AI News, 14 November 2026

Saturday 14 November

The Jensen Huang Scandal Has Shocked Even Scandal-Prone Silicon Valley

The revelation that Jensen Huang had a romantic relationship with Alice Huang changed the chatter in Silicon Valley from soaring I.P.O.s to tawdry affairs.

AI-generated summary Nvidia chief Jensen Huang confessed to cheating on his wife with Alice Huang, a woman U.S. officials believe was spying for China. Jensen Huang subsequently resigned from Nvidia by Thursday morning as the scandal raised implications for national security and the global artificial intelligence boom.

Earlier:

SaaStr AI App of the Week: SciFin. The Nutanix and Cohesity Founder Raised $44 Million to Build AI for Sales Teams.

AI-generated summary The founder of Nutanix and Cohesity raised forty four million dollars in seed funding to launch a sales tool called SciFin that features an AI companion named Pixie. SciFin connects tools to build a picture of the revenue business, and Pixie operates on top of it to track deal metrics and process meeting transcripts.

Anthropic can already retrain Mythos with "a fraction of the compute originally needed," says former researcher Coxon

AI-generated summary Jacob Coxon says a slide from an Anthropic IPO roadshow reveals that algorithmic efficiency is improving at an accelerating rate. The former researcher states that Anthropic can already retrain the Mythos model using only a fraction of the compute originally needed.

Earlier:

KubeCon NA 2026 Live: The Identity Crisis of AI Agents and the First Announcements from Salt Lake City

AI-generated summary KubeCon NA 2026 opened in Salt Lake City where discussions focused on tracking AI agent actions and workloads within Kubernetes clusters. Red Hat introduced OpenShift 5.0 with an integrated Model Context Protocol server while Cloudsmith presented an AI Agent Skills Registry.

Earlier:

A third of the tokens in the code you send an LLM are comments. Indentation is 2.5%.

AI-generated summary An analysis of the CPython 3.13 standard library using OpenAI's tiktoken and o200k_base measures that comments and docstrings account for 34.8% of the tokens in source code files. Re-indenting files and deleting blank lines save only 2.5% and 0.1% of tokens respectively, while GPT-5.6, Claude Sonnet 5.5, and Claude Haiku 5.5 process input tokens at stated list prices.

Nscale loses $1.02bn in six months on the road to a NYSE listing

AI-generated summary Nscale Limited filed a registration statement with the US Securities and Exchange Commission on September 18, 2026 for an initial public offering on the New York Stock Exchange under the ticker NSCL. The filing shows that the London-based builder of AI data centers lost $1,020.1 million in the first half of 2026 on revenue of $140.6 million.

Earlier:

Nebius Outlines $40B AI Cloud Spending Plan

Nebius Outlines $40B AI Cloud Spending Plan

AI-generated summary Nebius Group announced a capital expenditure roadmap projecting up to $45 billion in planned infrastructure spending for 2027 to expand its global graphics processing unit clusters and data centre operations. Nebius management indicated that roughly 75% to 85% of the projected capital outlay will be covered by customer prepayments on qualifying long-term contracts and asset-backed credit facilities.

Earlier: