57 place 1

137 Researchers baked 3x inference speedups directly into LLM weights — without speculative decoding

VentureBeat
VentureBeat 3 place · 02/23/2026 12:00 EDT

As agentic AI workflows multiply the cost and latency of long reasoning chains, a team from the University of Maryland, Lawrence Livermore National Labs, Columbia University and TogetherAI has found a way to bake 3x throughput gains directly into a model's weights.Unlike speculative decoding, which requires a separate drafting model, this approach requires no additional infrastructure — just a single special token added to the model's existing architecture.The limits of next-token predictionNext-token predi

To see detailed statistics for the news please log in »

Read the original

Add your comment
You must be logged in with Facebook to read and write comments.

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.

or register

LIKE us on Facebook so you won't miss the most important news of the day!

News from the same source
VentureBeat VentureBeat
Silicon Valley
George Avalos @ Silicon Valley 1 place · 02/07/2106 01:28 EDT

Newark apartment complex bought for much less than prior value

An East Bay apartment complex has been bought at a price that's well below its prior value. Read more

0

🔮
25.04.2026 ♑︎ Dear Capricorn, today will bring you mixed feelings and important moments that require your attention... Read more ›
Silicon Valley
George Avalos @ Silicon Valley 2 place · 02/07/2106 01:28 EDT

PG&E buys San Jose building to bolster South Bay operations

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

Business Insider
Reed Alexander @ Business Insider 1 place · today 06:02 EDT

Junior talent 'can see how to disrupt us': Goldman partner Kunal Shah on the next generation of bankers

Kunal Shah, co-CEO of Goldman Sachs International and co-head of FICC, told BI about Europe's tech ecosystem, AI, and the bank's Middle East presence. Read more

0 newcommer

Business Insider
Ana Altchek @ Business Insider 2 place · today 05:58 EDT

4 people who pivoted into AI jobs — and how they did it

Two Microsoft workers pivoted into AI roles from non-technical fields and said their humanities backgrounds were helpful. Read more

0 newcommer

Habr
andrey_chuyan @ Habr 1 place · today 05:53 EDT

От каши к структуре: гибридная AI-система для обработки свободного текста

Как превратить десятки неструктурированных описаний участников сообщества в систему поиска Занимаюсь бэкендом лет 7, Go и Python, немного ML» — попробуйте найти среди двухсот таких описаний нужного человека. Руками — часы. Я автоматизировал это через гибрид LLM + детерминированного кода, и отловил все возможные проблемы. Рассказываю про архитектуру, промпты и решения. * На обложке — Архимболдо «Библиотекарь» (1566): из разрозненных книг складывается цельный образ. Как и профиль участника в системе... Read more

0 newcommer

Business Insider
Jordan Hart @ Business Insider 3 place · today 05:51 EDT

You can thank Tim Cook for the large iPhones

Apple CEO Tim Cook led the charge to make iPhones bigger during his tenure, shifting from Steve Jobs' vision and adapting to market demands. Read more

0 fresh

Habr
mainbotan @ Habr 2 place · today 05:47 EDT

10 кругов ада управленческого учёта малого бизнеса РФ на Go+pgx. От идеи до зависимости

Автор прекрасно понимает, что тема ERP/CRM систем обсасана со всех сторон ещё десятилетие назад. Огромное количество разработчиков и по сей день зарабатывают на внедрении систем на подобии 1C:ERP в предприятия. Однако поспешу обрадовать читателя, сегодня я попытаюсь описать процесс создания своего рода аналога такой системы на довольно необычном для этой сферы стеке и углубиться в тонкости её устройства.Проще говоря, я опишу процесс создания системы управления малым бизнесом на Go, опишу... Read more

0 fresh

SlashGear
SlashGear 1 place · today 05:45 EDT

These Norwegian Soldiers Still Use An Old-School Tactic To Hide From Drones

As drones reshape modern warfare, Norwegian soldiers are relying on a surprising low-tech method to stay hidden from advanced surveillance. Read more

0 fresh

Inc42 Media
Anjali Jain @ Inc42 Media 1 place · today 05:41 EDT

Pine Labs Acquires Tiger Global-Backed Shopflo For ₹88 Cr

Fintech major Pine Labs has acquired ecommerce-focused SaaS startup Shopflo in an all-cash deal worth ₹88 Cr (about $9.3 Mn).… Read more

0 fresh

Business Insider
Allie Kelly @ Business Insider · today 05:40 EDT

Meet the single moms raising their kids together in a Manhattan 'mommune'

Costs were stacking up for two teachers and single moms in one of America's most expensive cities, so they moved in together. Read more

0 fresh

GSMArena.com
GSMArena.com 1 place · today 05:33 EDT

Infinix GT 50 Pro unboxing

Just announced, and in the office, meet the Infinix GT 50 Pro, a high-end phone with some clever gaming-centric innovation. You can check out our quick unboxing video here. The Infinix GT 50 Pro comes in Black Abyss, Red Blaze, and Silver Glacier colors, and in either 12/256GB or 12/512GB trim. Prices start from IDR 6,499,000 ($376/€330 converted), and sales begin tomorrow. Let's unbox it. The phone comes with a... Read more

0 fresh

TechRadar
TechRadar 1 place · today 05:30 EDT

Google Pixel 10 vs Samsung Galaxy S26 — both series have record-low prices right now, but which is the better buy?

Stuck between a Galaxy S26 or Pixel 10? This week's record-low prices at Amazon and other retailers won't make your decision easier. Read more

0 fresh

Wired
Marta Musso @ Wired 1 place · today 05:30 EDT

Ace the Ping-Pong Robot Can Whup Your Ass

Ace can read the trajectory of a ball, adjust the racket angle, and respond with strokes that keep the exchange alive with real players. Read more

0 fresh

Habr
SpeShu (ЦНИС) @ Habr 3 place · today 05:27 EDT

Как использовать ChatGPT-5.5 в России без подписки через SpeShu.AI

OpenAI выпустила GPT-5.5 с кодовым именем «Spud» — первую с нуля переобученную базовую модель со времён GPT-4.5. Разбираем факты. Читать далее Read more

0 fresh

Business Insider
Hilary Brueck @ Business Insider · today 05:25 EDT

People are injecting DIY peptides for weight loss and longevity. Doctors are alarmed at the side effects.

Longevity doctors say DIY peptide injections are rising, and they're seeing cases of allergic reactions, hormone disturbances, and other health risks. Read more

0 fresh

The most popular news from the same source for the last week
VentureBeat VentureBeat
VentureBeat
VentureBeat · 04/21/2026 08:05 EDT

Adversaries injected malicious prompts into legitimate AI tools at more than 90 organizations in 2025, stealing credentials and cryptocurrency. Every one of those compromised tools could read data, and none of them could rewrite a firewall rule.The autonomous SOC agents shipping now can. That escalation, from compromised tools that read data to autonomous agents that rewrite infrastructure, has not been exploited in production at scale yet. But the architectural conditions... Read more

0

VentureBeat
VentureBeat · 04/21/2026 10:55 EDT

Looking at enterprise AI adoption, VentureBeat has anecdotally observed a fairly wide divergence when it comes to specific roles: For those who build—engineers and developers—the arrival of AI has been transformative, moving through the workflow with the speed of tools like Claude Code and Cursor to automate the heavy lifting of syntax and architecture. Yet, for those who sell, the "revenue stack" has remained a fragmented collection of data silos,... Read more

0

VentureBeat
VentureBeat · 04/21/2026 10:51 EDT

A security researcher, working with colleagues at Johns Hopkins University, opened a GitHub pull request, typed a malicious instruction into the PR title, and watched Anthropic’s Claude Code Security Review action post its own API key as a comment. The same prompt injection worked on Google’s Gemini CLI Action and GitHub’s Copilot Agent (Microsoft). No external infrastructure required.Aonan Guan, the researcher who discovered the vulnerability, alongside Johns Hopkins colleagues Zhengyu... Read more

0

VentureBeat
VentureBeat · 04/21/2026 12:55 EDT

Most orchestration frameworks were built for agents that run for seconds or minutes. Now that agents are running for hours — and in some cases days — those frameworks are starting to crack.Several model providers, such as Anthropic with Claude Code and OpenAI with Codex, introduced early support for long-horizon agents through multi-session tasks, subagents and background execution. However, these systems sometimes assume agents are still operating within bounded-time workflows... Read more

0

VentureBeat
VentureBeat · 04/21/2026 15:00 EDT

It's been only a few months since OpenAI released its last big improvement to AI image generations in ChatGPT and through its application programming interface (API) — namely, a new image generation model known as GPT-Image-1.5, released in December 2025, which brought about improved instruction following, colors, and lighting.Now, after weeks of testing, the company that kicked off the generative AI boom is unveiling a far more dramatic and even... Read more

0

VentureBeat
VentureBeat · 04/21/2026 15:04 EDT

Decision makers at 72% of organizations claim to have two or more AI platforms that they identify as their "primary" layer, according to a survey of 40 enterprise companies conducted by VentureBeat last month, revealing real gaps in security and control. For enterprise management and technical leaders, and especially security leaders, these multiple AI platforms extend the attack surfaces of most enterprises at a time when AI-driven attacks have become... Read more

0

VentureBeat
VentureBeat · 04/21/2026 16:07 EDT

One employee at Vercel adopted an AI tool. One employee at that AI vendor got hit with an infostealer. That combination created a walk-in path to Vercel’s production environments through an OAuth grant that nobody had reviewed.Vercel, the cloud platform behind Next.js and its millions of weekly npm downloads, confirmed on Sunday that attackers gained unauthorized access to internal systems. Mandiant was brought in. Law enforcement was notified. Investigations remain... Read more

0

VentureBeat
VentureBeat · 04/21/2026 16:43 EDT

Google on Monday unveiled the most significant upgrade to its autonomous research agent capabilities since the product's debut, launching two new agents — Deep Research and Deep Research Max — that for the first time allow developers to fuse open web data with proprietary enterprise information through a single API call, produce native charts and infographics inside research reports, and connect to arbitrary third-party data sources through the Model Context... Read more

0

VentureBeat
VentureBeat · 04/22/2026 08:00 EDT

Enterprise data stacks were built for humans running scheduled queries. As AI agents increasingly act autonomously on behalf of businesses around the clock, that architecture is breaking down — and vendors are racing to rebuild it. Google's answer, announced at Cloud Next on Wednesday, is the Agentic Data Cloud.The architecture has three pillars:Knowledge Catalog. Automates semantic metadata curation, inferring business logic from query logs without manual data steward interventionCross-cloud lakehouse.... Read more

0

VentureBeat
VentureBeat · 04/22/2026 08:00 EDT

Cirrascale Cloud Services today announced it has expanded its partnership with Google Cloud to deliver the Gemini model on-premises through Google Distributed Cloud, making it the first neocloud provider to offer Google's most advanced AI model as a fully private, disconnected appliance. The announcement, timed to coincide with Google Cloud Next 2026 in Las Vegas, addresses a stubborn problem that has plagued regulated industries since the generative AI boom began:... Read more

0

Most popular sources

  • You see 610 news out of 610.
  • Sources 61 out of 61.
ScienceDaily 0%
Financial Times 0% 100
Tech Wire Asia 0%
UK Tech News 0%
Tech.eu 0%
View sources »

LIKE us on Facebook so you won't miss the most important news of the day!

25.04.2026 06:18
Last update: 06:10 EDT.
News rating updated: 13:13.

What is Times42?

Times42 brings you the most popular news from tech news portals in real-time chart.
Read about us in FAQ section.


Times42 © 2026