An Astral Codex Ten essay argues that concern about catastrophic AI risks extends beyond major companies and predates their calls for coordination.
AI Safety Is Not A Corporate Plot
Browsing Category
Myth-Busting
35 posts
What Sun Got Wrong
A detailed analysis of inaccuracies in Sun’s recent coverage, what has been confirmed, and why it matters for readers and the media landscape.
Evidence Of Fraud In An Influential Study About Procrastination
Search and coverage interest is spiking in reported evidence of fraud in an influential procrastination study. The trigger remains unconfirmed.
The Dunning-Kruger Effect May Just Be A Data Artefact (2020)
Research indicates the perceived overconfidence linked to the Dunning-Kruger effect could result from data bias rather than psychological phenomenon.
Is AI Reasoning Right For The Wrong Reasons?
New research questions whether AI reasoning reflects true understanding or if models are just mimicking correct answers for the wrong reasons.
Truth Is Not A Direction: A Tarski Attack On LLM Probes
A recent analysis applies Tarski’s semantic theory to critique the validity of large language model probing methods, questioning their ability to determine ‘truth’.
The Hugging Face Incident
Hugging Face confirms a data breach affecting user data; investigation ongoing. Impact on users and AI community under scrutiny.
What Happens To Schrödinger’s Cat?
An exploration of the current understanding and debates surrounding Schrödinger’s cat, highlighting confirmed facts and ongoing uncertainties in quantum physics.
Human Mathematicians Are Being Outcounterexampled
Advanced AI systems now regularly identify counterexamples to mathematical conjectures, surpassing human mathematicians’ capabilities in this area.
Myth: Water-Based Always Means Nonflammable
Perceptions of water-based products being nonflammable can be misleading, as certain chemicals and conditions may still pose fire risks—discover why.
