Running 70B models on Apple Silicon vs cloud APIs and GPU rentals: a cost analysis for March 2026.

The companies building frontier AI models are telling you who they are. In its most recent IRS filing, OpenAI quietly removed the word "safely" from its mission statement[^7] — capping two years of dissolving its own safety teams, from the Superalignment team in May 2024 to the Mission Alignment team in February 2026, while researchers who departed warned publicly that "safety culture and processes have taken a backseat to shiny products."[^8] That same month, Anthropic — widely considered the most safety-focused of the frontier labs — published a statement from CEO Dario Amodei clarifying that the company does not oppose autonomous weapons in principle: "Even fully autonomous weapons," he wrote, "may prove critical for our national defense." The objection is not moral; it is technical. Frontier AI, in Anthropic's view, is simply not reliable enough yet.[^9]

Here it is:

{
  description = "nix-darwin system flake";

  inputs = {
    nixpkgs.url = "github:NixOS/nixpkgs/nixpkgs-25.11-darwin";
    claude-code.url = "github:sadjow/claude-code-nix";
    nix-darwin = {

Source: Redwood Center for Theoretical Neuroscience Seminar
Author: Ben Dongsung Huh (IBM Research)
Venue: Redwood Center for Theoretical Neuroscience, UC Berkeley
Date: February 11, 2026
Length: 1:15:07

Source: 20 People. $100M Revenue. The 5 Operations Behind Every Tiny Team Beating a Giant One.
Author: Nate B Jones (AI News & Strategy Daily)
Date: March 1, 2026
Length: 28:44

Systems-Theoretic Process Analysis (STPA) is a hazard analysis method developed by Nancy Leveson at MIT. Unlike traditional methods that decompose systems into failure-prone components, STPA models systems as control structures and asks: under what conditions do control actions become unsafe?

Traditional safety analysis (fault trees, FMEA) assumes accidents result from component failures in a chain of events. STPA recognizes that accidents can occur without component failure — through unsafe interactions between components that individually function correctly. This matters for AI systems, where the LLM may behave exactly as designed yet produce unsafe outcomes.

Date: January 15, 2026
Attendees: Sarah Chen, Marcus Rivera, Priya Patel, David Kim

• Q1 target: $4.2M (up 18% YoY)
• Enterprise pipeline: 12 qualified opportunities
• Self-serve ARR growth: 22% projected

Robert Moore and Douglas Gillette

King, Warrior, Magician, Lover (1990) argues that four Jungian archetypes — King (ordering), Warrior (discipline), Magician (knowledge), Lover (passion) — constitute the deep structure of mature masculinity, and that the modern crisis of manhood stems from the collapse of ritual initiation processes that once helped boys access these energies. The book's bipolar shadow model (every archetype fails in two directions: grandiose inflation or passive collapse) is a genuinely useful clinical tool, its diagnosis of arrested male development was prescient, and its practical recommendations are grounded in real therapeutic technique. However, the framework is undermined by consistently unreliable historical illustrations (Cortes, Chernobyl, Hitler, Vietnam) that psychologize systemic events, unsupported empirical claims presented as established science, an unjusti

ISO 9001:2015 helps organizations deliver products and services that satisfy customers and meet regulations. The standard demands consistent quality and continual improvement.

1. Customer Focus — Understand what customers need. Exceed their expectations.
2. Leadership — Leaders set direction and create conditions where people can contribute.