The field notes
The Daily AI Drop - Evening Edition, October 11, 2026
Two afternoon AI videos: Fahd Mirza tests Microsoft's new Decision-1 decision-scoring model — fast, cheap structured judgments now rivaling TypeSafe's Jev — and Nate B. Jones argues that taste, agency, focus, resourcefulness, and follow-through are the five human skills getting more valuable as AI gets smarter.
Two afternoon videos, one theme: decisions are getting cheaper, judgment is getting dearer.
Microsoft Decision-1: Model for Fast Decision-Making — Fahd Mirza
Fahd tests Microsoft's new Decision-1 model, announced October 9, a purpose-built "decision model" that takes a situation plus a fixed set of options and returns a calibrated probability for each in a single pass — yes/no, multiple choice, ratings, and rubric-based grading of AI responses or agent actions, all through a structured API call. Microsoft post-trained Qwen3.5-9B for single-pass scoring and reports 83.5% accuracy across 36 benchmarks (~150k questions kept blind from training), with 85ms median latency — 4.5x faster than runner-up Quyet-1.0-Large and ~35x faster than GPT-6 Sol. Why it matters: this is the same "decision models" category Fahd covered two days ago with PG-Jev (TypeSafe's Jev), and Microsoft is now a direct rival — Decision-1 is available in Microsoft Foundry and headed to OpenRouter, priced around $0.042 per million input tokens with no output charge. Xbox Research already used it to label 10,000+ pieces of player feedback 14x faster and 200x cheaper than GPT-6 Sol. For agent builders, routing, verification, and safety screening just got millisecond-cheap.
The AI Paradox: 5 Human Skills That Are More Valuable Than Ever — Nate B. Jones
Nate argues that as AI gets smarter, five human skills are getting more valuable, not less: taste (acting on a problem you actually care about), agency (closing the gap between what you say you'll do and what you ship), focus (depth when anyone can spin up ten prototypes by lunch), resourcefulness (flexibility on unfamiliar projects), and follow-through (staying with a problem until it actually helps someone). Why it matters: it's the human mirror of the Decision-1 story — when the mechanical part of deciding gets commoditized, judgment about what matters and the discipline to finish become the bottleneck. He frames it as a bet on 2027, including advice to revisit projects you shelved as too much work before AI, and warns that a flood of cheap prototypes rewards finishers and punishes collectors.