RE
Research2026-06-19
The 70-Year-Old Test That Breaks Every LLM
A psychology test from 1935 just exposed a fundamental flaw in transformer attention. GPT-4o went from 91% accuracy to 15%. Here's what that actually means.
8 min read
// TAG
#attention
3 articles
AI
AI2026-03-30
Attention Is Memory-Bound, Not Compute-Bound
Everyone knows attention scales quadratically. Almost nobody talks about why that's a memory problem, not a math problem — and why it matters.
8 min read
RE
Research2025-03-01
The Attention Mechanism: Why Transformers Changed Everything
A deep technical dive into the self-attention mechanism that powers every modern LLM — from the original 'Attention Is All You Need' paper to today's multi-head architectures.
1 min read