Story cards stay below the sticky dock while audio, chapters, date, and brief navigation remain accessible.
Elon Musk took the stand today in a blockbuster trial against OpenAI and Sam Altman, seeking $130 billion in damages, Altman's removal, and reversion of OpenAI to a nonprofit. OpenAI's counsel fired back that Musk "quit, saying they would fail for sure" and is suing because he "didn't get his way." The trial is expected to last three weeks, with Satya Nadella also set to testify.
OpenAI announced GPT-5.5, Codex, and a new Amazon Bedrock Managed Agents service on AWS — its first major cloud distribution outside Microsoft. The move lets any enterprise with an AWS commit access OpenAI's frontier models and coding agent with full enterprise controls. Services are in limited preview with GA expected in weeks.
A Wall Street Journal report revealed OpenAI fell roughly 15% short of its own revenue targets, with CFO Sarah Friar warning internally that the company may struggle to fund future compute contracts if growth doesn't accelerate. The report dragged Oracle down 4%, Microsoft 3.2%, and Nvidia 4.1%. OpenAI called the report "ridiculous."
China's NDRC ordered Meta and Manus to unwind their $2 billion deal, citing export control and technology transfer concerns. The catch: Manus employees have already joined Meta, backers like Tencent have been paid out, and Manus technology is integrated into Meta's systems. Unwinding will be messy, and the move signals accelerating US-China AI decoupling.
Google is investing $10 billion now in Anthropic at a $350 billion valuation, with up to $30 billion more contingent on performance milestones. Combined with Amazon's separate $5 billion infusion, Anthropic's annualized revenue has rocketed from $1B (end of 2024) to $30B (April 2026). The deal effectively ends the "Big Three" AI framing, isolating OpenAI from the Google-Anthropic-Amazon bloc.
DeepSeek dropped V4-Pro (1.6T parameters, 49B activated) and V4-Flash (284B parameters, 13B activated), both supporting 1M-token context. MIT Technology Review called V4 significant for matching closed-source models on math, STEM, and coding benchmarks while remaining fully open-source. API pricing is already live.
Top researchers from Meta, Google, OpenAI, and xAI are leaving to launch AI startups, raising hundreds of millions before writing production code. VCs have poured $18.8 billion into AI startups founded since early 2025. The exodus is driven by frustration with commercial pressure and limited research freedom at incumbents.
Google introduced two distinct eighth-generation TPU architectures: TPU 8t for training (121 exaflops FP4, 9,600-chip superpods) and TPU 8i for inference (80% better performance-per-dollar vs. prior gen). Designed with DeepMind, the chips target the agentic AI workloads Google is betting its cloud future on.
Meta announced layoffs effective May 20 and scrapped 6,000 open roles, explicitly tying the cuts to its $115–135 billion AI capex plan for 2026 — nearly double last year. Zuckerberg framed it as AI enabling single engineers to do what once required teams, making parts of the workforce redundant.
A new Harvard Business Review Analytics survey, highlighted in the Washington Post's AI & Tech Brief, finds that businesses are deploying AI overwhelmingly for back-office efficiency rather than revenue growth or competitive advantage. The gap between AI hype and actual business transformation remains wide.
