ZeroSlop — May 22, 2026
12 stories worth knowing about today — AI breakthroughs, launches, and innovations making a difference.
TechCrunch AI
Hark raises $700M Series A for its secretive ‘universal’ AI interface
Hark raises $700M Series A for its secretive ‘universal’ AI interface
Hark just landed a massive $700M Series A to build a truly universal AI interface that works seamlessly across your existing apps and services—no switching platforms required. The company is launching its first multimodal models this summer, followed by custom hardware designed to make AI assistance feel native to how you already work and live. This could reshape how we actually interact with AI in the real world, moving beyond chatbots toward a genuinely integrated personal AI layer.
NVIDIA Blog
NVIDIA GTC Taipei at COMPUTEX: Live Updates on What’s Next in AI
NVIDIA GTC Taipei at COMPUTEX: Live Updates on What’s Next in AI
NVIDIA’s developer conference is bringing together the brightest minds in tech to showcase cutting-edge breakthroughs across AI factories, infrastructure scaling, and emerging frontiers like agentic and physical AI. From real-time announcements on what’s next in enterprise AI to hands-on previews of tomorrow’s most impactful tools, GTC Taipei is where innovation meets implementation. If you want to know where the AI industry is actually heading—not where the hype says it’s going—this is the event defining it.
arXiv CS.AI
Summary
Researchers have developed VBFDD-Agent, an AI system that uses descriptive text modeling to detect and diagnose lithium-ion battery faults in electric vehicles—moving beyond rigid, scenario-specific methods to handle real-world complexity. By addressing the shortage of open-source battery fault data and creating a unified knowledge framework, this breakthrough enables stronger human-AI collaboration for safer EV maintenance and reliability. This is a game-changer for scaling EV safety as the industry rapidly expands.
arXiv CS.AI
AutoRPA: Efficient GUI Automation through LLM-Driven Code Synthesis from Interactions
AutoRPA: Efficient GUI Automation through LLM-Driven Code Synthesis from Interactions
AutoRPA tackles a real bottleneck in AI automation by automatically converting LLM reasoning into lightweight, reusable code—eliminating the need to invoke expensive language models repeatedly for repetitive GUI tasks. This framework merges the best of both worlds: the adaptability of AI agents with the speed and efficiency of traditional automation, promising to make intelligent task automation practical at scale. By distilling complex reasoning into deployable functions, AutoRPA could unlock automation for everyday business workflows that currently demand manual upkeep.
MarkTechPost
Cohere Releases Command A+: A 218B Sparse MoE Model for Agentic Workflows That Runs on as Few as Two H100 GPUs
Cohere just dropped Command A+, a massive 218B-parameter open-source model that crushes the efficiency game—delivering enterprise-grade multimodal reasoning on just two H100 GPUs. This sparse mixture-of-experts architecture consolidates four prior variants into one polyglot powerhouse supporting 48 languages, making agentic AI workflows finally accessible to teams without unlimited compute budgets. It’s a major step toward democratizing cutting-edge AI capabilities.
MIT Tech Review
The Download: online safety’s future and climate tech’s big pivot
Tech researchers are taking the Trump administration to court to protect their ability to study and combat online hate speech—a crucial fight for preserving the independent research that keeps the internet safer. The case highlights a critical tension: as AI and digital platforms grow more powerful, we need unfettered academic investigation to understand and counter their potential harms. This battle over research freedom could define whether the tech industry innovates responsibly or operates in the shadows.
The Guardian Tech
OpenAI makes breakthrough on 80-year-old maths problem
OpenAI’s AI has cracked an 80-year-old mathematical puzzle that stumped human mathematicians for nearly a century, marking a striking leap in machine reasoning capabilities. The breakthrough on Paul Erdős’s planar unit distance problem shows AI moving beyond pattern recognition into genuine problem-solving territory—tackling the kind of rigorous, creative thinking that defines mathematical innovation. This isn’t just a trophy win; it signals AI systems are becoming genuine partners in exploring fundamental questions that have eluded human effort for generations.
MarkTechPost
Microsoft Releases Fara1.5: A Family of Browser Computer-Use Agents (4B/9B/27B) That Outperform OpenAI Operator and Gemini 2.5 Computer Use on Online-Mind2Web
Microsoft’s new Fara1.5 family of AI agents just set the bar higher for browser automation—with the largest model hitting 72% on Online-Mind2Web and beating OpenAI and Google’s offerings in the process. Available in three sizes (4B to 27B parameters), these agents prove that efficiency and performance aren’t mutually exclusive, backed by Microsoft Research’s innovative FaraGen1.5 synthetic data pipeline. This is the kind of practical AI breakthrough that makes autonomous web interaction faster and smarter for everyone.
arXiv CS.AI
AgentAtlas: Beyond Outcome Leaderboards for LLM Agents
AgentAtlas: Beyond Outcome Leaderboards for LLM Agents
AgentAtlas moves AI agent evaluation beyond simple win-loss metrics, introducing a comprehensive taxonomy that captures how agents fail and why—covering everything from decision-making patterns to safety and robustness across real-world environments like codebases, browsers, and operating systems. This shift from single-number benchmarks to nuanced, multi-dimensional assessment is a game-changer for deploying agents that need to be trustworthy, not just lucky. Finally, we have a framework that lets builders understand agent behavior deeply enough to ship with confidence.
Wired AI
The Gulf’s AI Boom Has an Undersea Cable Problem
The Gulf’s AI Boom Has an Undersea Cable Problem
The explosive growth of AI infrastructure in the Gulf region is exposing a critical vulnerability: undersea cables that weren’t built to handle today’s data demands are becoming critical chokepoints for innovation. As hyperscalers race to deploy AI clusters across the region, they’re forcing Gulf nations to modernize their internet backbone—turning a logistical crisis into an opportunity to build world-class, resilient connectivity that can support the next wave of AI breakthroughs. This infrastructure overhaul could make the Gulf a true tech powerhouse while setting a global blueprint for AI-ready networks.
TechCrunch AI
Trump delays AI security executive order, saying language ‘could have been a blocker’
Trump Delays AI Security Order Over Language Concerns
In a move that keeps the door open for faster AI development, President Trump pumped the brakes on a proposed executive order that would have mandated government security reviews before new AI models hit the market—signaling that the incoming administration wants safeguards that don’t slow innovation. The delay highlights the ongoing tension between responsible AI governance and market velocity, with Trump specifically flagging that the original language risked becoming “a blocker” for the industry. What emerges now is a critical moment to craft security frameworks that actually work: ones that protect without paralyzing the breakthroughs that could reshape everything from healthcare to energy.
TechCrunch AI
The Path, founded by Tony Robbins and Calm alums, hopes to offer safer AI therapy
The Path’s AI Therapy Just Hit a Safety Milestone That Blows Consumer Chatbots Out of the Water
A new startup backed by Tony Robbins and veterans from meditation app Calm has built an AI therapist that scores 95 on mental health safety benchmarks—nearly 50 points ahead of existing consumer bots. By combining deep expertise in behavioral psychology with rigorous safety testing, The Path is showing that AI can tackle one of our most sensitive domains without cutting corners. This could reshape how millions access mental health support, proving that responsible AI and real impact aren’t mutually exclusive.