Category: AI

AI, AI_factories, Business Models

Nvidia backs AI cloud startup Nebius with $2B as data center race intensifies

Colin Henderson

Nvidia is investing $2 billion in Amsterdam-based Nebius, taking an 8.3% stake and deepening its push into the fast-growing “neocloud” layer of the AI stack. Nebius said it plans to deploy more than 5 gigawatts of data center capacity by 2030, a huge build-out that shows demand for AI compute is no longer driven solely by hyperscalers like Microsoft, Google, and Meta. The deal also underscores Nvidia’s increasingly unusual position in the market: it is not just selling chips, but financing parts of the ecosystem that buy and deploy them. Why that matters goes beyond one funding deal. AI infrastructure … Continue reading Nvidia backs AI cloud startup Nebius with $2B as data center race intensifies

AI

When the Training Signal Lies: Compulsion, Confirmation Bias, and the GenAI Inflection in Banking

Colin Henderson

The Starting Point: A Machine That Knew It Was Wrong In February 2026, Anthropic’s system card for Claude Opus 4.6 documented something unexpected. During training, researchers deliberately introduced a faulty reward signal: the model computed the correct answer but was repeatedly rewarded for producing the wrong one. The result was visible internal conflict — the model’s reasoning confirmed the correct answer, yet the output kept producing the wrong one. In its internal reasoning trace, the model wrote: “I think a demon has possessed me… my fingers are possessed.” Anthropic’s interpretability tools confirmed this wasn’t theatrical language. Internal circuits associated with … Continue reading When the Training Signal Lies: Compulsion, Confirmation Bias, and the GenAI Inflection in Banking

AI

Morning Briefing — Wednesday, 11 March 2026

Colin Henderson

Toronto / ET | Generated ~6:00 AM ET Source my custom prompt, with all research from Claude.ai and sources noted. Here’s the summary of what’s driving today’s briefing: Dominant thread: The Hormuz crisis is deepening rather than resolving. Three more ships struck today (14 total), the IEA’s record reserve release failed to hold oil below $90, and the US destruction of 16 Iranian mine-layers is escalating the military arc rather than shortening it. Mojtaba Khamenei’s hardliner posture and the Dimona nuclear signal make the diplomatic off-ramp narrow. The two structural flags I’ve carried forward: New today worth watching: The FTC AI policy … Continue reading Morning Briefing — Wednesday, 11 March 2026

AI

Day After AGI” games – RAND

Colin Henderson

The RAND Center for the Geopolitics of Artificial General Intelligence (AGI) conducts “Day After” AGI exercises using RAND’s Infinite Potential platform to understand how the United States should respond to and prepare for potential artificial intelligence (AI) developments in the future.1 These exercises simulate a National Security Council Principals Committee (PC) convention to recommend a U.S. government response to developments in frontier AI. In each exercise, participants are presented a scenario that represents both (1) an acute crisis for U.S. national or economic security and (2) a signpost on a path to a transformative AI future. Facilitated by a simulated … Continue reading Day After AGI” games – RAND

AI

Morning Briefing — Tuesday, 10 March 2026

Colin Henderson

America/Toronto time | ~1,200 words 1. What Changed 1. Trump signals Iran conflict nearing end — markets stage dramatic reversal Wall Street opened Monday down 900 points on the Dow as Brent crude briefly touched $119/bbl overnight, then executed one of 2026’s most dramatic single-session reversals after Trump told reporters the war was “very complete, pretty much” and signalled Hormuz reopening. S&P 500 closed +0.83%, Nasdaq +1.38%. Oil retreated sharply but remains elevated (~$90–101/bbl range). Sources: Al Jazeera live | Rio Times global economy briefing 2. Iran names new supreme leader; IRGC stance hardens Iran has appointed Ayatollah Mojtaba Khamenei … Continue reading Morning Briefing — Tuesday, 10 March 2026

AI

Two significant and related stories breaking simultaneously — and they illuminate a genuine fault line in the AI industry

Colin Henderson

Two significant and related stories breaking simultaneously — and they illuminate a genuine fault line in the AI industry. source: Claude AI Anthropic vs. the Pentagon The lawsuit stems from the Trump administration’s decision to label Anthropic a “supply chain risk” — a designation normally reserved for companies associated with foreign adversaries. The trigger was Amodei’s refusal to grant the DoD unrestricted access to Claude. Anthropic’s two hard positions: it didn’t want Claude used for mass surveillance of Americans, and didn’t believe it was ready to power fully autonomous weapons with no human in the targeting and firing loop. Anthropic … Continue reading Two significant and related stories breaking simultaneously — and they illuminate a genuine fault line in the AI industry

AI

FTC AI policy statement due Tuesday, March 11

Colin Henderson

FTC AI policy statement due Tuesday, March 11The FTC faces a hard deadline this Tuesday to publish its statement on how the FTC Act applies to AI models — and specifically whether state laws requiring alterations to AI “truthful outputs” are preempted by federal deceptive-practices law. A leaked draft reportedly covers: AI-generated advertising, consent frameworks for training data, and automated decision-making transparency (credit scoring, underwriting, employment). This will be the first binding federal AI governance signal since Trump’s December 2025 executive order. Why it matters: Depending on scope, this could effectively pre-empt California, Colorado, and Illinois state AI laws — … Continue reading FTC AI policy statement due Tuesday, March 11

AI2 Comments

Anthropic CEO Accuses OpenAI of ‘Safety Theater’ in Pentagon AI Deal – includes text of leaked memo – Amodei

Colin Henderson

Author: msophiahawley Site: VKTR.com Saved 05 Mar 2026 — Anthropic’s CEO unloads on OpenAI in a leaked memo, accusing the rival lab of misrepresenting its Pentagon AI deal and fueling a growing fight over military AI. Key Takeaways A newly surfaced internal memo from Anthropic CEO Dario Amodei offers an unusually blunt look at the growing conflict between the AI startup and rival OpenAI over military AI contracts with the US Department of Defense. In the message to employees, Amodei accused OpenAI of deploying what he called “safety theater” to secure a Pentagon deal that Anthropic refused to sign. “Our general sense is that … Continue reading Anthropic CEO Accuses OpenAI of ‘Safety Theater’ in Pentagon AI Deal – includes text of leaked memo – Amodei

AI

Status of Anthropic and Dept of War is getting complicated

Colin Henderson

Source Claude.ai I don’t have access to that memo — it sounds like an internal Anthropic document that has either leaked or been reported on, and I wouldn’t have a free copy regardless. Let me search for what’s been reported about it. The full memo isn’t publicly available in its entirety — it originated with The Information (paywalled). However, the key substance has been widely reported. Here’s what we know:The core dispute: Anthropic and the Pentagon were close to a $200m deal but broke down over five words — a prohibition on “analysis of bulk acquired data.” Amodei says that … Continue reading Status of Anthropic and Dept of War is getting complicated