✦FireflyAI news, ranked by what matters
✦ Night SkyListLatest
✦ TopLatest
Model ReleaseFundingRegulationResearchInfrastructureProductOpinion
5
ProductTechCrunch AI·2d ago

X relaunches a rebuilt Android app after year-long effort

X has released a globally available, rebuilt version of its Android application.

Read at techcrunch.com ↗
5
FundingSCMP Tech·10d ago

Chinese state-owned firms set up semiconductor funds amid calls for ‘patient’ capital

Chinese state-backed firms and provincial governments are establishing new investment funds, including a 5 billion yuan partnership, to provide long-term capital for the domestic semiconductor industry.

Read at scmp.com ↗
5
ProductQbitAI 量子位·11d ago

老黄RTX Spark真机现身Bilibili World!CPU和GPU直接焊在一起,笔记本跑120B大模型

Nvidia's new RTX Spark chip, which integrates the CPU and GPU to run 120-billion-parameter large language models locally on laptops, was showcased at Bilibili World.

Read at qbitai.com ↗
5
Product36Kr 36氪·7d ago

亚马逊Leo与南非企业达成协议,预计明年启动卫星互联网服务

Amazon's Project Kuiper satellite internet initiative has partnered with South African provider Herotel to launch connectivity services in the region by 2027.

Read at 36kr.com ↗
5
Product36Kr 36氪·8d ago

越过湖畔,AI在阳澄湖半岛「跑进」真实生活

The inaugural Yangcheng Lake Peninsula AI Carnival was held in Suzhou, showcasing real-world testing and deployment of AI applications in tourism, sports, and daily life.

Read at 36kr.com ↗
5
ResearcharXiv cs.AI·9d agoPrimary

Semantic Drift and the Stability of Operator Control in Reasoning-Class Decision Support Systems

A longitudinal study identifies and analyzes semantic context drift in reasoning-focused large language models, highlighting risks to human-machine decision support systems.

Read at arxiv.org ↗
5
ResearcharXiv cs.AI·9d agoPrimary

Adaptive Model Compression (AMC): Saliency-Driven Resource Allocation for Ultra-Low-Power Transformer Inference

The proposed Adaptive Model Compression framework dynamically allocates hardware resources during transformer inference based on token importance to reduce energy and memory usage on edge devices.

Read at arxiv.org ↗
5
ResearcharXiv cs.AI·9d agoPrimary

Can Agentic Trading Systems Pay for Their Own Intelligence?

Researchers introduced TradeLens to evaluate whether the operational costs of LLM-based trading agents are offset by the incremental profits they generate.

Read at arxiv.org ↗
5
ResearcharXiv cs.AI·9d agoPrimary

When Does Depth Survive Composition? Compute--Quality Regimes in Latent World Models

This research investigates how the precision of variable-depth predictions in adaptive-compute world models scales and survives during autoregressive rollouts.

Read at arxiv.org ↗
5
ResearcharXiv cs.AI·9d agoPrimary

Bringing Back Rule Induction to Fluid Intelligence Research? An Initial Validation of the ARC-AGI Benchmark in Humans

Researchers conducted an initial human validation of the ARC-AGI benchmark to explore how fluid intelligence relates to rule induction and working memory.

Read at arxiv.org ↗
5
ProductTechNode·9d ago

China’s electric vehicle exports rise 68.7% in H1 2026

China's customs data shows a significant rise in green technology exports for the first half of 2026, including the export of over 10,000 AI-integrated intelligent bionic robots.

Read at technode.com ↗
5
ResearcharXiv cs.AI·9d agoPrimary

Maximizing Human Efficiency in Large-Scale Robot Post-Training via VLAC-Cut Guided Pipeline

This work introduces a pipeline designed to optimize human labor efficiency during the iterative post-training and adaptation of Vision-Language-Action models for robotics.

Read at arxiv.org ↗
5
ProductPandaily·11d ago

Ant Group Robotics Subsidiary Lingbo Takes a New Approach to Building Robot Brains

Ant Group's robotics subsidiary, Lingbo, is developing an embodied AI intelligence platform that leverages data from Ant's extensive payment ecosystem.

Read at pandaily.com ↗
5
ProductTechNode·9d ago

ByteDance denies plans to enter smart driving business

ByteDance has refuted media reports claiming that the company plans to enter the autonomous driving and driverless logistics sectors through its cloud platform, Volcano Engine.

Read at technode.com ↗
5
Funding36Kr 36氪·7d ago

资金逆势布局,科技主题与宽基ETF同获加仓

Investors are increasingly allocating capital into AI and semiconductor-themed ETFs, viewing current market volatility as a long-term entry opportunity.

Read at 36kr.com ↗
5
OpinionThe Next Platform·7d ago

ShinyHunters: How to close the Credential Rotation Gap

An analysis explores the technical challenges and failure points associated with maintaining secure credential rotation across organizations.

Read at intelligence.theregister.com ↗
5
Opinion36Kr 36氪·7d ago

基金二季报科技调仓现分歧,多只绩优基金维持高仓位

Public fund managers are showing divergent strategies regarding technology sector investments despite the ongoing global AI boom.

Read at 36kr.com ↗
5
Product36Kr 36氪·11d ago

机构:HBM4价格今年下半年或涨至4-5美元/千比特

High-bandwidth memory HBM4 prices are projected to rise significantly in the second half of 2026 due to surging AI demand and manufacturing complexities.

Read at 36kr.com ↗
5
Regulation36Kr 36氪·10d ago

科技行业裁员浪潮加剧,多数美国劳动者支持设立人工智能财富基金

A recent survey reveals that a majority of Americans support legislation to mandate that major AI companies contribute half of their shares to a public sovereign wealth fund to distribute AI dividends.

Read at 36kr.com ↗
5
ProductTechCrunch AI·6d ago

Roblox launches an AI-powered game-creation feature in its mobile app

Roblox has integrated a text-to-game generation feature into its mobile application to simplify the game creation process.

Read at techcrunch.com ↗
5
InfrastructureData Center Dynamics·2d ago

Telecom Italia board approves Poste Italiane takeover bid

Telecom Italia's board has approved a takeover bid that will return the company to state ownership.

Read at datacenterdynamics.com ↗
5
ResearcharXiv cs.AI·2d agoPrimary

RealDESED: A Real-World Domestic Sound Event Detection Benchmark

Researchers present RealDESED, a real-world audio benchmark dataset collected from homes to improve domestic sound event detection models.

Read at arxiv.org ↗
5
ResearcharXiv cs.AI·2d agoPrimary

Spatiotemporal Facial Action Unit Detection using Twin Cycle Autoencoders for Driver Monitoring

A twin cycle autoencoder approach is proposed for spatiotemporal facial action unit detection to enhance driver monitoring systems under challenging conditions.

Read at arxiv.org ↗
5
ResearcharXiv cs.AI·8d agoPrimary

MaxSAT-Based Feedback for Guiding Vision-Language Models in Sudoku

Researchers propose a neuro-symbolic approach that uses formal constraint reasoning to improve the logical consistency of vision-language models on Sudoku puzzles.

Read at arxiv.org ↗
← NewerPage 121Older →
Firefly aggregates headlines and links to original sources. All content belongs to its respective publishers.