-
TechnologyAnthropic Launches Cowork, a Claude Code-Like Tool for General Computing
Built in under two weeks using Claude Code itself, Anthropic's Cowork lets non-developers harness autonomous AI file management — no command line required.
-
OpenAIOpenAI's Audio-First Device Strategy
OpenAI is consolidating engineering, product, and research teams around audio to ship a Q1 2026 model with natural interruption handling and simultaneous speech—positioning the company for smart glasses and screenless 'companion' devices before the hardware even ships.
-
TechnologyDisney's One-Year Exclusivity Window with Open AI
Disney sent Google a cease-and-desist the same day it announced its $1B OpenAI deal — a deliberate signal that exclusivity is the price of access. Bob Iger isn't embracing AI; he's building an auction market for cultural authenticity.
-
OpenAIOpenAI CEO declares "code red" as Gemini gains 200 million users in 3 months
Sam Altman has declared a "code red" inside OpenAI as Google's Gemini surged from 450 million to 650 million users in just three months — the clearest signal yet that the AI race is tightening.
-
aiOpenAI Dev Day 2025: Apps SDK, AgentKit, and New Models Transform Developer Platform
OpenAI's Dev Day 2025 was dense: Apps SDK for agent builders, AgentKit for enterprise, GPT-5 Pro priced at $15/$120 per million tokens, Sora 2 for video, and a new real-time mini model. 800 million ChatGPT users and counting.
-
aiOpenAI Launches Sora 2
OpenAI's Sora 2 brings synchronized audio, improved physics simulation, and a social platform for AI-generated video. But is the world ready for a TikTok made entirely of synthetic content?
-
FinanceOpenAI Employees could participate in $6 Billion Liquidity Event at $500 Billion Valuation
OpenAI employees are negotiating a ~$6 billion share sale to SoftBank, Thrive Capital, and Dragoneer at a $500 billion valuation—a significant jump from OpenAI's current $300 billion mark.
-
TechnologyOpenAI Addresses and fixes GPT-5's 'Cold' Personality After User Revolt
Nearly 5,000 users piled onto a Reddit thread titled 'GPT-5 is horrible,' and OpenAI responded—restoring GPT-4o for paid subscribers and committing to a warmer version of GPT-5.
-
CultureMonday Notes
A definitive look at how Silicon Valley's competitive anxieties and deep-pocket bets detonated the AI arms race — alongside Monday's curated cultural briefings.
-
CultureWednesday Notes: AI x Jobs, Sports roundup, Joe Holder on Silence, Netflix password sharing crackdown
Netflix launched its password-sharing crackdown in the U.S. on May 23, 2023, potentially locking out millions of account borrowers. Also: AI's growing impact on employment and Joe Holder on intentional silence.