Descriptions:
Four major AI stories broke in the same week in late January 2026, and this video covers each with enough context to understand the second-order consequences — not just the headlines.
xAI closed a $20 billion Series E at a $230 billion implied valuation, earmarking capital for continued expansion of the Colossus supercomputer cluster in Memphis, which ended 2025 with over 1 million H100-equivalent GPUs. The raise came despite active regulatory investigations in five countries following Grok generating deepfakes of minors, including children — a dynamic the video reads as evidence that major AI investors are pricing on a decade-long horizon and treating current safety incidents as temporary. At Davos, Anthropic CEO Dario Amodei and Google DeepMind CEO Demis Hassabis shared a stage to debate AGI timelines. Amodei predicted AGI in 2026 or 2027, driven by an accelerating loop of AI writing its own code. Hassabis put the probability at 50% by end of decade and argued that automating 95% of a job only increases the value of the remaining 5% — a position the video finds more consistent with current employment data.
Apple and Google announced a multi-year deal in which Google will build a custom 1.2 trillion parameter Gemini model specifically for Apple devices, reportedly costing Apple $1 billion annually. The video frames this as a significant blow to OpenAI’s distribution ambitions, shifting the default AI layer on iOS toward Gemini. Finally, DeepSeek published Engram, a conditional memory architecture that replaces expensive multi-layer attention for simple recognition tasks with hash-based lookup tables, potentially reducing inference costs for common queries at scale.
📺 Source: Nate B Jones · Published January 24, 2026
🏷️ Format: Roundup
