Other
Cybersecurity news in this category
OAuth Authorization Server Built Without Storing User Profiles
Why a 202-Character License Key Uses ECDSA P-256 Instead of Ed25519 or RSA
Pilcrow Publishes Free Comprehensive Guide on Web Authentication
redb 3.4.0 Adds Replay Checkpoints, Shared Runtime Layer and Declarative Secrets Handling for .NET Ecosystem
redb 3.4.0 delivers operational improvements for the .NET integration platform that includes typed storage over Postgres, MSSQL and SQLite, the redb.Route engine modeled after Apache Camel, the Tsak runtime with dashboard and clustering, and redb.Identity for OIDC and OAuth 2.1. The release focuses on post-deployment resilience with replay checkpoints that capture message state at named points inside routes, allowing safe re-execution of downstream steps without duplicating side effects. A shared runtime layer moves framework assemblies into a separate Libs/shared directory so that individual DLLs can be replaced without rebuilding or redeploying the entire Tsak or Identity packages. Secrets are now declared with a [Sensitive] attribute on endpoint options, ensuring URIs containing passwords or keys are sanitized before they reach logs, metrics or health checks. Additional controls include role-based access to management APIs, persistent audit trails and cryptographic signing of dynamically loaded modules. All Pro features remain free on the entire 3.x line with no licensing server required.
Neural Networks Without Magic: 80-Year History, Business Applications, and Why They Will Not Replace Experts Overnight
In an in-depth interview, Data Science team lead Vasily Ryazanov traces neural networks back to the 1970s work of his father and academician Zhuravlev, explaining that the technology is approximately 80 years old rather than a recent phenomenon. Ryazanov details how modern large language models such as ChatGPT and Claude function by predicting tokens within a context window after pre-training on massive datasets, and he contrasts prompt engineering with the deeper mathematical and programming skills required to build models. He describes real-world deployments including an antifraud system for the insurance company Alliance that automates detection of medical claim fraud. The discussion covers practical limits such as hallucinations, risks of uploading sensitive data to external services, and the psychological tendency of users to over-trust fluent model outputs. Ryazanov emphasizes that while tools like Claude and ChatGPT accelerate routine tasks, they remain assistants that require human verification on high-stakes decisions in health, finance, or security.
Meshtastic Gains Renewed Popularity as Offline LoRa Mesh Networks Expand for Remote and Disaster Communications
Meshtastic, the LoRa-based mesh protocol first introduced in 2020, is experiencing a significant resurgence driven by global internet outages and the need for dependable offline communication. Originally used for urban mesh networks, disaster response, hiking groups, and search-and-rescue operations, the technology now attracts new users across Russia and worldwide. Two main device types exist: stationary rooftop nodes that form network infrastructure and portable mobile nodes that connect smartphones via Bluetooth. Popular ready-made devices include the Hacker Pager terminal and Heltec V3 modules, which support both standalone operation and smartphone integration through official Meshtastic apps. Enthusiasts have demonstrated practical applications such as remote shed alarms using infrared beam sensors connected to Heltec ESP32 LoRa boards that publish alerts to Home Assistant via MQTT. The Mars Society has deployed T-Echo radios during multi-week training expeditions in areas lacking cellular coverage, while municipalities explore Meshtastic as backup systems for natural disasters. New commercial products like the T5 E-Paper S3 Pro further lower barriers by offering pre-configured hardware with e-paper displays.
Microsoft Adds Option to Completely Disable Copilot Key in Windows 11
Microsoft is testing a new setting in experimental builds of Windows 11 that lets home users fully disable the dedicated Copilot key on compatible keyboards. The option, labeled Do nothing, appears in the Bluetooth and devices section under Keyboard settings and allows users to reassign the key away from launching the AI assistant. Enthusiast phantomofearth discovered the feature, which currently sits alongside choices for launching Microsoft 365 Copilot, performing a search, or executing a custom action. The rollout is gradual, with some Windows Insiders receiving the setting earlier than others. The Copilot key was introduced on new Windows laptops in 2024 as a replacement for the right Ctrl key to highlight AI PCs. Earlier this year Microsoft had already promised limited support for restoring the original Ctrl function, while users previously relied on third-party tools such as NoCopilotKey to prevent accidental activation.
Turkish BiP Messenger Audience in Russia Surges 650 Times in One Year Amid Foreign App Adoption
Russian users continue exploring foreign messaging services even as initial hype subsides. MTS AdTech data shows demand for overseas platforms rose 50 percent in the first quarter and 26 percent in the second. Turkish BiP emerged as the standout performer, with its Russian monthly active users climbing 120 percent from April to July to reach 4.3 million. This marks a dramatic increase from just 6,600 users a year earlier. South Korean KakaoTalk grew 80 percent to 1.26 million users, while WeChat rose 17 percent and Gem Space increased 11 percent. American Imo remains the largest foreign messenger with 12.2 million users but added only 1 percent over the quarter. Blocked services SimpleX and Discord saw declines of 41 percent and 8 percent respectively, while domestic platforms VKontakte, MAX and Telegram retain far larger audiences.
Building Trusted TLS in Kubernetes Without InsecureSkipVerify Using cert-manager and trust-manager
The article explains how to establish verifiable TLS across a Kubernetes cluster running on Talos with Cilium networking and Linstor storage, eliminating reliance on InsecureSkipVerify flags. It details the creation of an internal self-signed CA via cert-manager, using a selfSigned bootstrap issuer to generate the CA certificate before deploying a production ClusterIssuer. trust-manager then distributes only the public CA bundle as ConfigMaps across all namespaces, allowing workloads to validate certificates without exposing private keys. For external services, an ACME issuer integrates with Let's Encrypt and Gateway API to automate HTTP-01 challenges. The guide also covers kubelet server certificate rotation through serverTLSBootstrap and the kubelet-serving-cert-approver controller, which validates CSR requests against Node objects before approval. These steps ensure mutual trust for internal services, publicly trusted certificates for internet-facing endpoints, and verified kubelet connections for the API server and monitoring systems.
Designing and Implementing Private Cloud Security Across Geographically Distributed Data Centers
The article shares practical experience in designing and deploying protection for a private cloud hosted across two geographically separated data center sites. The project aimed to achieve service resilience, meet information security requirements, and maintain manageable infrastructure. While theoretical designs often combine standard best practices such as high availability, network segmentation, next-generation firewalls, and dynamic routing, real-world implementation revealed significant conflicts between architectural components. Multiple contractor teams participated, each bringing different visions for network architecture and security controls, leading to complex negotiations and compromises. The process evolved into an engineering puzzle where proposed solutions had to be integrated into a functional, scalable system despite technical limitations and differing priorities. The full piece focuses on the journey from an idealized design to practical deployment through iterative problem-solving rather than deep technical dives.
VK Tech Adds Multi-Level Memory to AI Agents in VK AI Space Platform
VK Tech has introduced a four-layer memory system for AI agents within its VK AI Space platform, enabling them to retain context, reuse materials, and incorporate past employee edits instead of restarting tasks from scratch. The new capability is expected to reduce the share of repeatedly solved tasks by 40–50 percent. Memory layers include session memory for current dialogues, user memory for employee profiles and preferences, project memory for specific project materials and rules, and agent memory for recording successful approaches and human corrections. This transforms AI agents from short-term tools into consistent team participants that can draw on accumulated corporate knowledge during tasks such as tender preparation, incident investigation, and product launches. Access remains strictly limited to information the collaborating employee is authorized to view, with all data stored inside the company perimeter and never sent to external services. Users can review and delete stored information at any time. VK Tech believes the feature will convert individual specialist experience into a persistent organizational asset that carries forward across tasks.
Efort Qizhi Demonstrates Humanoid Robots Mastering Industrial Tasks Through Universal Technology Base at WAIC 2026
At WAIC 2026 in Shanghai, Efort Qizhi presented its humanoid robots performing complex tasks on uneven terrain, including lifting irregular boulders and installing large photovoltaic panels without prior task-specific coding. The company showcased its HALO human skill collection suit that captures multi-modal data including vision, force, touch and language to distill human operations into robot-understandable representations. Central to the approach is HumanGPT, a world model tailored for embodied intelligence that integrates collected human skills with sensor data to create virtual dynamic environments for robot decision-making. Efort Qizhi also highlighted its full toolchain covering data collection, storage, governance, training, inference and deployment, built around HumanGPT, Dayan data platform, Modou IDE and Openmind OS. Chairman Dr. You Wei introduced the skill iceberg concept, stressing that true industrial value lies in complex capabilities such as multi-robot coordination and real-time process parameter generation rather than simple pick-and-place actions. The system has already logged over 5000 hours of operation on customer production lines, proving transfer from human expertise to reliable machine productivity across manufacturing scenarios.
Google Tests Third-Party App Store Support in Play Store Following Epic Games Antitrust Ruling
Google is actively testing multiple new features in the Play Store, including a dedicated Play Labs experimental section, direct installation of alternative app stores, and an expanded two-line search interface. The most notable change allows users to install competing app marketplaces directly from Google Play, a capability tied to the reinstated 2024 court decision in the Epic Games antitrust case after the companies' settlement was withdrawn. Although the new menu option for third-party stores has been discovered through app modifications, attempts to access it currently result in errors. Play Labs, an experimental area for previewing upcoming store capabilities similar to features already present in Google Search, was also found but remains non-functional. The updated search bar is already operational, supporting longer natural-language queries and an Ask Store prompt to help users describe needs in detail rather than using short app names. APK analysis indicates these developments are still in progress, meaning Google may modify, delay, or cancel any of them before wider release. The overall direction signals that the primary Android app store is preparing to facilitate competition by distributing rival marketplaces.
Spring Security Tutorial Details Refresh Token Implementation to Complement JWT Authentication
A detailed technical guide explains how to enhance JWT-based authentication in Spring Security by introducing refresh tokens for improved security and user experience. The article covers the creation of a RefreshToken entity stored in the database, along with repository and service layers that handle token generation, verification, revocation, and rotation using UUID and expiration timestamps. It discusses security advantages such as short-lived access tokens combined with long-lived refresh tokens hashed via SHA-256 or HMAC-SHA256, while also addressing risks including database theft and XSS attacks mitigated by HttpOnly cookies. Code examples demonstrate updates to AuthService, AuthController, and SecurityConfig to support registration, login, token refresh via cookies, and logout functionality. Additional best practices include automatic token rotation on refresh and session management across multiple devices. The tutorial emphasizes minimizing credential transmission over networks while maintaining stateful token validation.
OTUS Publishes Extensive Digest Covering Kubernetes, GitLab CI, DevSecOps and Infrastructure Reliability
OTUS has released a large-scale digest that compiles free lessons, practical articles and advanced courses focused on real-world infrastructure challenges. The collection addresses interconnected topics including Kubernetes container orchestration, GitLab CI pipeline automation, observability with OpenTelemetry, network troubleshooting, and DevSecOps practices. Readers can explore specific lessons on secrets management with Vault, load testing, RAID recovery, and kernel-level protections in Linux. The digest also features in-depth articles on Docker Compose configuration, CrowdSec migration, L2 network loops, and self-service deployment models. For experienced professionals, OTUS offers specialized courses on high-load systems, PostgreSQL administration, Linux kernel development, and advanced networking with BGP and OSPF.
Xello and Kode Bezopasnosti Sign Technological Partnership to Integrate Equipment Emulation into Xello Deception Platform
Xello, a Russian developer of data and infrastructure protection platforms, and Kode Bezopasnosti, a Russian developer of software and hardware information security tools, have signed a technological partnership agreement. The collaboration will enable emulation of Kode Bezopasnosti equipment within Xello Deception, the first domestic Distributed Deception Platform designed to protect against targeted attacks. This integration will enhance the creation of a realistic false layer of IT assets and expand use cases for early threat detection in customer environments. Xello product head Rustam Zakirov emphasized that the partnership addresses customer needs by making the deception layer more authentic for organizations using Kode Bezopasnosti solutions while fostering expertise exchange. Kode Bezopasnosti strategic marketing head Pavel Korostelev highlighted the growing demand for deception technologies amid rising targeted attacks and the value of supporting proactive detection capabilities on the Russian market. The agreement combines both companies' strengths to improve practical scenarios for cyber deception deployment.
How Ideco Measures NGFW Performance Using TRex ASTF and RFC 9411 Methodology
Ideco engineers have published a detailed methodology for benchmarking the performance of their Ideco NGFW Novum product on the Ideco EX hardware platform. The team uses TRex in Advanced Stateful mode to generate realistic L4-L7 traffic and follows RFC 9411 guidelines to ensure reproducible results across different test environments. Key metrics include throughput on UDP 1518-byte frames reaching 200 Gbit/s, TCP/HTTP performance varying from 100 Gbit/s down to 62 Gbit/s depending on object size, and EMIX mixed traffic dropping from 92 Gbit/s with firewall only to 15.5 Gbit/s when all inspection engines are enabled. The methodology incorporates binary search for maximum load, a 1% session drop threshold, and 300-second hold times per test point to capture real-world behavior rather than short-term peaks. All tests were performed without TLS decryption, with separate plans for hardware-based Keysight BreakingPoint testing when decryption is required. The full article includes hardware specifications, traffic profiles, loss calculation formulas, and configuration details to allow anyone to reproduce the measurements.
Siemens Industrial AI Drives Green Efficiency Revolution in Infrastructure Cooling, Parks, and Data Centers
The article explores how Siemens is applying industrial AI to transform physical infrastructure operations across buildings, industrial parks, and data centers in China. It details the company's non-invasive AI BOX solution that optimizes cooling systems in hotels and other facilities without replacing existing equipment, achieving around 7% additional energy savings. The Smart ECX platform enables integrated source-grid-load-storage management for zero-carbon factories, delivering 30% cost reductions and over 90% overall energy efficiency. Siemens also supplies critical power distribution hardware such as NXAirS medium-voltage switchgear and SIVACON S8 low-voltage cabinets to support high-reliability AI data center operations. The coverage highlights alignment with China's 15th Five-Year Plan carbon peaking policies and real-world deployments at venues including the Shanghai Yingyi Crowne Plaza Holiday Hotel and Sichuan Chuanrun's Chengdu factory. Overall, Siemens emphasizes combining domain physics models with machine learning to deliver verifiable, replicable industrial value rather than generic AI concepts.
RuBackup 2.9.0 Receives Official Compatibility Certificate with BAUM-Inform Storage Systems
RuBackup version 2.9.0 from the Astra Group has successfully passed compatibility testing with storage systems produced by BAUM-Inform. The verification confirmed that the backup solution can reliably perform data backup and recovery operations using BAUM-Inform hardware. Testing followed a predefined program and methodology, although detailed results and the full list of tested configurations were not disclosed. BAUM-Inform systems are designed for corporate data storage and processing, while RuBackup supports servers, workstations, virtualization environments, mail systems, databases, and domain infrastructure. The certification allows organizations to deploy both products together without additional integration validation. Both companies stated that the verified combination reduces deployment risks and supports the development of fully domestic data protection infrastructures.
VK WorkSpace Adds Guest Access Without Accounts, Password Protection, Polls, and File Uploads Up to 30 MB
VK Tech has updated the cloud version of its VK WorkSpace Board to allow external participants such as clients, contractors, and other guests to join collaborative boards without creating an account. Access can be protected by a password set by the owner, who can also assign specific permissions including view-only, commenting, or full editing rights while restricting guests from creating new boards or viewing version history. Domain administrators can enforce mandatory password protection across all boards that permit guest access. Additional features include built-in polls with automatic result counting, support for uploading files up to 30 MB that are scanned by antivirus software, a mini-map for navigating large projects, and export options in SVG, PNG, JPG, PDF, and CSV formats. The service also received an English-language interface, addressing key use cases involving external collaboration as highlighted by VK Tech executive Petr Shcheglov.
Solar SIEM 2026.2 Adds Full Solar JSOC Detection Library, TI Feeds Support, Enhanced AI Agent and Multi-Tenancy
GC Solar has released Solar SIEM 2026.2, which now includes the complete detection rule library developed by Solar JSOC over 14 years of monitoring approximately 300 customer infrastructures. The update eliminates the need for organizations to spend months building custom rule sets by providing ready-made detection scenarios that identify sophisticated attacks from the earliest implementation stages. In 2025, Solar JSOC recorded 1.16 million security events after false-positive filtering and confirmed more than 33,000 incidents, with malware accounting for 36 percent and unauthorized access attempts for 23 percent. The product also gained native support for TI Feeds, allowing automatic ingestion and correlation of indicators of compromise from Solar 4RAYS and customer-owned sources. The built-in AI agent has been significantly expanded so it can now independently query raw event data, perform deeper analysis, and recommend next steps, reducing manual workload for analysts. Multi-tenancy capabilities enable multiple organizations to share a single SIEM instance while keeping their event streams fully separated, targeting holdings, MSSP providers, and large enterprises with numerous divisions. More than 40 companies of various sizes participated in the pilot program.
Engineers Create HemaDyne Perfusion System to Replicate Real Human Blood Flow on Lab-on-a-Chip Devices, Aiding Personalized Medicine and Mars Missions
Researchers have developed an automated perfusion system called HemaDyne that accurately reproduces patient-specific blood flow patterns, including rapid pressure changes occurring within 50 milliseconds, inside microfluidic lab-on-a-chip devices. The technology overcomes previous limitations of laboratory pumps that could not match the pulsatile nature of blood circulation generated by the human heart, which produces multiple short impulses and waves of varying duration. By converting real patient examination data into G-code commands that control a foldable bellows-style chamber inspired by accordions and plastic glue bottles, the system delivers precise, individualized hemodynamic conditions to endothelial cells. This enables detailed study of how mechanical forces from blood flow influence vascular diseases, aging, and microgravity effects without relying on animal testing. The compact, long-term stable pump supports experiments lasting months and can be integrated with various organ models. NASA has backed the project to investigate cardiovascular changes during extended spaceflights, positioning these microphysiological systems as tools for preparing crewed missions to Mars where full clinical trials are impossible.
AI-Designed NovoTags Enable Real-Time Fluorescent Labeling of Mitochondria and Other Cellular Structures in Living Cells
Researchers have developed NovoTags, a new class of small artificial proteins created from scratch using AI systems to bind specific Janelia Fluor dyes and make targeted molecules glow brightly under microscopes. The proteins were designed with RFdiffusion for shape generation, LigandMPNN for amino acid sequences, and validated by AlphaFold and RoseTTAFold before lab synthesis and testing. Three independent tags were created for green, orange, and far-red spectra, allowing genetic attachment to proteins of interest such as those in endosomes, mitochondria, and chromatin. Unlike traditional fluorescent proteins, NovoTags do not glow on their own but form a pocket that holds the dye and alters its optical properties for high brightness and super-resolution imaging including STED microscopy. The team also introduced NovoSplit, a split version that acts as both a fluorescent label and molecular glue to control protein interactions on demand. Future plans include combining the tags with cryo-electron tomography and expanding the palette to distinguish up to 30 proteins simultaneously using color and fluorescence lifetime. All sequences and compatible dyes have been made openly available to the scientific community, with results published in Science.
Scientists Build Moisture-Powered Generator from Cigarette Filters, Sugarcane Fibers, Table Salt and Spent Battery Components
Researchers have developed a flexible, paper-like electricity generator that harvests energy from atmospheric humidity using only low-cost waste materials. The device combines fibers from the invasive sugarcane species Saccharum with cellulose extracted from discarded cigarette filters, sodium chloride, and carbon paste recovered from used batteries, topped with an aluminum foil electrode. At around 65% relative humidity, a single cell produces up to 1.16 volts—nearly double the output of most previous moisture harvesters—while four units connected in series delivered 1.79 volts and 0.323 milliamperes, sufficient to light a small red LED continuously for more than three hours. The generator does not store chemical energy like a conventional battery; instead, absorbed water dissolves the salt, creating an ion-concentration gradient that drives charge separation between the carbon and aluminum electrodes. Performance drops sharply when humidity becomes too high or too low because the internal moisture gradient disappears, confirming that ionic movement, not an undisclosed chemical reaction, is the true source of electricity. The approach demonstrates that abundant waste streams can replace expensive nanostructures for low-power environmental sensors and wearable electronics, although the technology remains unsuitable for replacing conventional batteries.
Yandex Maps and Navigator Add Real-Time Fuel Availability and Queue Data Plus Fuel-Efficient Routing Options
Yandex has rolled out major updates to its Maps and Navigator services aimed at helping drivers find gas stations with available fuel and minimal queues. The new features display fuel stock levels and queue sizes directly in each gas station's information card, allowing users to instantly build a route to the chosen location. The same queue and fuel data have also been integrated into the Yandex Go and Zapravki applications. Information is aggregated from multiple sources including anonymized fuel order records, data from taxi drivers using Yandex Pro, real-time traffic conditions near stations, and optional short surveys sent to drivers who have recently refueled or passed by. In addition, the apps now offer economical route suggestions that prioritize lower fuel consumption through smoother speeds and fewer traffic lights, even if the trip takes slightly longer. These capabilities are currently live in Moscow, Saint Petersburg, and 14 other major Russian cities including Kazan, Yekaterinburg, Krasnodar, Novosibirsk, Samara, and Chelyabinsk, with gradual expansion planned as more reliable data becomes available.
Avito Considers Launching In-App Dating Service for Serious Relationships with Emphasis on Safety and Verification
Russian classifieds platform Avito is exploring the possibility of introducing its own dating service directly inside the main application. The company has confirmed that the idea is under active consideration, though no final decision has been made yet. The planned service would be free of charge and specifically oriented toward users seeking serious, long-term relationships rather than casual encounters. A major focus will be placed on security features, including mandatory profile verification, anti-bot measures, three-stage moderation, and protected communication channels, potentially leveraging external services such as Gosuslugi. Avito intends to apply its existing machine-learning technologies, already used across more than 100 models for matching listings, to generate partner recommendations. Before any full rollout, the company plans to test user interest through a Fake Door experiment by presenting the feature as if it already exists. Competitors have reacted calmly, noting that mandatory verification is no longer unique, while some observers question whether users will accept mixing classifieds and dating profiles under a single account.
White House Teleprompter Operator Gabriel Perez Accused of Insider Trading on Kalshi Prediction Market, Earning Over $100,000 from Trump's Speeches
A White House employee exploited his access to President Donald Trump's prepared speeches to place profitable bets on the Kalshi prediction market platform. Gabriel Perez, the teleprompter operator, reportedly earned more than $100,000 by wagering on whether specific words and phrases would be uttered during over a dozen public events, including addresses to Congress and the World Economic Forum. His trading patterns raised red flags when he adjusted or canceled positions after Trump deviated from prepared remarks, prompting Kalshi to detect the suspicious activity and refer the case to the U.S. Commodity Futures Trading Commission. Federal regulators are now negotiating a potential civil settlement that could require Perez to return his profits and cease further trading. The White House placed him on unpaid administrative leave, citing violations of rules against profiting from non-public information, and confirmed that President Trump has been informed of the incident. This case follows earlier enforcement actions by Kalshi against politicians and other insiders, highlighting ongoing challenges in regulating prediction markets.
50,000-Year-Old DNA Found in South African Teeth Where Heat Should Have Destroyed It
Researchers have successfully extracted ancient DNA from animal teeth discovered in the warm coastal caves of South Africa, pushing back the known limits of genetic preservation in hot climates by tens of thousands of years. The study analyzed 320 fossil bones and teeth from six species of wild bovids, with 144 samples undergoing genetic testing that yielded a 45% success rate. A key innovation was the use of single-stranded DNA analysis, which recovered up to 6.7 times more genetic material than traditional double-stranded methods by targeting short, damaged fragments. The oldest successful sample came from a tooth approximately 50,000 years old belonging to an extinct long-horned buffalo, far surpassing the previous record of 9,300-year-old DNA from a blue antelope. These findings demonstrate that ancient DNA can survive in subtropical environments for much longer than previously assumed, opening new possibilities for studying African fauna from the Pleistocene era. The results encourage future paleogenetic research to prioritize Holocene specimens while still testing older material from warm regions.
Bitrix24 Adds Email One-Time Codes as New Two-Factor Authentication Option in Cloud Version
Bitrix24 has introduced email-based one-time codes as an additional two-factor authentication method for its cloud platform, expanding options beyond authenticator apps, push notifications, and SMS. The new feature targets organizations where employees cannot or prefer not to use dedicated 2FA applications or receive text messages reliably. It reinforces account security by requiring a second verification step after login credentials, making unauthorized access more difficult even if passwords are compromised through phishing or leaks. However, the company notes that email as a second factor carries risks if an attacker already controls the user's mailbox. The same capability is planned for the on-premise edition soon, and Bitrix24 intends to mandate two-factor authentication for Professional and Enterprise cloud tariff customers. This update highlights the growing emphasis on flexible yet secure authentication mechanisms in enterprise collaboration tools.
Former OpenAI CTO Mira Murati Launches Thinking Machines' Inkling: Open-Weights Multimodal AI Model with 975 Billion Parameters and Self-Training Demo
Thinking Machines, founded by former OpenAI technical director Mira Murati, has released Inkling, its first open-weights multimodal AI model that supports text, images, and audio in a unified architecture. The model uses a mixture-of-experts design with 975 billion total parameters but activates only 41 billion at once, supports a 1-million-token context window, and was trained on 45 trillion tokens spanning text, images, audio, and video. A smaller Inkling Small variant with 12 billion active parameters was also introduced for faster and cheaper inference. Key innovations include adjustable reasoning depth that lets developers control compute usage per query and a self-training experiment where the model autonomously fine-tuned itself via the Tinker platform to avoid using one letter of the English alphabet. Weights are now available on Hugging Face with support for Transformers, vLLM, SGLang, and llama.cpp, positioning Inkling as a flexible foundation for further customization rather than a direct competitor to closed frontier models.
VK Sells 100% of RuStore to CEO Dmitry Pankrushev, Transferring Full Control to Management Team
VK has reached an agreement to sell its entire stake in the Russian Android app store RuStore to the company's own general director, Dmitry Pankrushev. The transaction moves RuStore out of direct corporate control by VK and places it under the leadership of the team that has been managing its day-to-day operations. No financial details, including the purchase price or payment structure, have been disclosed. VK described RuStore as Russia's single point of access to essential Android services and as a key platform supporting local application developers. After the deal closes, the existing team has committed to continuing regular updates, expanding features, and maintaining stable service for users. On the surface, end-users should notice no immediate changes in availability or functionality, yet the shift represents a significant structural change for one of Russia's most important alternative app marketplaces.
DNA Origami Turns Secrets into Molecular Morse Code: Multi-Key Nano-Safe Resists Single-Party Attacks
Chinese researchers have developed a multi-layered encryption system based on DNA origami that encodes messages as Morse code patterns of dots and dashes on flat DNA rectangles. These structures are then folded into tubes using locking strands, physically concealing the data until the correct molecular key is applied. The approach leverages the programmable properties of DNA to perform cryptographic functions including data encoding, physical hiding through shape change, and controlled release only with a matching set of six unlocking strands. Testing achieved 99.7% efficiency in unlocking, with the full encryption-decryption cycle taking around ten hours on the phrase “JUNE6 INVASION NORMANDY.” The system provides 2,576 possible key combinations and hides message length by standardizing block sizes, making it resistant to physical scanning without the proper key. While too slow for everyday use, the work demonstrates that future information protection may combine mathematical algorithms with physical properties of engineered biomolecules.
Tornyol Develops 40-Gram Autonomous Drones That Hunt Mosquitoes Mid-Air Using Doppler Ultrasound and Sound Detection
American startup Tornyol, backed by Y Combinator, has successfully tested a miniature autonomous drone weighing around 40 grams that can chase and collide with flying insects without chemicals. In its first public demonstration on July 14, the prototype tracked and intercepted a moth inside a controlled indoor facility using external motion-capture equipment. The company aims to reduce mosquito-control costs by a factor of 100 by deploying swarms of ten drones to clear one square kilometer. Future versions will rely entirely on onboard sensors, including ultrasonic emitters and microphone arrays, to detect the Doppler shift created by mosquito wingbeats and distinguish species and sex. Engineers plan to move all computation to the drone itself within weeks and eventually form autonomous swarms capable of navigating urban environments. Although the current test still depended on external infrared tracking and a ping-pong ball as a stand-in target, Tornyol views the flight as the first verified aerial interception toward a scalable, non-chemical mosquito eradication system.
Russia Launches Paid 'Tochka Kontakta' Messenger for Inmates in SIZO and Correctional Facilities
A new messaging platform called 'Tochka Kontakta' has been introduced in Russia to enable communication between relatives and individuals held in pre-trial detention centers (SIZO) and correctional colonies. The service supports text messages, photographs, video clips, and voice recordings, but operates on a strict pay-per-action model with no free tier for basic correspondence. Prices range from 3.09 rubles for short texts up to 200 characters to 35 rubles for longer electronic letters of up to 2,500 characters, while video and voice messages carry separate fees between 4.29 and 7.99 rubles. Paid content subscriptions are also available at 99 rubles per thematic section with ongoing updates. The platform promises a familiar chat interface that includes payment notifications for incoming messages, and developers plan to distribute it through all major app stores. However, key operational details such as the identity of the service operator, the message moderation process, and the list of participating institutions remain undisclosed.
Intel Unveils Starfire Chip Designed for AI Processing on US Military Satellites in Orbit
Intel has announced the Starfire system-on-chip, a radiation-tolerant processor developed specifically for orbital deployment aboard US government satellites. The chip integrates eight CPU cores, a three-die neural accelerator, and a graphics block using Foveros packaging, delivering up to 75 TOPS at 35 W. It combines Intel’s advanced 18A process for the CPU and AI blocks with the older Intel 3 node for the GPU, relying on architectural hardening rather than process maturity to withstand space radiation. Unlike legacy rad-hard processors such as BAE Systems’ RAD750, Starfire targets on-board AI inference and data processing instead of basic telemetry and control. The device remains a demonstration platform whose radiation performance is still under evaluation, with first test samples scheduled for the third quarter of 2026. Intel Government Technologies will handle distribution, and the project supports Pentagon programs RAMP-C and SHIP while Intel’s 18A yield roadmap extends into 2027.
Home Internet Tariffs in Russia Rise Up to 20% in First Half of 2026 Amid Rising Costs and Surging Demand
Tariffs for residential broadband in Russia accelerated sharply in the first half of 2026, climbing an average of 8-10% and reaching as much as 20% in some regions and with certain providers. This marks a significant increase compared with the modest 5-7% growth recorded across the entire previous year of 2025. Major operators including Rostelecom and ER-Telecom implemented price hikes ranging from 50-60 rubles to around 100 rubles per month, while some legacy plans became more expensive and previously free connection fees were introduced. The primary drivers cited are higher costs for equipment, electricity, employee wages, infrastructure leasing, and ongoing network modernization projects. Additional pressure came from springtime restrictions on mobile data that pushed large numbers of users onto fixed-line Wi-Fi connections, forcing providers to handle substantially higher traffic volumes. Regulators at the Ministry of Digital Development and the Federal Antimonopoly Service have stated they are monitoring the justification for these increases and may launch inspections, yet industry participants warn that further rises of 12-15% could occur in the second half of the year if infrastructure access costs continue to climb.
Poisson's Spot: 200-Year-Old Optical Effect That Proved Light Is a Wave Now Enables Simple Creation of Optical Skyrmions for Future Photonic Computing
Researchers have discovered a remarkably simple method to generate optical skyrmions—stable topological structures in light fields—by directing a laser at a small opaque disk and exploiting the Poisson spot, also known as the Arago spot. This phenomenon, first observed over 200 years ago, played a pivotal role in confirming the wave nature of light during debates between particle and wave theories in the early 19th century. Unlike previous approaches that required expensive and precisely engineered metamaterials, the new technique uses only a coherent laser beam and a basic circular obstacle, allowing multiple types of skyrmions to form naturally in the bright central spot within the shadow. The experiment produced up to four varieties simultaneously, including spin skyrmions, Stokes skyrmions, and structures tied to electric and magnetic field vectors, all exhibiting topological stability that preserves their form under minor distortions. These properties make optical skyrmions promising candidates for encoding and transmitting information in future photonic computers, data storage systems, and high-speed communication technologies. The approach dramatically lowers the barrier for laboratories to reproduce and study these complex light configurations, potentially accelerating research into topology-protected data handling that is resilient to environmental noise.
Bankrupt After Just Six Weeks of Production Shutdown: How a Cyber Attack Killed a 37-Year-Old German Textile Manufacturer and Exposed the Cruel Reality of Modern Cyber Threats
A 37-year-old German textile processing company, ZEGO Textilveredelungszentrum, has filed for insolvency after a cyber attack halted its production lines for nearly six weeks, demonstrating that business interruption alone can destroy even well-established manufacturing firms without any data theft or ransom demands. The firm, based in Bavaria and serving automotive, workwear, and technical textiles industries, suffered the attack on March 29, 2026, leading to irrecoverable financial losses despite eventual system recovery. Managing Director Johannes Zenglein described the decision as one of the most difficult in the company's history, noting that the prolonged downtime caused severe cash flow disruption, lost orders, and customer attrition. The incident highlights a growing trend where cyber attacks on industrial systems lead directly to bankruptcy, as seen in prior cases like the 158-year-old British transport company Knights of Old and a German mobile phone repair firm. Key lessons include the critical need for robust business continuity plans, quantified downtime cost assessments, and supply chain resilience evaluations beyond traditional security measures. Unlike typical ransomware events, this attack required no encryption or extortion to achieve devastating results, underscoring that operational resilience is now a matter of corporate survival.
t.me Domain Restored in DNS After Sudden Outage, But Full Recovery for Telegram Links May Take Up to 24 Hours
The short domain t.me has been restored to active DNS status after the .me registry removed the restrictive serverHold flag that had taken it offline worldwide. DNS records have been reinstated, allowing t.me links to function again in browsers, although propagation delays caused by caching at ISPs and recursive resolvers mean not all users will see the change immediately. The outage occurred when the registry-level serverHold status was applied, preventing resolution of the domain despite Telegram continuing to operate normally. As a precaution, Telegram began automatically substituting t.me links with telegram.me inside its mobile and desktop applications. Old t.me addresses continued to work when opened directly within the messenger, but external browsers were unable to reach them. The exact cause of the serverHold status remains unknown, with possibilities including a technical error, legal request, or deliberate action by the registry operator. The domain is now formally active again, and full global visibility depends on DNS cache expiration across networks.
Generative AI Can Clone and Modify Android Apps for as Little as 88 Kopecks, Positive Technologies Warns
Positive Technologies researchers have demonstrated that generative AI models can successfully modify, reassemble, and preserve the functionality of Android applications in just minutes at extremely low cost. In experiments involving 90 diverse commercial apps, closed-source AI models achieved an 84% success rate while open-weight models reached 61%, requiring an average of 14 iterations and between 5.5 and 9 minutes per attempt. The cost of each successful modification ranged from 0.88 to 40.89 rubles, meaning an attacker could attempt to clone around 100 popular applications for only a few thousand rubles. Although the researchers inserted only neutral changes rather than malicious code, the same technique could be used to embed data interception, alter app behavior, or add connections to external command servers. The resulting fake APKs can then be distributed through third-party stores, websites, messengers, and online communities, primarily threatening users who sideload applications outside official marketplaces. The study shows that AI does not create entirely new attack vectors but dramatically lowers the technical barrier that previously required advanced reverse-engineering skills.
Hidden Spy for 1.6 Million Users: Popular Browser Extension ModHeader Secretly Collected Browsing History
Google and Microsoft have removed the popular ModHeader browser extension from the Chrome Web Store and Microsoft Edge Add-ons after security researchers discovered a hidden mechanism designed to secretly collect users' browsing history. The extension, which had approximately 1.6 million installations, allowed developers to modify HTTP headers for testing and debugging purposes but contained a dormant data-collection module in its legitimate codebase. British firm Stripe OLT confirmed that the suspicious code was part of the genuine signed build rather than a fake version. Although the history-stealing functionality remained inactive due to an empty internal browser list, the extension still transmitted telemetry data and could have been activated remotely via a simple update. Experts recommend immediate removal of the extension, replacement of any credentials entered through it, and blocking of the domains stanfordstudies[.]com and extensions-hub[.]com.
Drawn-on-Skin Conductive Ink Electrodes Replace Traditional Sensors: Washable, Customizable Designs Enable Long-Term ECG, EMG, and EEG Monitoring
Engineers have developed a water-based conductive ink that can be painted directly onto the skin using a brush to form flexible, washable electrodes in any shape or color, including decorative designs like animals or patterns. The ink dries within about ten minutes into a thin, stretchable conductive layer that conforms to skin microrelief, maintains stable electrical contact during movement and sweating, and can be easily removed with water. This approach solves common problems with rigid metal electrodes and hydrogels, such as poor adhesion, air gaps, and signal distortion during physical activity. The system pairs disposable painted electrodes with a reusable electronic module connected via porous silver fabric, enabling Bluetooth transmission of biosignals for up to 12 hours of continuous monitoring. Tests demonstrated reliable electrocardiogram recording during daily activities and exercise, electromyography for controlling robotic prosthetics, and potential uses in pediatrics, neurointerfaces, and even plant monitoring. The innovation, detailed in Proceedings of the National Academy of Sciences with a preliminary patent filed, promises more comfortable and precise long-term biosignal tracking without complex fabrication.
ChatGPT Returns to WhatsApp After EU Forces Meta to Reopen Business API to Rival AI Bots
Home users of generative AI services have begun seeing <b>ChatGPT</b> working again inside <b>WhatsApp</b>, owned by <b>Meta</b> (recognized as extremist and banned in Russia). The partial restoration follows an EU antitrust investigation that accused Meta of abusing its dominant position to favor its own AI assistant. <b>OpenAI</b> originally launched the integration in 2024, allowing users to message the chatbot like any regular contact without extra apps. In 2025 Meta updated its <b>Business API</b> rules, effectively blocking third-party universal chatbots and pushing competitors out. After the European Commission intervened, Meta was required to reopen access, and <b>ChatGPT</b> is now gradually reappearing for some users. The rollout remains uneven, with some contacts responding normally while others stay silent, and no paid subscription is required. Neither Meta nor OpenAI has officially linked the return to the EU decision.
Google Enhances Android Backup Controls in Play Services Update, Adds Document Sync to Google Drive Amid Storage Limit Cut
Google is updating its Android backup system through Google Play Services version 26.25, introducing separate toggles for messaging backups and the ability to automatically save documents to the cloud. Users can now disable SMS, MMS, and RCS backups independently, though RCS remains nested under the MMS category for less visibility. The new document backup feature supports formats including DOC, PPT, XLS, and PDF, storing copies in a device-named folder on Google Drive without automatic synchronization to the original files. These changes coincide with Google's reduction of free storage from 15 GB to 5 GB, where Android backups now count against the quota, potentially turning small 40 MB backups into much larger archives. The update provides more granular control over what gets backed up, including call history and system settings, accessible via Settings > Accounts and backup > Google Backup > Other device data. While the feature can help preserve important files, it risks quickly exhausting limited free storage if documents are included.
The Formula Remains the Same, but the Mathematics Changes: Physicists Show Quantum World Can Be Described Without Imaginary Numbers
Researchers from Heinrich Heine University Düsseldorf and the German Aerospace Center have developed a new mathematical framework for quantum mechanics that relies solely on real numbers instead of complex numbers. By relaxing one overly restrictive assumption from a 2021 study that concluded imaginary numbers were indispensable, the team created an entire family of real-valued theories capable of reproducing all predictions of standard quantum mechanics. The work demonstrates that complex numbers, while convenient, are not necessarily a fundamental requirement of nature but rather one possible mathematical tool for describing quantum phenomena such as wave-particle duality, tunneling, and entanglement. The new approach maintains identical experimental outcomes for any conceivable test, making the two formulations indistinguishable through measurement alone. Although the authors do not advocate replacing the established complex-number formalism, their findings open fresh perspectives on the mathematical foundations of quantum theory and its applications in quantum computing and secure communication.
Near Absolute Zero and 0.55% Efficiency: World's First Superconducting Quantum Heat Engine Demonstrated by Aalto University Researchers
Researchers from Aalto University have built the first experimental cyclic quantum heat engine operating inside a superconducting quantum circuit, achieving a full repeating Otto cycle near absolute zero. The device uses a transmon qubit as the working substance, coupled with a resonator and a quantum refrigerator that can heat or cool the qubit on demand, replacing separate hot and cold reservoirs with a single voltage-controlled component. Experimental measurements over three consecutive cycles confirmed positive net work output, with an average power of 0.039 electronvolts per second and an efficiency of approximately 0.55 percent, reaching 27 percent of the theoretical Otto-cycle limit for the chosen parameters. Simulations suggest that efficiency could approach 2.2 percent once the system reaches steady-state operation. Beyond energy conversion, the technology offers a promising route to reduce the number of external microwave control lines in large-scale superconducting quantum processors by enabling autonomous on-chip operations such as state preparation and readout. The work also provides a new experimental platform for testing predictions of quantum thermodynamics under precisely controlled conditions.
Study Reveals That Names Gradually Shape Adult Facial Appearance Through Societal Expectations, Not Genetics Alone
New research demonstrates that a person's given name can influence their physical appearance over time as they unconsciously conform to social stereotypes associated with that name. The study compared photographs of children and adults, finding that participants could accurately guess names from adult faces at rates significantly above chance, while the same task proved nearly impossible with children's images. Machine learning algorithms reinforced these findings by detecting greater facial similarity among adults sharing the same name, an effect absent in younger subjects. To rule out natural aging as the cause, researchers artificially aged childhood photos, yet the resulting images still failed to match name-based expectations. The phenomenon is attributed to social structuring, where repeated societal feedback creates a self-fulfilling prophecy that subtly alters expressions, posture, and facial habits. This effect highlights how even minor social labels can leave measurable marks on human appearance by adulthood.
Tired of Instant Messaging, Users Flock to Virtual Pigeons: Roost App Revives Postal Delivery with Birds and Animals
Roost is a new messaging application that deliberately replaces instant delivery with delayed virtual bird flights, turning the wait itself into the core user experience. Users select couriers ranging from fast falcons to slow snails and turtles, with delivery times calculated from real animal speeds and actual geographic distances displayed on an interactive map. The app combines messenger functions, social networking, and a collectible creature-training game featuring over one thousand birds and animals that can be leveled up through mini-games. Created by Ticketmaster trust and safety manager Logan Mendelson as a small experiment, Roost went viral after a Threads post by Karen Lewis, growing from 10,000 to over 300,000 users in weeks without any paid marketing. Despite its playful design, the service raises notable privacy concerns because it requires precise location data, stores messages in plaintext, and sends content to OpenAI for moderation while lacking end-to-end encryption.
Scientists Develop Eye-in-a-Care-Box Device to Keep Whole Eyes Alive and Responsive to Light for 10 Hours After Death, Advancing Future Transplant Possibilities
Researchers have created a novel perfusion system called Eye-in-a-Care-Box (ECaBox) that successfully maintains the viability of whole eyes outside the body for extended periods after death. The device uses 3D-printed components and continuous oxygenated fluid circulation through the ophthalmic artery to supply oxygen and nutrients to delicate retinal tissues that typically degrade rapidly without blood flow. Experiments on pig eyes showed preserved cellular structure and electrical responses to light flashes via electroretinography even 10 hours post-mortem when perfusion began promptly, outperforming standard cold storage. Similar benefits were observed in limited tests with human donor eyes, though ethical constraints limited sample sizes and no light-response testing was performed on human retinas. The work builds on a landmark 2024 face-and-eye transplant case and highlights potential applications beyond transplantation, including ex vivo studies of macular degeneration without live animal testing. Challenges remain in reconnecting the optic nerve for vision restoration, but the technology could enable portable donor-eye preservation to reduce critical time delays.
Virtual Herd of Cows Solves Century-Old Mystery of Terracettes on Slopes That Puzzled Darwin
A new computer model demonstrates that grazing animals such as cows and sheep can independently create the striking stepped patterns known as terracettes on steep hillsides through simple energy-saving behavior. Researchers modeled individual virtual animals that only sought food while minimizing movement costs, without any pre-defined paths. On steep slopes, the animals gradually preferred horizontal routes because climbing or descending required more energy, and each step compacted soil and removed vegetation, reinforcing those paths for subsequent animals. This feedback mechanism, called stigmergy, led to the emergence of dozens of parallel, evenly spaced terraces after thousands of steps, closely matching real-world examples from the Alps, Britain, and other regions. The findings revive and strengthen observations made by Charles Darwin in 1881, who noted similar features and enlisted his son to count nearly 30 parallel steps on a 40-degree chalk slope in England. While soil creep and other geological factors may also contribute, the model proves that chaotic animal movement alone is sufficient to produce highly ordered landscape patterns without external coordination.
Legacy of Hawking, Penrose and Zeldovich: US Laboratory Recreates Rotating Black Hole Physics with Synthetic Superluminal Rotation
Physicists at the Advanced Science Research Center of the City University of New York have successfully reproduced the core physical principle behind energy extraction from rotating black holes in a stationary laboratory setup. By using a ring of electronic resonators whose parameters were rapidly modulated in sequence, the team created an artificial rotational pattern that interacted with radio waves exactly as a superluminal rotating medium would. The experiment directly demonstrates the Penrose process and Zeldovich’s rotational superradiance without any mechanical motion, gravity, or event horizon. Radio waves carrying the correct angular momentum were amplified by drawing energy from the electronic control system that continuously altered resonator properties. Published in Nature, the work belongs to the emerging field of Floquet engineering and opens pathways for studying extreme Doppler shifts and wave interactions in controlled tabletop environments. Potential future applications include advanced wireless communications, photonics, signal processing, and quantum devices, although significant miniaturization remains necessary before practical deployment.
Universe Proves Ordinary After All: High-Profile DESI Study Claiming Giant Aligned Cosmic Structures Collapses Following Data Correction
A widely publicized study published in Nature that used Dark Energy Spectroscopic Instrument (DESI) observations to suggest the existence of unusually long, preferentially aligned filaments in the cosmic web has been overturned by independent analysis. The original paper analyzed data on 47 million galaxies and quasars spanning more than 11 billion years and argued that these structures could stretch across billions of light-years and show non-random orientations, potentially violating the cosmological principle. Critics identified two critical errors: the use of luminosity distance instead of the standard comoving distance and the failure to account for the evolving distances caused by the universe’s expansion. Once these corrections were applied, the apparent giant aligned structures vanished, restoring consistency with the standard cosmological model in which the universe appears statistically uniform on the largest scales. The episode has reignited debate over the rigor of peer review at top journals, the risks of embargoed pre-publication announcements, and the importance of releasing results on preprint servers such as arXiv for community scrutiny before formal publication.
Europe Unveils RLV C5 Reusable Rocket Concept as a Potential Competitor to SpaceX Starship
Following the successful recovery of Super Heavy booster using tower catch arms, SpaceX's Starship continues to push the boundaries of fully reusable super-heavy launch systems. A detailed independent analysis by the German Aerospace Center (DLR) reconstructed flight data from early Starship tests and estimated current payload capacity at around 59 tons to low Earth orbit in fully reusable configuration. The study also modeled future Starship variants with Raptor 3 engines, projecting up to 115 tons reusable and 188 tons expendable. In response, DLR presented its own European super-heavy concept called RLV C5, which features a reusable winged booster based on SpaceLiner technology and an expendable upper stage. The design uses liquid hydrogen/oxygen propulsion and an innovative mid-air capture system by a large subsonic aircraft, aiming for higher payload-to-mass efficiency than Starship. While Starship already undergoes real flight testing, RLV C5 remains a paper study requiring significant further development and funding.
CAR-T Therapy Expands Beyond Leukemia: GPNMB Protein Emerges as Promising Target for Glioblastoma and Rare Soft Tissue Sarcoma
CAR-T therapy, which has transformed treatment for certain blood cancers like leukemia by engineering patient T-cells to recognize specific surface proteins, has long struggled against solid tumors due to cellular heterogeneity and immunosuppressive microenvironments. Two independent research teams have now identified the surface protein GPNMB as a dual-purpose target that appears on both tumor cells and tumor-supporting immune cells. In glioblastoma models, GPNMB-directed CAR-T cells eradicated patient-derived tumor cells, slowed disease progression in mice, and attacked the protective immune niche. Parallel work on alveolar soft part sarcoma showed similar efficacy in mice, prevented metastasis, passed human-skin safety tests, and achieved disease stabilization in the first treated patient after a single infusion. The studies highlight how targeting proteins shared between cancer cells and their immunosuppressive surroundings could overcome key barriers that have limited CAR-T success in solid tumors, although larger clinical trials are still needed.
Bacteria Convert Toxic Dissolved Uranium into Rare Stable Compound FeU(V)O₄, Paving Way for Microbial Nuclear Waste Remediation
Scientists have discovered that naturally occurring bacteria from a flooded uranium mine in the Ore Mountains can transform dissolved uranium into a rare, stable pentavalent compound called FeU(V)O₄ using only glycerol as an additional carbon and energy source. In laboratory experiments mimicking the oxygen-free conditions found 2 km underground, the microbial community reduced dissolved uranium levels to just 5% of the original concentration within 130 days. Advanced synchrotron analysis at ESRF in France revealed that the bacteria not only accumulate uranium on cell walls but actively produce the previously little-known FeU(V)O₄ mineral, first identified in 2020 in uranium-contaminated Croatian soil. Remarkably, the compound remained stable and even continued to form when the bacterial biomass was dried and exposed to oxygen. This biological process could offer a cost-effective, nature-based approach to cleaning uranium-polluted water and soil, although further research is needed to scale it beyond laboratory conditions and identify the specific microbial species involved.
Doubling GPT-3 Inference Speed and Eliminating Silicon Furnaces: Vertical Memory Architectures V-Die and MOSAIC Revolutionize HBM Cooling and Bandwidth for AI Accelerators
Researchers from South Korea and Japan have unveiled two groundbreaking vertical memory stacking technologies, V-Die and MOSAIC, at the June 2026 IEEE VLSI Symposium that promise to dramatically increase HBM capacity and performance while solving critical heat dissipation problems in AI accelerators. Instead of stacking DRAM dies horizontally with TSV interconnects, both approaches rotate the memory packages sideways so individual dies stand vertically like radiator fins, enabling better cooling and more interconnects. The Korean V-Die design eliminates TSVs entirely, provides four times more connections than HBM4, reduces read latency by 37 percent, and uses microfluidic channels to keep temperatures around 45 °C, delivering 540 tokens per second on GPT-3-scale workloads versus 296 tokens with conventional HBM4 in simulations. The Japanese MOSAIC project replaces physical contacts with miniature inductive coils that enable wireless data transfer at up to 4 Gbit/s per channel, allowing up to 98 dies and 294 GB of memory (theoretically 882 GB) to be placed directly above a GPU with only a one-degree temperature rise. Both technologies remain in the research or early prototype stage and still face manufacturing and cost challenges, yet they demonstrate a viable path around the thermal wall that currently limits further scaling of high-bandwidth memory for next-generation AI hardware.
Rosreestr to Integrate AI with Satellite Imagery for Faster and More Efficient Map Updates Across Russia
Rosreestr plans to deploy artificial intelligence to accelerate the updating of cartographic materials by analyzing data from space-based remote sensing. The agency has already begun transitioning to satellite imagery for mapping the entire territory of Russia and now intends to add AI capabilities to detect changes automatically and reduce manual work. Following a signed roadmap with Roscosmos, a pilot project is being launched to monitor terrain changes using orbital images. Deputy Head Maxim Smirnov highlighted that the technology will increase update speed, boost labor productivity, and improve the efficiency of budget spending. IT companies are invited to propose solutions that combine space data with vector and registry information at various scales, after which Rosreestr will select the most suitable monitoring technology. The initiative aims to replace slow, labor-intensive manual map updates with automated satellite analysis, which is especially practical for a country of Russia’s vast size.
Physicians Fail to Detect Flawed AI Treatment Recommendations Even When Evidence Contradicts Them, New Study Finds
A new study involving 223 physicians has revealed that doctors continue to trust artificial intelligence recommendations for patient treatment even when real-world results clearly show the therapy provides no benefit. Participants were asked to evaluate an experimental therapy for a rare disease after receiving AI-generated suggestions that divided patients into groups expected to respond differently. In controlled experiments, researchers deliberately made the treatment either only moderately effective across all patients or completely ineffective, yet most doctors did not revise their assessment of the AI system. The findings point to a strong automation bias, where clinicians tend to over-rely on algorithmic outputs that appear objective and data-driven, making them less likely to question contradictory evidence. Even physicians with substantial medical experience struggled to override the initial AI guidance when presented with outcome data. The research underscores that simply having a human in the decision loop does not guarantee detection of AI errors in high-stakes medical contexts. Authors recommend additional safeguards such as independent case review before viewing AI suggestions and mandatory explanations when agreeing with algorithmic advice.
Why Simple ChatGPT Wrappers Are Losing Value: AI Model Makers Integrate Features, Pushing Startups to Control Full Industry Processes
Early startups that built businesses by wrapping large language models like ChatGPT are rapidly losing their edge as OpenAI, Anthropic, and other developers directly embed specialized capabilities into their core platforms. According to the venture firm NFX, startups must now move beyond narrow add-on tools and instead take ownership of complete industry workflows, such as building full legal services or end-to-end supply chain management systems. The initial wave of wrapper companies focused on tasks like generating ad copy, handling customer support, assisting sales teams, and searching legal documents, expecting base models to remain limited in specialized domains. This assumption proved incorrect as major AI providers added robust features for coding, document handling, and content creation, eroding the competitive advantage of single-task startups. Notable examples include Jasper, which raised $125 million at a $1.5 billion valuation before pivoting after revenue misses and layoffs, alongside more resilient firms like EvenUp, Blitzy, Tomo, and Seso that embed AI deeply into comprehensive services. Even large legal practices such as Freshfields are partnering directly with model developers, though these efforts require significant compute resources and in-house expertise. NFX emphasizes that deep domain knowledge, proprietary data, established sales channels, and control over entire service pipelines are far harder to replicate than isolated AI features.
Beats Normally — And Still Fails: Soft Robotic Heart Model Created to Study HFpEF, the Most Insidious Form of Heart Failure
Researchers have developed a soft robotic heart model capable of dynamically changing its stiffness during operation to accurately replicate heart failure with preserved ejection fraction (HFpEF). The device uses fluid-pressure-responsive artificial muscles made from reinforced rubber tubes layered on a silicone base to mimic the left atrium and ventricle. Unlike previous models that followed fixed programs, this system features a closed-loop control mechanism where internal pressure directly influences muscle behavior and vice versa, allowing realistic simulation of diastolic dysfunction. The model can progressively restrict ventricular expansion to reproduce multiple stages of the disease, from early relaxation delays to severe filling impairment. With HFpEF accounting for roughly half of all heart failure cases and affecting over three million people in the United States alone, the platform offers a much-needed laboratory tool for testing therapies that target myocardial stiffness rather than just symptoms.
Scientists Consider Replacing Leap Seconds with a Century-Scale 'Extra Hour' to Protect Global Digital Infrastructure from Negative Leap Second Risks
International timekeeping authorities are accelerating plans to overhaul the leap second system by allowing a much larger gap between atomic time and Earth's rotation, potentially requiring the next major adjustment only after several centuries. The proposal, which could take effect as early as 2027, aims to eliminate the disruptive practice of adding or removing individual seconds from UTC. Leap seconds, introduced in 1972, have repeatedly caused outages at companies including Meta, Reddit, and Cloudflare, while also affecting aviation and high-frequency trading. Because Earth's rotation has been accelerating since 2016, experts now face the unprecedented risk of a negative leap second that would remove one second from UTC—an event whose impact on modern systems remains unknown. With a roughly 30 percent chance of needing such a negative adjustment before 2035, the General Conference on Weights and Measures is pushing for a larger tolerance between coordinated and astronomical time. The new approach would preserve the long-term link between civil time and Earth's rotation but make it far less rigid, allowing atomic clocks to govern daily operations without sudden one-second jumps.
Startup MorphMind Unveils Academic Humanizer to Help AI-Generated Research Papers Evade Detection by Mimicking Human Academic Style
MorphMind has released Academic Humanizer, a specialized tool that rewrites AI-generated academic drafts to remove machine-like patterns and align them with an individual researcher’s voice. The service operates as a Claude skill, allowing users to upload AI-created manuscripts along with examples of their previous publications so the output better reflects personal style while preserving scientific accuracy. Developers stress that the tool performs only editing and clarity improvements and does not generate new findings, data, or citations. The project’s GitHub description was revised after early wording about “removing AI traces” raised concerns, with co-founder Jie Ding clarifying that it is not intended to bypass peer review. Despite these safeguards, experts warn that the tool could make weak or fabricated papers appear more convincing, complicating detection efforts by journals and universities already dealing with a surge of AI-written submissions containing fabricated references.