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

đź”®
04.04.2026 ♏︎ Dear Scorpio, today interesting opportunities and challenges will open up before you. Try to listen... 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

Silicon Canals
Christian Kelly @ Silicon Canals 3 place · today 22:46 EDT

Psychology says the reason older people stop caring isn’t emotional withdrawal – it’s that they’ve finally learned to distinguish between what actually matters and what they were only caring about out of social obligation

Research reveals that when older adults stop attending every social event or having opinions on every topic, they're not becoming antisocial—they're demonstrating a sophisticated psychological skill younger people desperately need but rarely develop until faced with their own mortality. Read more ›

0 fresh

SlashGear
SlashGear 1 place · today 22:45 EDT

How Many AA Batteries Does It Take To Power A PC? This YouTuber Found The Answer

As it turns out, it is possible to run a desktop PC on AA batteries - although it is, predictably, wildly impractical. Here's how many you'd need. Read more ›

0 fresh

Silicon Canals
Christian Kelly @ Silicon Canals · today 22:22 EDT

The hardest part of growing up lower middle class wasn’t the lack of money. It was learning to want things quietly, because visible desire in a household running on tight margins felt like an accusation against the people who were already giving everything they had.

Growing up lower middle class taught a specific emotional skill: how to suppress desire so that the people already stretched thin wouldn't feel the weight of what they couldn't provide. That skill persists long after the economic constraints disappear. Read more ›

0 fresh

Mashable
Mashable 1 place · today 22:00 EDT

NYT Strands hints, answers for April 5, 2026

The NYT Strands hints and answers you need to make the most of your puzzling experience. Read more ›

0 newcommer

Mashable
Mashable 2 place · today 22:00 EDT

NYT Connections hints today: Clues, answers for April 5, 2026

Connections is a New York Times word game that's all about finding the "common threads between words." How to solve the puzzle. Read more ›

0 newcommer

Mashable
Mashable 3 place · today 22:00 EDT

Wordle today: Answer, hints for April 5, 2026

Here's the answer for "Wordle" #1751 on April 5 as well as a few hints, tips, and clues to help you solve it yourself. Read more ›

0 newcommer

Silicon Canals
Sarah Mitchell @ Silicon Canals · today 21:52 EDT

Most people don’t realize that the dishonest people in their lives rarely lie about facts — they lie about their intentions, and that specific distinction is why you keep feeling confused rather than simply hurt

When someone leaves you feeling unsettled after a conversation where every fact they shared was verifiably true, you're not going crazy—you're experiencing the most sophisticated form of deception that good people rarely see coming. Read more ›

0 fresh

SlashGear
SlashGear 2 place · today 21:45 EDT

This New $20 Harbor Freight Product Can Take The Mess Out Of DIY Oil Changes

Changing the oil in your car result in spillages and stains across your garage floor and clothes, but there are products that can make the job easier. Read more ›

0 fresh

Slashdot
EditorDavid @ Slashdot 1 place · today 21:34 EDT

Microsoft Pulls Then Re-Issues Windows 11 Preview Update.  Also Begins Force-Updating Windows 11

Nine days ago Microsoft released a non-security "preview" update for Windows 11 — not mandatory for the average Windows user, notes ZDNet, "but rather as optional, more for IT admins and power users who want to test them." TechRepublic adds that the update "was to bring 'production-ready improvements' and generally ensure system stability by optimizing different Windows services." So it's ironic that some (but not all) users reported instead that... Read more ›

0 fresh

SlashGear
SlashGear 3 place · today 21:30 EDT

4 Reasons Why Some Drivers Regret Buying A Tesla

Tesla looks great on paper, but living with one can be different. Some drivers share what didn't quite live up to the promise. Read more ›

0 fresh

GSMArena.com
GSMArena.com 1 place · today 21:01 EDT

Weekly poll results: the Galaxy A57 is interesting but pricey, the Galaxy A37 gets shown the door

There may be trouble brewing on the horizon for the Galaxy A57 and A37 – the results from last week’s poll are in and the consensus is that the two phones are overpriced. And that there is some serious competition in the mid-range market. Luckily, Galaxy A phones don’t stay at MSRP for long, so the first issue will correct itself soon enough. If it hasn’t already – we’re seeing... Read more ›

0 fresh

Silicon Canals
Sarah Mitchell @ Silicon Canals · today 20:46 EDT

Psychology says people who reply to messages within seconds aren’t just efficient – they’ve built their sense of safety around being reachable, because somewhere in their past, being slow to respond had consequences

That instant urge to reply — the one that makes your fingers fly across the keyboard before you've even finished reading the message — might be your nervous system protecting you from a consequence that stopped being real years ago. Read more ›

0 fresh

SlashGear
SlashGear · today 20:45 EDT

TikToker's Pep Boys Visit Shows Why You Should Always Double-Check Your Estimate

When an expected car problem leads to a costly estimate, one driver decides to dig deeper - and what he finds turns into a lesson worth sharing. Read more ›

0 fresh

CNET
Katelyn Chedraoui @ CNET 1 place · today 20:35 EDT

NASA's Artemis II Astronauts Are More Than Halfway to the Moon: Day 4 Live Updates

The four-person Artemis II crew passed the halfway point to the moon late yesterday. Here's everything you need to know about the historic mission entering its fourth day. Read more ›

0 fresh

The most popular news from the same source for the last week
VentureBeat VentureBeat
VentureBeat
VentureBeat 1 place · 03/29/2026 12:00 EDT

Last week, one of our product managers (PMs) built and shipped a feature. Not spec'd it. Not filed a ticket for it. Built it, tested it, and shipped it to production. In a day.A few days earlier, our designer noticed that the visual appearance of our IDE plugins had drifted from the design system. In the old world, that meant screenshots, a JIRA ticket, a conversation to explain the intent,... Read more ›

0

VentureBeat
VentureBeat · 03/30/2026 14:00 EDT

Enterprises building voice-enabled workflows have had limited options for production-grade transcription: closed APIs with data residency risks, or open models that trade accuracy for deployability. Cohere's new open-weight ASR model, Transcribe, is built to compete on all four key differentiators — contextual accuracy, latency, control and cost.Cohere says that Transcribe outperforms current leaders on accuracy — and unlike closed APIs, it can run on an organization's own infrastructure.Cohere, which can... Read more ›

0

VentureBeat
VentureBeat · 03/30/2026 15:30 EDT

“You can deceive, manipulate, and lie. That’s an inherent property of language. It’s a feature, not a flaw,” CrowdStrike CTO Elia Zaitsev told VentureBeat in an exclusive interview at RSA Conference 2026. If deception is baked into language itself, every vendor trying to secure AI agents by analyzing their intent is chasing a problem that cannot be conclusively solved. Zaitsev is betting on context instead. CrowdStrike’s Falcon sensor walks the... Read more ›

0

VentureBeat
VentureBeat · 03/30/2026 18:20 EDT

For three decades, the web has existed in a state of architectural denial. It is a platform originally conceived to share static physics papers, yet it is now tasked with rendering the most complex, interactive, and generative interfaces humanity has ever conceived. At the heart of this tension lies a single, invisible, and prohibitively expensive operation known as "layout reflow." Whenever a developer needs to know the height of a... Read more ›

0

VentureBeat
VentureBeat · 03/31/2026 06:00 EDT

ThinkLabs AI, a startup building artificial intelligence models that simulate the behavior of the electric grid, announced today that it has closed a $28 million Series A financing round led by Energy Impact Partners (EIP), one of the largest energy transition investment firms in the world. Nvidia’s venture capital arm NVentures and Edison International, the parent company of Southern California Edison, also participated in the round.The funding marks a significant... Read more ›

0

VentureBeat
VentureBeat · 03/31/2026 07:00 EDT

Softr, the Berlin-based no-code platform used by more than one million builders and 7,000 organizations including Netflix, Google, and Stripe, today launched what it calls an AI-native platform — a bet that the explosive growth of AI-powered app creation tools has produced a market full of impressive demos but very little production-ready business software.The company's new AI Co-Builder lets non-technical users describe in plain language the software they need, and... Read more ›

0

VentureBeat
VentureBeat · 03/31/2026 10:28 EDT

For the modern enterprise, the digital workspace risks descending into "coordination theater," in which teams spend more time discussing work than executing it. While traditional tools like Slack or Teams excel at rapid communication, they have structurally failed to serve as a reliable foundation for AI agents, such that a Hacker News thread went viral in February 2026 calling upon OpenAI to build its own version of Slack to help... Read more ›

0

VentureBeat
VentureBeat · 03/31/2026 11:00 EDT

Anthropic appears to have accidentally revealed the inner workings of one of its most popular and lucrative AI products, the agentic AI harness Claude Code, to the public.A 59.8 MB JavaScript source map file (.map), intended for internal debugging, was inadvertently included in version 2.1.88 of the @anthropic-ai/claude-code package on the public npm registry pushed live earlier this morning. By 4:23 am ET, Chaofan Shou (@Fried_rice), an intern at Solayer... Read more ›

0

VentureBeat
VentureBeat · 03/31/2026 14:00 EDT

Slack today announced more than 30 new capabilities for Slackbot, its AI-powered personal agent, in what amounts to the most sweeping overhaul of the workplace messaging platform since Salesforce acquired it for $27.7 billion in 2021. The update transforms Slackbot from a simple conversational assistant into a full-spectrum enterprise agent that can take meeting notes across any video provider, operate outside the Slack application on users' desktops, execute tasks through... Read more ›

0

VentureBeat
VentureBeat · 03/31/2026 14:15 EDT

“Your AI? It’s my AI now.” The line came from Etay Maor, VP of Threat Intelligence at Cato Networks, in an exclusive interview with VentureBeat at RSAC 2026 — and it describes exactly what happened to a U.K. CEO whose OpenClaw instance ended up for sale on BreachForums. Maor's argument is that the industry handed AI agents the kind of autonomy it would never extend to a human employee, discarding... Read more ›

0

Most popular sources

  • You see 392 news out of 392.
  • Sources 61 out of 61.
ScienceDaily 0%
The Fintech Times 0%
Financial Times 0%
Vox 0%
Inc42 Media 0%
View sources »

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

04.04.2026 23:21
Last update: 23:10 EDT.
News rating updated: 06:10.

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