Fresh

Research Overview

Every article below comes from original research -- decompiled source code, network traffic analysis, controlled experiments, or large-scale data analysis. No speculation dressed up as insight. If a claim is made, the methodology is documented.

---

AI Search Engine Internals

Article	What It Covers
CC Rank & Hidden Authority Signal	How Common Crawl data shapes LLM training and why WebGraph metrics predict AI citations
RRF & AI Citations	The Reciprocal Rank Fusion formula ChatGPT uses and why topic clusters get a mathematical advantage
Google AI Mode Architecture	The 4-stage pipeline (Prepare, Retrieve, Signal, Serve) and 7-signal ranking stack
Google Discover Architecture	8-stage content pipeline, NAIADES personalization, and 87,498 decompiled classes
ChatGPT Search Configuration	The ret-rr-skysight-v3 reranking model, freshness scoring, and multi-layer filtering
Perplexity's Ranking Patterns	L3 XGBoost reranking, authoritative domain lists, and topic multipliers

Content & Token Analysis

Article	What It Covers
GPT-5 Tokenizer Analysis	Vocabulary evolution from 50k to 200k tokens, multilingual efficiency, and AEO content implications
Token Economics & Log Probabilities	How log probabilities reveal entity alignment, content gaps, and competitive positioning
Temperature Zero Trick	Using deterministic mode to expose LLM default associations and implicit biases

Platform Behavior & Patterns

Article	What It Covers
ChatGPT & Public Suffix List	How the PSL shapes domain authority attribution in ChatGPT search
1,827 ChatGPT Prompt Analysis	What real users actually ask ChatGPT -- behavioral patterns from site:chatgpt.com data
Recency Bias in AI Search	How freshness affects citation selection across all major AI search platforms

---

Reading Order Recommendations

For SEO professionals moving into AEO:

1. RRF & AI Citations (understand the math)
2. CC Rank (understand the data source)
3. Google AI Mode Architecture (understand the pipeline)
4. Token Economics (understand measurement)

For content strategists:

1. GPT-5 Tokenizer Analysis (understand how AI reads your content)
2. RRF & AI Citations (understand why clusters win)
3. 1,827 ChatGPT Prompt Analysis (understand what users actually want)
4. Recency Bias (understand the freshness factor)

For technical SEOs:

1. Google AI Mode Architecture (full pipeline breakdown)
2. ChatGPT Search Configuration (parameter-level detail)
3. Perplexity's Ranking Patterns (alternative platform intelligence)
4. Public Suffix List (domain attribution mechanics)