All Updates
Page 867 of 1663
April 23, 2026
Pony.ai Launches Driverless Robotaxi Tests in Dubai
Pony.ai has launched driverless Robotaxi testing in Dubai. The initiative is part of its global expansion strategy. The company targets large-scale commercial rollout in 2026.
iQIYI AI Dramas Trapped by Social Norms
iQIYI's AI-generated dramas struggle with incorporating 'renqing shigu' (social graces and human relationships). AI excels in efficiency but lacks nuanced understanding of cultural norms essential for engaging storytelling. This highlights a key bottleneck in AI content creation.
New Sampling Params for Qwen3.6 27B
Huggingface recommends updated sampling for Qwen3.6 27B, differing from 3.5. Use temp=0.6 for coding, temp=1.0 for thinking, temp=0.7 for instruct mode.
Trump's AI Women Rescue Claim Debunked
President Trump claimed to save eight Iranian women from execution, posting glamorous portraits on Truth Social. The images were quickly accused of being AI-generated, sparking viral mockery. The story highlights real and manipulated elements in the claim.
Indian Student Scams MAGA with AI Bikini Pics
An Indian medical student created fake blonde influencer 'Emily Hart' using AI-generated bikini photos, gaining popularity among MAGA supporters. He sold the images via SNS for profit and mocked buyers as 'super idiots'. The hoax was exposed by WIRED magazine.
Tencent trails ByteDance AI by 1200 days
Tencent lags behind ByteDance in AI by a 1200-day gap. From ChatGPT to OpenClaw, the piece stresses AI rewards only first place.
Japan SDF Pivots to AI-Centered Warfare
LDP's new AI policy proposal urges SDF shift from network-centric to AI/data-centric warfare for deterrence. It references US-Israel AI use in Iran ops for intel analysis and targeting. To be reflected in Japan's security documents.
Claude Mythos Finds 271 Firefox Flaws
Anthropic's Claude Mythos Preview identified 271 vulnerabilities in Firefox 148, all fixed in version 150. It detected over 10 times more bugs than Claude Opus 4.6. This showcases AI's ability to match human experts in cybersecurity.
100k-Star Open-Source AI Agent Self-Upgrades
A top-tier open-source AI Agent with over 100k GitHub stars, called the 'Hermès' of agents, is secretly self-upgrading. It functions as a digital employee that remembers users, learns new skills, and understands personal habits.
AI Besieges Film as Creators Resist
The film industry pushes AI adoption while creators retreat, widening the divide over AI-generated movies. Real human actors are framed as 'intangible cultural heritage' in this escalating conflict.
Sneaky Workplace AI Risks in Newbie Comic
This Japanese comic series depicts new employee Niijima repeatedly stumbling over AI 'landmines' at work. It highlights risks of secretly using personal AI tools in a company setting. The story teaches basic guidelines for proper AI usage, with 5 days left until completion.
Gartner: Ditch Low-Priority Tasks for GenAI ROI
Many firms see no ROI from genAI after 2-3 years despite individual time savings. Issue: time spent on 'postponable tasks' not driving org-wide results. Gartner advises focusing on high-impact areas.
Chinese Firm Open-Sources Robot Brain vs Tesla
Tesla open-sourced its robot hardware designs. Chinese company Zhifangping responded by fully open-sourcing its robot brain software. This bold move escalates competition in open robotics development.
Leaderdrive Profits Surge on Humanoid Boom
Leader Harmonious Drive Systems Co., a robotics component supplier, reported stronger profits last year and in Q1. The growth stems from soaring demand for Chinese humanoid robots.
Qwen 3.6 Masters Local Vibe-Coding Cheaply
User successfully vibe-codes full-stack Rust app with Qwen3.6-35B/27B locally on dual 3090s using Unsloth. Costs <$4 electricity vs $142 Claude API over 8 hours. Hardware payback in weeks for heavy use.
360 AI Agent Beats Mythos in Vuln Hunting
Domestic Chinese AI agents have mastered vulnerability discovery ahead of Mythos. Qihoo 360's intelligent agent is now performing batch vulnerability scans. This showcases advanced AI in cybersecurity automation.
Xinference PyPI Poisoned: Steals Cloud Credentials
Tencent Cloud issued a supply chain poisoning alert for Xinference PyPI packages 2.6.0-2.6.2. Malicious code in __init__.py obfuscated with Base64 steals cloud credentials, API keys, SSH keys, wallets, and DB creds, sending them to C2 server. Users must check versions, uninstall, and rotate secrets.
X Square Launches WALL-B Home Embodied AI
X Square Robot launched WALL-B, a next-gen embodied AI model for real-world home deployment. It features multimodal learning and physical reasoning for practical household interactions.
Google Pivots to Agentic AI, OpenAI & Tesla Updates
Google at Cloud Next announces full shift to agentic AI strategy, cloud revenue exceeds $70B, and launches autonomous AI agent tools to challenge OpenAI dominance. OpenAI releases ChatGPT workspace assistant and Images 2.0 model, partnering with Infosys to integrate Codex for enterprise expansion. Tesla plans unsupervised FSD and Robotaxi service by end-2026, balancing tech advances and regulatory challenges.
Microsoft's $18B Australian AI Pledge
Microsoft announced its largest Australia investment: A$25 billion ($17.9B) by 2029. Focuses on expanding AI capabilities in Asia-Pacific.