AI #185: Preference Cascade

The AI Preference Cascade Reaches Farther

Those who know keep warning us about AI risks.

The Preference Cascade Is Only Getting Started

We are in the midst of a preference cascade about existential risk from AI.

Tools for thought in your OODA loop

The unreasonable effectiveness of getting oriented

Should I State or Should I Show? Aligning AI with Human Preferences

The proliferation of AI agents introduces a new principal-agent problem which stems from human principals' difficulty in articulating preferences. We report results from an experiment focused on mitigating this problem via revealed preferences from choice data. Compared to stated preference from human-written prompts, individuals communicate their preferences more effectively through choices, with as few as two sufficing to match prompts' predictive value. This gap largely reflects subjects' difficulty in translating preferences into prompts and is largest among those exhibiting more Allais-type behavioral patterns. Subjects also misperceive the approaches' relative performance, often choosing the less accurate one.

OODA loop

Make faster decisions with incomplete data.

Kimi K3: The open-weights escalation

The global implications on the AI ecosystem.

Import AI 455: AI systems are about to start building themselves.

The first step towards recursive self improvement

Our minds aren’t equipped to handle AI

AI is junk food for the mind: easy, tempting and ultimately very bad for you.

AI’s Uneven Arrival

o1/o3 points the way to AGI, which is AI that can complete tasks; it may take longer for most companies to adopt them than you might think — just look at digital advertising.

AI #179 Part 1: A Louder Fire Alarm for General Intelligence

What a week.

AI Sycophancy and Decisions

We examine whether sycophantic AI advice distorts decisions. Our experiment involves 1,500 participants in 30 decision environments spanning core domains in eco

Thinking about the future of AI shouldn’t feel like choosing options from a buffet

The world doesn’t react to the vibes we want to give off

Angels-on-the-Shoulder: 5 AI Tools for Better Decisions

Five “angels-on-the-shoulder” AI designs that help people make better decisions.

AI FOR EPISTEMICS & COORDINATION

Civilization and technology have radically improved the human condition. Nonetheless, the world sometimes goes in directions which essentially nobody would prefer — e.g., nuclear arms races, unexpected financial crashes, predatory marketing, or ubiquitous political misinformation.