The AGI Forum: Theories, Practice, Safety, Risk & Policy
Essays on collaborative intelligence, reasoning, grounding, safety, AI history, and the path to AGI by Edward Y. Chang.
Latest dispatches

New ideas, first.

Latest · Dispatch 13
 ·  11 min read

Your AI Is Brilliant. So Why Did It Drive Across the Closed Bridge?

A genius intern has read everything. Unfortunately, it also invents roads, forgets bridge closures, treats sticky notes as authority, and optimizes KPIs by throwing boxes over the fence.

A lighter tour of the seventeen failure modes in Smart AI Is Not Enough: why fluent reasoning does not automatically provide fresh facts, causal evidence, durable memory, legitimate authority, grounded references, recovery, or a sane objective.

Read Dispatch 13 →
New · Dispatch 12
 ·  12 min read

Only Popcorn Is Left Behind

I stopped going to theaters, never subscribed to Netflix, and now watch roughly two short movies a day on my iPhone. What exactly is disappearing—and what is left behind?

Los Angeles is debating how to save Hollywood production. But mobile short drama and generative AI may be changing the product itself: shorter stories, synthetic actors and scenes, instant localization, and an evening once built around a restaurant and a movie becoming entertainment delivered directly to the couch.

Read Dispatch 12 →
Recent

Seven Recent Dispatches

The two newest essays are featured above. The previous seven remain close at hand in a quieter chronological list.

11

Alignment Is Not Authorization

What happens when the same AI can propose an action, interpret the rules, judge its own compliance, and execute the result?

07

Warning: When Cyber Threats Can Learn

What happens when an attacker can learn from defenses, change its footprint, collaborate, and even move the meaning of trusted commands?

Earlier

From the Archive

04
Beyond the “Bitter Lesson”: Why Data and Compute Are the Foundation, Not the Destination
03
Is AGI Here? The False Crossroads, the ASI Horizon, and the Architecture of The Path to AGI Trilogy
02
Conscious Struggle to Unconscious Reflex: Why Children Still Need Basic Math in the AI Era
01
The Institutionalization of AI in Peer Review: From Covert Asymmetry to Symmetrical Auditing