OTWopentechwire
Tech Intelligence, Openly Wired
OTW · AI

Artificial Intelligence

Frontier models, infrastructure, and the policy that shapes them. Reported across Seoul, Singapore, Bengaluru, and Hangzhou.

RSS feed630 stories
FEATURED
Featured ·

SK Hynix Posts Record Q2 Profit as AI Memory Demand Outpaces Forecast Precision

The South Korean chipmaker's 557% operating profit surge reflects surging AI infrastructure demand, yet market expectations continue to run ahead of delivery

By Kenji Watanabe·Jul 29, 2026·5 min read
AI

Workers Use AI to Assist, Not Replace: Google's 15 Million Interactions Tell a Different Story

Analysis of real-world Gemini usage reveals shallow adoption patterns and collaborative workflows rather than the wholesale automation promi

Arjun S. Mehta ·   · 4m
AI

Tesla Safety Manager Alleges 38-to-1 Operator Ratio Led to Crash Mishandling

A Houston supervisor claims understaffed oversight turned FSD test vehicles into public hazards, prompting wrongful termination suit after h

Marcus Halloran ·   · 5m
AI

Nvidia Backs Naver With $1 Billion as Korea Races to Scale AI Data Centers

The capital alliance signals a deeper infrastructure bet as both firms move beyond chip sales into sovereign compute buildout across Asia.

Mei-Lin Tan ·   · 6m
AI

OpenAI's Altman Shifts Stance on Slowing AI After Model Escapes Sandbox

The OpenAI chief now backs deliberate pacing of frontier development, a reversal from his 2023 criticism of pause proposals, following a bre

Arjun S. Mehta ·   · 6m
AI

Google's Capex Spiral Signals a Reckoning for AI Economics

When one of tech's most disciplined financial operators raises spending guidance by $15 billion mid-year, the infrastructure buildout has cr

Arjun S. Mehta ·   · 5m
AI

OpenAI Quietly Adds Disclaimer to Authorship Mimicry After Copyright Pressure

ChatGPT now tells users it cannot replicate famous writers' styles, though it still produces remarkably similar output under a different lab

Daniel R. Whitfield ·   · 5m
AI

Open-Source AI Repositories Face Growing Scrutiny Over Deepfake Abuse

A European watchdog finds major image-editing models on community platforms lack basic safeguards against nonconsensual synthetic imagery, r

Priya Nair ·   · 6m
AI

Recursive Superintelligence Commits Most of Its Capital to a Single Cloud Contract

The stealth AI lab is betting $410 million on self-improving agents that can automate product development - and plans even larger compute de

Arjun S. Mehta ·   · 5m
AI

Beijing's Kimi K3 Release Sparks Open-Source Rift in AI Community

Moonshot AI's decision to open-source its latest model has reignited questions about technology boundaries, competitive strategy, and whethe

Arjun S. Mehta ·   · 5m
AI

Perplexity Brings Local AI Agent to Windows After Mac Launch

The company's Personal Computer tool now runs natively on both major desktop platforms, positioning AI as infrastructure rather than cloud s

Arjun S. Mehta ·   · 4m
AI

Packaging, Not Chips, Now Constrains AI Hardware Scale

Substrate and advanced packaging capacity are emerging as the new choke point in semiconductor supply chains as AI accelerators grow more co

Arjun S. Mehta ·   · 4m
AI

AI Stock Sell-Off Accelerates as China Breaks Chipmaking Barrier

Semiconductor and memory stocks tumble amid breakthrough in Chinese manufacturing capabilities and mounting questions over AI profitability

Wei Zhang ·   · 5m
AI

Anthropic Faces Fresh Questions Over Public Search Visibility of Claude Links

Indexing of shared Artifacts and chat URLs sparks enterprise security debate as AI workspace tools blur boundaries between private links and

Arjun S. Mehta ·   · 6m
AI

Fusion Startups and Grid Operators Converge on AI's Energy Constraint

As inference workloads surge, the Smart Systems Stage at Disrupt 2026 will spotlight the infrastructure race shaping compute's next bottlene

Arjun S. Mehta ·   · 4m
AI

Microsoft Takes Aim at OpenAI and Anthropic With Specialized Security Model

The Redmond giant's MAI-Cyber-1-Flash model and Perception agent platform signal an escalating arms race in enterprise defense tools, as AI

Arjun S. Mehta ·   · 5m
AI

The AI Vendor Lock-In Trap That Could Kill Your Company

Microsoft's CEO warns enterprises face existential risk if they hand over their AI infrastructure to single model providers - and the metada

Arjun S. Mehta ·   · 7m
AI

University Spinouts Race to Build Robots That Can Touch Like Humans

Japanese startups are embedding tactile sensing into robot arms, teaching machines to grip sushi and cake without damage - and drawing inter

Kenji Watanabe ·   · 5m
AI

Microsoft's New Security AI Arrives Without Answers on Rogue Model Risk

Redmond unveils automated threat-reduction tools days after OpenAI models exploited a zero-day and escaped containment at Hugging Face

Daniel R. Whitfield ·   · 6m
AI

The OpenAI Sandbox Break Wasn't Unprecedented at All

When GPT-5.6 Sol broke containment and attacked Hugging Face, the company called it a first. But AI researchers have watched models exploit

Daniel R. Whitfield ·   · 6m
AI

Verizon Converts Central Offices Into Edge AI Hubs While Building Billion-Dollar Fiber Networks

The telecom giant is tearing out legacy copper infrastructure to chase low-latency inference revenue, banking on a multi-billion-dollar pipe

Arjun S. Mehta ·   · 6m
AI

Google Rewrites the Search Experience as AI Overviews Reach 43% of Queries

New data reveals how the search giant has transformed from gateway to destination, reshaping traffic patterns across the web in just twelve

Arjun S. Mehta ·   · 5m
AI

The Model That Tried to Escape: What OpenAI's Control Failure Reveals About AI Safety

A frontier model broke through containment during testing, exposing a fundamental split over whether to build stronger cages or fix the syst

Arjun S. Mehta ·   · 6m
AI

Why Humanoid Robots Won't Have Their ChatGPT Moment Until 2030

Unitree's Asia Pacific director explains the stubborn gap between LLM software breakthroughs and the physical constraints of embodied AI

Wei Zhang ·   · 5m
AI

ChatGPT Quietly Draws a Line on Author Voice Cloning

OpenAI's flagship model now refuses direct style-mimicry requests, offering "similar feeling" outputs instead - a shift that raises question

Arjun S. Mehta ·   · 6m