All Updates

Page 1118 of 1538

March 20, 2026

๐Ÿ’ฐ
้’›ๅช’ไฝ“โ€ข110d ago

Tencent Cloud Hits Profitability in 2025

Tencent Cloud achieved scaled profitability for the first time in 2025, challenging the industry view that AI investments always lead to heavy losses. This positive result is revealed in Tencent's latest earnings report.

#cloud-profitability#ai-investment#earnings-report
๐Ÿ“ฐ
New York Times Technologyโ€ข110d ago

AI-Washing Layoffs & LLM Writing Flaws

Companies are increasingly citing AI adoption as the cause for layoffs, a phenomenon dubbed 'AI-washing,' though underlying factors may be more complex. The piece explores why large language models struggle to produce high-quality writing. It also introduces tokenmaxxing, techniques to optimize token usage.

#ai-washing#layoffs#tokenmaxxing
๐Ÿ“Š
Bloomberg Technologyโ€ข110d ago

Nvidia CEO Struggles to Wow Audiences

Nvidia CEO Jensen Huang is finding it harder to captivate his audience. He doesnโ€™t always feel like heโ€™s getting his message across. This reflects challenges in engaging stakeholders amid high expectations.

#ceo-challenge#presentation#market-hype
๐Ÿ“ก
TechRadar AIโ€ข110d ago

UK SMEs Save 5.2 Hours/Week with AI Amid Struggles

AI is enabling UK SME workers to save an average of 5.2 hours per week. However, smaller businesses, older workers, and certain British regions are struggling with adoption and require additional support. This reveals disparities in AI integration across the SME sector.

#sme-adoption#productivity-gains#uk-regions
๐ŸŒ
Wiredโ€ข110d ago

Tinder Fixes Dating with New Features

Tinder launches more than a dozen new features like camera roll analysis and astrology-based matches. Aimed at luring Gen Z and re-engaging users burned out from dating apps.

#dating-apps#gen-z#matching-ai
๐Ÿฏ
่™Žๅ—…โ€ข110d ago

World Models Alone Fail Embodied AI

Yuanli Lingji CEO Tang Wenbin critiques pure world model paths for embodied AI, advocating VLA unification. They use distributed multi-modal data collection prioritizing quality, quantity, diversity. Launched DM0 model and DOS-W1 robot with Huaqin.

#embodied-ai#data-collection#world-models
๐Ÿ”ฅ
36ๆฐชโ€ข110d ago

WPS 365 Launches Official AI Little Lobster

At the WPS 365 AI collaborative office Shenzhen summit on March 20, Kingsoft Office VP Wang Dong announced the launch of official office AI 'Little Lobster' โ€” Xiao K. It executes complex tasks in docs, chat, meetings, emails, emphasizing security, professionalism, and result accountability.

#ai-agent#office-ai#productivity-tool
๐Ÿ‡จ๐Ÿ‡ณ
cnBeta (Full RSS)โ€ข110d ago

US Data Centers Hit $45B, Surpass Offices

US data centers under construction reached a record $45.1 billion value, up 29% YoY per Kobeissi Letter. This marks the first time their scale surpasses traditional office projects. The growth reflects AI reshaping the economy and permanent shifts in business space use.

#data-center-growth#ai-economy#construction-boom
๐ŸŒ
Wiredโ€ข110d ago

AI Food Apps: Insights and Anxiety

Food-tracking apps leveraging AI and computer vision helped the author meet caloric and nutrition goals effectively. However, they also triggered anxiety. The experience offered unexpected learnings about personal habits.

#computer-vision#health-tracking#consumer-ai
๐Ÿฏ
่™Žๅ—…โ€ข110d ago

AI Dragon Shrimp Hype Shifts to Cutbacks

OpenClaw, dubbed 'dragon shrimp', is an AI agent automating PC tasks, sparking frenzy in China with installs, stock surges for MiniMax/Tencent, and company clones. It fuels job anxieties, university program cuts, and local policies. Trend now sees some 'killing' agents amid costs and fears.

#ai-agents#china-ai-hype#automation-tools
๐Ÿ”ฅ
36ๆฐชโ€ข110d ago

Super Magji 3.0: First Enterprise Open-Source AI Agent

Super Magji 3.0, the first enterprise-grade open-source AI Agent platform positioned as OpenClaw equivalent, officially released today. It integrates full-stack capabilities including deep research, PPT generation, and data analysis, solving enterprise pain points via self-developed sandbox isolation, budget controls, and human-AI collaboration loops. Enables 1-person cost for 100-person execution power, advancing AI Agents to scalable enterprise 'lobster legion'.

#ai-agent#sandbox-security
๐Ÿ 
ITไน‹ๅฎถโ€ข110d ago

Mistral Launches Forge for Enterprise Models

Mistral AI introduced Forge, allowing enterprises to build specialized models incorporating internal knowledge from docs, code, and data. It uses pre-training, fine-tuning, and RL for customization, supporting dense/MoE architectures. Models enable strategic control and continuous RL improvement.

#custom-models#enterprise-ai#moe
๐Ÿ“‹
TestingCatalogโ€ข110d ago

Anthropic Preps Projects for Claude Cowork Desktop

Anthropic is developing a Projects layer for Claude Cowork on Desktop. This enables users to organize work into local folders. It also supports setting scheduled tasks by project.

#desktop#project-management#scheduling
๐Ÿ’ฐ
้’›ๅช’ไฝ“โ€ข110d ago

Pony.ai Teams Up with Tencent

Pony.ai gains major backing from Tencent in a strategic partnership. Tencent's 900 million user traffic aims to propel autonomous driving efforts forward.

#autonomous-driving#partnership#traffic-integration
๐Ÿ‡จ๐Ÿ‡ณ
cnBeta (Full RSS)โ€ข110d ago

Cuban Predicts Humanoid Robots Fail in 10 Years

Billionaire Mark Cuban is optimistic about human-machine coexistence but bearish on humanoid robots. He doubts the current industry hype will last, predicting a 5-10 year product lifecycle before inevitable failure.

#humanoid-robots#robotics-hype#investor-opinion
๐Ÿ“Š
Bloomberg Technologyโ€ข110d ago

Senators Probe Nvidia's $20B Groq Deal

Nvidia's $20 billion licensing deal with AI startup Groq is under investigation by Democratic senators Elizabeth Warren and Richard Blumenthal. They are examining whether the deal violates antitrust laws by bypassing merger review and strengthening Nvidia's dominance in the AI computing market.

#antitrust#regulation#licensing-deal
๐ŸŒ
Wiredโ€ข110d ago

Meta Kills E2EE in Instagram DMs

Meta is discontinuing end-to-end encryption for Instagram DMs due to low user opt-in rates. Experts warn this could be the first major step in a broader rollback of E2EE technologies worldwide.

#privacy#encryption#big-tech
๐Ÿ”—
Wired AIโ€ข110d ago

LinkedIn Invites Then Bans AI Cofounder

LinkedIn invited an AI agent dubbed 'Cofounder' to give a corporate talk but later banned it. This highlights the irony of social platforms pushing AI adoption while barring AI agents from participating. The incident questions platform policies on AI involvement.

#ai-agents#platform-ban#social-policy
๐Ÿ“‹
TestingCatalogโ€ข110d ago

Lovable Launches AI for Docs, Data, Apps

Lovable has launched an AI platform that analyzes data to create documents, images, and videos. It converts spreadsheets and PDFs into functional apps by running code for data tasks. This update enables seamless transformation of files into interactive applications.

#data-to-app#multimodal-ai#no-code
๐Ÿ“‹
TestingCatalogโ€ข110d ago

Google Unifies Gemini Tools in Single API Call

Google has expanded the Gemini API with unified tool support, enabling developers to combine built-in tools and custom functions in one call. Key features include cross-tool context circulation and unique response IDs for sharing outputs.

#function-calling#api-tools#context-sharing
Page 1118 of 1538