitectes and Jordi Sala. The fire began during the early hours of September 22 and spread across the seven-building comple

"After all, time is a key ingredient of my profession," states Francis Kéré in his book Francis Kéré: Building Stories. Architecture is constantly evolving and transforming over time. Bricks, doors, windows, walls, stairs, and other elements become the protagonists of a process t

The new Zando Central Market of Kinshasa, was conceived above all as an African market, not a mall. Despite its scale and thanks to the attention to comfort, the project remains open to the city, its climate, its people and its everyday life. Its architecture provides a robust an

The Fondation Beyeler in Riehen, Switzerland, is expanding its museum ensemble with three new buildings by Swiss architect Peter Zumthor, an enlarged and restored park, and the adaptive reuse of several historic structures. The expansion extends the existing museum designed by Re

Arcipressi is an architectural residence nestled in the green heart of Florence, a portion of a historic villa that has undergone a careful renovation aimed at restoring a timeless elegance to the spaces, in harmony with the surrounding landscape.

This week began with the United Nations' commemoration of the International Day of Peace, an invitation to ask what conditions are needed to cultivate nonviolence, social stability, and respect for human dignity. Against a backdrop of growing individual isolation, authoritarianis

In most architectural projects, vertical circulation is expressed as a purely functional element, tucked away behind fire-rated doors or an elevator core. However, sometimes architects also use spirals when trying to resolve site elevation changes. Given the right scale, these el

For some time, I have been wondering whether it might be possible to create an architectural system without a fixed, predetermined outline—one that remains open and flexible with regard to both space and time. Of course, the Dom-ino system is well-established: a structural concep

Commissioned by the chairman to commemorate his parents and ancestral home, Tianlai Art Museum transforms private memory into a public space for exhibition, set within its rural context. The original old house had deteriorated beyond repair, so a full reconstruction was chosen ov

Swiss studio Herzog & de Meuron has utilised the existing concrete foundation of an older structure for the Student Center & Dining Commons at Amherst College, adding mass-timber levels on top. The Student Center & Dining Commons at Amherst College in Massachusetts rises from the

Swiss architect Peter Zumthor has expanded the Fondation Beyeler museum in Riehen, Switzerland, with a cluster of art buildings designed to nestle into the park setting. Built in the gardens outside the original Fondation Beyeler art museum, which was designed by Renzo Piano in 1

As a wealthy nation, Norway has a responsibility to tackle climate change by developing a nature-first approach to designing buildings, Oslo Architecture Triennale director Line Ramstad tells Dezeen in this interview. At this year's Oslo Architecture Triennale, built environment

The conversion of a Victorian hospital is one of three reuse projects that make up the shortlist for the 2026 RIAS Andrew Doolan Best Building in Scotland award. Designed by Bennetts Associates, the former hospital building now houses a mix of teaching spaces for the University o

Washington DC nonprofit The National Trust for Historic Preservation, which sued the government last year over the White House ballroom extension, has announced a $10 million fundraising campaign for legal defence in support of US landmark preservation. National Trust for Histori

Sobokuya builds The Shrine Office in mountain valley in Hokuto   Set within a mountain valley in the precincts of Fujitake Shrine, in Hokuto, Japan, The Shrine Office by Sobokuya is conceived around the way natural materials change and deepen with age. Two generous openings bring

MPNYC transforms Piraeus warehouse through adaptive reuse   Manhattan Projects New York City (MPNYC) transforms a former storage and workshop building on Polydefkous Street in Piraeus, Greece, into a combined office, showroom, and gallery. Located diagonally opposite Castor Place