All Updates
Page 1718 of 1962
March 7, 2026
OpenAI Robotics Head Resigns Over DoD Deal
OpenAI's head of robotics, Caitlin Kalinowski, resigned criticizing the company's rushed partnership with the Department of Defense without adequate guardrails on surveillance and lethal autonomy. OpenAI confirmed the resignation and stated their deal prohibits domestic surveillance and autonomous weapons. This marks a high-profile fallout after Anthropic rejected similar terms.
Govts Can Cheaply Slow AI Training?
Article analyzes if governments can quickly slow AI training via methods like bandwidth limits, cable unplugging, and inference-verification. Current prototypes ineffective against RL due to low communication needs, allowing covert compute use. Aggressive measures like proof-of-work or memory wipes could delay progress by 1+ year.
OpenAI Robotics Lead Resigns Over Surveillance Fears
Caitlin Kalinowski, OpenAI's hardware and robotics engineering lead since November 2024, announced her resignation. She voiced concerns about AI in national security, including surveillance of Americans without judicial oversight and lethal autonomous weapons lacking human authorization. The move follows contract disputes.
Indonesia Bans Social Media Under 16
Indonesia will require high-risk social platforms to delete accounts of users under 16 starting March 28 in a phased rollout. Initial targets include YouTube, TikTok, Facebook, Instagram, Threads, X, Roblox, and Bigo Live. Meta awaits official regulation details amid global trends in Australia, Spain, and Malaysia.
OpenAI Bellevue Office Opens, CTO on Codex
GeekWire podcast visited OpenAI's new Bellevue office on opening day, touring its retro-modern space. They interviewed CTO of Applications Vijaye Raji on Codex, company culture, and Seattle's AI momentum. The episode highlights OpenAI's expansion in the region.
PS6 Launch Unchanged Despite Memory Surge
Rumor states PS6 remains on track for late 2027 to early 2028 release. Memory price hikes unlikely to cause major delays, as delay costs far exceed pricier memory expenses. Contradicts Bloomberg report of postponement to 2028-2029 over 'memory doomsday'.
iPhone 17e Benchmarks Lag on GPU vs iPhone 17
First Geekbench 6 scores for iPhone 17e with A19 chip show multi-core CPU at 9241, matching iPhone 17's 9249 average. GPU performance drags it behind overall. Early leak reveals design trade-offs.
Roblox AI Rephrases Profane Chat Real-Time
Roblox launched an AI-powered chat rephraser that replaces inappropriate language, like profanity, with suitable alternatives in real-time instead of hashes. Users see notifications for changes, and repeat offenders are still penalized. Rolled out for age-similar users across supported languages amid child safety lawsuits.
Personal Brand Trumps AI Tools Like OpenClaw
In the AI era, chasing tools like OpenClaw fuels anxiety without building core competitiveness; personal branding provides lasting differentiation. Article draws parallels to Swiss watch industry's quartz crisis and highlights Paul Graham's 'Brand Age' essay signaling shift to branding over tech.
AI Deanonymizes 66% Forum Users
ETH Zurich and Anthropic researchers demonstrate LLMs can identify real identities of anonymous Hacker News and Reddit users from posts and profiles alone, succeeding in about two-thirds of cases. The technique scales to thousands of candidates and works across sources like LinkedIn and interviews. This erodes traditional online anonymity protections.
Pentagon Recruits Ex-Uber Exec in Anthropic AI Dispute
The Pentagon is enlisting a former Uber executive to address its feud with Anthropic over AI technologies. This involvement signals escalating tensions in military AI procurement and development. Details of the dispute remain limited.
Bilibili Profits via AI-Resilient Differentiation
Bilibili's 2025 annual report shows 13% revenue growth, 8% DAU rise, and 2.59B RMB profit without cost cuts. Community-driven boutique content resists AI replacement, while AI boosts matching efficiency. Ad revenue surges 23%, driven by unique content depth.
OpenClaw ClawCon Meetup Buzzes with Optimism
Hundreds gathered in Manhattan for ClawCon, a vibrant superfan meetup celebrating OpenClaw, the open-source AI assistant launched by Peter Steinberger in November 2025. The event featured lobster headdresses, pink-purple lighting, demos, and sponsor stations. Previously known as Clawdbot and Moltbolt, OpenClaw is gaining rapid popularity in the tech industry.
Circle & Stripe Build Stablecoin for AI Agents
Circle Internet Group and Stripe are developing payment systems for autonomous AI agents transacting in stablecoins. They anticipate millions of daily AI-driven settlements replacing credit cards. This infrastructure targets a nascent AI agent economy.
Google AI Overview Crushes Tech Media Traffic 85%
Google's AI Overview has triggered over 85% drops in search traffic to major tech media sites, with some like Digital Trends losing 97%. Growtika's analysis of Ahrefs data shows monthly Google referrals falling from 112M to under 50M. Google attributes shifts to user preferences for podcasts and forums.
Huawei Zero-Outage Net Hits 99.999% Reliability in 60+ Nets
Huawei's Zero-Outage optical network, enabling break-fiber-no-net, deployed in over 60 global operator networks with 99.999% reliability. Driven by AI app demands for low latency and high bandwidth. Features OTNSpirit AI for 90% risk reduction and 40% O&M cost cuts.
AI Usage Guide for Humanities
Practical principles for humanities workers to integrate AI into writing, research, and production. Stresses traceability, multi-model use, and step-by-step tasks over wishful prompting. Ensures outputs remain controllable and personally attributable.
AI Agents Flip Coding: 80% Natural Language Control
AI coding agents like OpenClaw and Claude Code enable 80% natural language-driven workflows, topping GitHub stars and handling complex 100k-line codebases autonomously. Sequoia declares AGI via long-horizon agents; Block slashes 40% staff citing AI efficiencies. Parallels Gutenberg press commoditizing knowledge replication.
YouTube Expands DM Testing in Europe
YouTube is scaling up its private messaging test to over 30 European countries after starting in Ireland and Poland. Users 18+ can share videos, Shorts, and lives via 7-day invite links in the mobile app. Spam concerns arise, but reporting and blocks are available.
AI Cuts Content Creation to Voice Only
Author built an AI-powered content system reducing work to voicing ideas; AI handles transcription, organization, drafting, while human edits final version. System eliminates waiting for mature ideas and topic research, focusing on authentic insights. Result: consistent output without emotional barriers.