119 place 0
Processing 200,000 tokens through a large language model is expensive and slow: the longer the context, the faster the costs spiral. Researchers at Tsinghua University and Z.ai have built a technique called IndexCache that cuts up to 75% of the redundant computation in sparse attention models, delivering up to 1.82x faster time-to-first-token and 1.48x faster generation throughput at that context length.The technique applies to models using the DeepSeek Sparse Attention architecture, including the latest De
A newsletter a day!
You may get 10 most important news around midday in daily newsletter. Press the button and we will send you the most important news only, no spam attached.
LIKE us on Facebook so you won't miss the most important news of the day!
VentureBeat
An East Bay apartment complex has been bought at a price that's well below its prior value. Read more ›
0
A PG&E Corp. unit has bought a San Jose building in a move to bolster the utility's South Bay operations. Read more ›
0
I tested Chipotle's portion sizes after the CEO's viral comments and found that whether you can get extra add-ons for free depends on who you ask. Read more ›
0 newcommer
Если вы прямо сейчас думаете, внедрять вам ИИ или нет, то сначала ответьте себе: а вы уже посчитали ROI? Читать далее Read more ›
0 newcommer
Dell Technologies has introduced a keyboard and mouse combo that charges in five seconds and delivers a full day of use. The new Dell Pro 7 Rechargeable Compact Keyboard and Mouse rely on supercapacitor technology rather than traditional lithium-ion batteries. The Dell Pro 7 Rechargeable Compact Keyboard and Mouse replace traditional lithium-ion batteries with supercapacitor […] Read more ›
0 fresh
Статья для руководителей, которые хотят нанимать сотрудников быстро и не терять новичков на испытательном сроке. Делимся своим опытом, рассказываем, когда есть смысл проводить групповые интервью и как это делать бережно по отношению ко всем участникам. Читать далее Read more ›
0 newcommer
Ethos, the AI platform that matches skilled professionals to income opportunities, has secured a $22.75M (£16.7m) Series A funding round. While AI has made it easier to generate CVs and job applications, it has also made it harder to identify expertise. Organisations are facing volumes of increasingly similar profiles, with limited ways to assess what ... Read more ›
0 fresh
CodeWords, a London-based AI agent platform that lets users run their business on autopilot, has raised a €7.6 million ($9 million) Seed round to fund the expansion of its go-to-market and engineering teams. The round was led by Visionaries, with participation from Firstminute Capital (which first backed the company at pre-Seed), Sequel, and Illusian, the ... Read more ›
0 fresh
На связи Сергей Скирдин, технический директор компании «Белый код». Мы занимаемся проектами в сфере управления данными: интеграции, хранилища, BI. В прошлой статье про DevCon я писал, что спрашивал про поддержку «1С:Шины» в БСП, чтобы не делать на каждом проекте отдельный коннектор. Конкретных сроков от вендора не прозвучало, и в конце статьи я оставил приглашение к сотрудничеству для тех, у кого есть готовый коннектор. Откликнулась компания «Денвик». Мы с ними давно... Read more ›
0 fresh
Claude's new Dreaming feature is a between-session memory refinement system that reviews past agent behaviour, identifies recurring mistakes and workflow patterns, and updates memory. Read more ›
0 fresh
В части 8, когда мы говорили о том, что нет универсального инструмента для работы в различных концах географии, я объяснял это тем, что имеются местные особенности, и то, что работает в одном месте, не работает в другом. — Как быть, если нужно быстро найти инструменты для исследования в различных регионах, которые находятся на противоположных концах географии? Скажу сразу, швейцарского ножа не существует. Это плохая новость. Хорошая же новость состоит в... Read more ›
0 fresh
Пока мы воспринимаем свои промпты как обычный текст из символов, для LLM они в виде токенов «выглядят» совсем иначе. И если не осознавать этого, порой можно наткнуться на проблемы. Поэтому полезно (и интересно) понимать: что вообще представляют собой токены? По какому алгоритму текст преобразуют в них и обратно? Какие важные нюансы при этом возникают?Возможно, подробнее и понятнее всех объяснил пару лет назад ИИ-рисерчер Андрей Карпатый, записав двухчасовое видео на английском.... Read more ›
0 fresh
Qualcomm's new Snapdragon 6 Gen 5 and 4 Gen 5 chips bring AI cameras, smoother displays, and better gaming to mid-range phones launching later this year. Read more ›
0 fresh
Sedivention,a Germany-based medtech startup, has raised €2.9 million in a seed fundinground led by bmp Ventures alongside the IBG funds. Additional investors includethe strategic investment arm of a g... Read more ›
0 fresh
Зимой клубника в Великобритании долгое время была почти экзотикой. Полки супермаркетов заполнялись импортом из Северной Африки и Ближнего Востока по заоблачным ценам. Для большинства руководителей это просто рыночный факт. Для Джеймса Дайсона — задача, которую можно решить инженерно. Рассказываем, как у производителя пылесосов появился один из самых необычных агропроектов в Европе: ферма, где клубнику выращивают в подвесных конструкциях, собирают с применением роботов и сохраняют в идеальных условиях с помо Read more ›
0 fresh
Rumors about the iQOO 15T and iQOO Pad6 Pro surfaced late last month, pointing to a May launch. iQOO has now officially teased both devices and opened pre-reservations in China. While no details about the devices have been confirmed, iQOO has teased the design of the 15T. It appears to have a square rear camera module. Meanwhile, tipster Digital Chat Station had recently shared an image of the alleged iQOO... Read more ›
0 fresh
Stockholm-based Pit, anAI-native platform that replaces the patchwork of spreadsheets, inboxes, andrigid SaaS tools used in enterprise operations, has announced its publiclaunch. The company also rais... Read more ›
0 fresh
VentureBeat
Netomi, the San Francisco-based startup building AI systems for enterprise customer service, said Thursday that it has raised $110 million in new funding in a round led by Accenture Ventures, with participation from Adobe Ventures, WndrCo, Silver Lake Waterman, NAVER Ventures, Metis Strategy and Fin Capital. Jeffrey Katzenberg, managing partner of WndrCo and co-founder of DreamWorks, has joined the company's board. The round builds on early backing from a roster... Read more ›
0
Writer, the enterprise AI agent platform backed by Salesforce Ventures, Adobe Ventures, and Insight Partners, today launched event-based triggers for its Writer Agent platform, enabling AI agents to autonomously detect business signals across Gmail, Gong, Google Calendar, Google Drive, Microsoft SharePoint, and Slack — and execute complex multi-step workflows without any human initiating the process.The release, which also includes a new Adobe Experience Manager connector and a suite of enhanced... Read more ›
0
On March 30, BeyondTrust proved that a crafted GitHub branch name could steal Codex’s OAuth token in cleartext. OpenAI classified it Critical P1. Two days later, Anthropic’s Claude Code source code spilled onto the public npm registry, and within hours, Adversa found Claude Code silently ignored its own deny rules once a command exceeded 50 subcommands. These were not isolated bugs. They were the latest in a nine-month run: six... Read more ›
0
AI is more than a technology — it's magic.Don't believe me? Why, then, is one of the leading companies in the space, OpenAI, publishing entire official, corporate blog posts about goblins?To understand, we first have to go back to earlier this week, on Monday, April 27, 2026, when a developer under the handle @arb8020 on the social network X posted a snippet from the OpenAI open source Codex GitHub repository,... Read more ›
0
Runpod, the high-performance cloud computing and GPU platform designed specifically for AI development, today launched a new open source, MIT licensed, enterprise-friendly Python programming tool called Runpod Flash — and it is poised to make creation, iteration and deployment of AI systems inside and outside of foundation model labs much faster. The tool aims to eliminate some of the biggest barriers and hurdles to training and using AI models today,... Read more ›
0
One of the key challenges of building effective AI agents is teaching them to choose between using external tools or relying on their internal knowledge. But large language models are often trained to blindly invoke tools, which causes latency bottlenecks, unnecessary API costs, and degraded reasoning caused by environmental noise. To overcome this challenge, researchers at Alibaba introduced Hierarchical Decoupled Policy Optimization (HDPO), a reinforcement learning framework that trains agents... Read more ›
0
Presented by TeamViewerEnterprise technology failures are largely invisible. Research from TeamViewer, based on a global survey of 4,200 managers and employees, finds that the majority of digital dysfunction never reaches the IT help desk. Employees work around slow applications, failed logins, and intermittent glitches rather than reporting them, leaving organizations without an accurate picture of how their technology is performing. The cumulative cost is significant: employees lose an average of... Read more ›
0
While Elon Musk faces off against his former colleague and OpenAI co-founder Sam Altman in court, Musk's rival firm xAI, founded to take on OpenAI, isn't slowing down on launching competitive new products and services.Last night, xAI shipped a new, proprietary base large language model (LLM), Grok 4.3, and a new voice cloning suite on the web. The new products arrive after months of tumult from xAI that saw all... Read more ›
0
The scaffolding layer that developers once needed to ship LLM applications — indexing layers, query engines, retrieval pipelines, carefully orchestrated agent loops — is collapsing. And according to Jerry Liu, co-founder and CEO of LlamaIndex, that's not a problem. It's the point.“As a result, there's less of a need for frameworks to actually help users compose these deterministic workflows in a light and shallow manner,” Jerry Liu, co-founder and CEO... Read more ›
0
Anthropic created the Model Context Protocol as the open standard for AI agent-to-tool communication. OpenAI adopted it in March 2025. Google DeepMind followed. Anthropic donated MCP to the Linux Foundation in December 2025. Downloads crossed 150 million. Then four researchers at OX Security found an architectural problem that affects all of them.MCP's STDIO transport, the default for connecting an AI agent to a local tool, executes any operating system command... Read more ›
0
Most popular sources
AlleyWatch
|
0% |
|
VentureBeat
|
0% |
StartupNation
|
0% |
ReadWrite
|
0% |
Tom's Hardware
|
0% |
| View sources » | |
LIKE us on Facebook so you won't miss the most important news of the day!
07.05.2026 04:23
Last update: 04:15 EDT.
News rating updated: 11:13.
What is Times42?
Times42 brings you the most popular news from tech news portals in real-time chart.
Read about us in FAQ section.
