Blog
Papers
AEO/GEO paper reviews — reading AI-search visibility research and turning it into practice.
Cutting LLM agent costs: delegate decisions to a 4B model
Cutting LLM agent costs: delegating choices to a small decision model saved 81 to 89 percent, but only with thresholds fitted on the agent's own runs.
36 posts
How to build a small LLM: Sheared-LLaMA's 3% compute path
How to build a small LLM, what LLM pruning is, whether small models are enough for AI agents, and the cost of building one, read through Sheared-LLaMA.
The world model that teaches a robot from video: V-JEPA 2
Whether you can teach a robot from video alone, what an AI world model is, and how it differs from a video generator, read through Meta's V-JEPA 2 paper.
Does AI search use your exact question? 70,000 queries show
How AI search turns a question into queries, why uncited pages are not irrelevant, and what retrievers miss, from Perplexity's Q2D-Web benchmark paper.
Is agent memory better than RAG over the raw conversation?
A 2026 UIUC study of why agent memory systems trail RAG over raw history, what LLM summaries lose, and how a path back to the source cuts the loss.
Why RAG misses the evidence an answer needs
A 2026 MBZUAI study of why RAG retrieves only part of the evidence multi-fact questions need and how chunking erases document identity.
Do AI search sources change with the user's location?
A 2026 University of Toronto audit of how user location changes the sources ChatGPT and Gemini cite, and which channel each engine reads location from.
How AI product recommendations vary across tools and APIs
A 2026 study of how decisively AI chatbots recommend products, which sources they display, and whether API audits represent the consumer screen.
How often do AI answers name people, and what rosters miss
A 2026 study of 2,400 grounded AI calls measures how often AI names individual professionals, and how much a roster-based count misses.
MMDocRAG: can AI pick evidence in multimodal document RAG?
MMDocRAG is a NeurIPS 2025 benchmark for multimodal document RAG. Even with perfect retrieval, the best model reaches only 70.2% quote-selection F1.
Video-RAG: retrieving OCR and speech text, not more frames
Video-RAG retrieves OCR, speech, and object text for long video QA instead of more frames. A NeurIPS 2025 paper that lifts Video-MME without training.
All posts in Papers
25 posts in Papers. Click a title to open it.
- DeepSeek V4.1 Flash: smaller KV cache, same post-training
- A reward ceiling for RL alignment, computed before training
- Rewriting content to get cited in AI search, and its risk
- When Google AI Overviews appear and what they cite
- How many measurements to trust AI search visibility
- How AI search picks which page to cite: four gatekeepers
- Why a Page Is Not Cited in AI Search, and How to Repair It
- How Much of a Cited Page Ends Up in an AI Answer?
- AI Answer Brand Mentions Track Exposure More Than Page Fit
- Does AI Search Cite AI-Written Content? About 16% of Sources
- Where Does AI Get Brand Information? 85.7% Is Third-Party
- Why Your Business Doesn't Show Up in AI Recommendations
- Adding a diversity reward to query fan-out: Google's R4T
- Which of two pages do AI answers cite? The 4 gatekeepers
- A survey of 45 GEO papers: what holds up and what doesn't
- Which Brands Do AI Search Engines Mention? A Benchmark
- FeatGEO: Do keywords or structure earn AI citations?
- Does structure alone raise AI citations? The evidence
- SAGEO Arena: edit only the body and you drop out of search
- GEO-16 Framework: What Pages Do AI Engines Cite?
- Content-Centric GEO (CC-GSEO): Agent Rewrites for Citations
- Which Documents Did RAG Use? Shapley Source Attribution
- C-SEO Bench: Does conversational search SEO actually work?
- Can AI recommendations be manipulated?
- The paper that defined GEO: visibility metrics, 9 methods