Category: Google
-
Reverse Prompting: Reconstructing Prompts from AI-Generated Text
We fine-tuned Google’s Gemma 3 (270M) to reverse the typical LLM workflow: given an AI-generated response, the model reconstructs the most likely prompt that produced it. We generated 100,000 synthetic prompt-response pairs using Gemini 2.5 Flash, trained for a single epoch on a consumer GPU, and built a Streamlit app that sweeps 24 decoding configurations…
-
Search Grounding is Transient
There is a fundamental misconception about how Google’s AI search and Gemini chatbot process retrieved web content. It is widely understood that these systems use Retrieval-Augmented Generation (RAG) to search the web, pull snippets from pages, and ground their answers in factual data. However, there is a pervasive assumption that once an AI pulls in…
-
What extraction method is Google using to build grounding snippets?
I’ve been reverse-engineering Google’s Gemini grounding pipeline (AI Mode, Gemini Chat…etc) by examining the raw groundingSupports and groundingChunks returned by the API. Specifically, I’m interested in the snippet construction step, the part where, given a query and a retrieved web page, the system selects which sentences to include in the grounding context supplied to the…
-
Implicit Queries in AI Search
Back in 2015 I wrote about Google’s reliance of user behaviours signals for ranking purposes. In that article I already covered their use of implicit signals, but now there’s an update! While investigating Google’s grounding pipeline (the system that feeds web content to Gemini before it generates an answer) I came across the same patent…
-
Sorry Google, I was wrong.
What Happened I run several tools on the Gemini API. One of them is a grounded search analysis tool that works in two stages: Gemini 2.0 Flash does a Google Search grounded query, then Gemini 3 Pro visits each source page using the URL Context tool to classify its content. Through a strange coincidence, two…
-
Google’s Trajectory: 2026 and Beyond
AI is shifting from tool to utility. Agentic AI Becomes the Default Interface 2026 prediction: Expect Google Search to become agentic by default. Not “here are 10 links” – more like “I booked the restaurant, here’s the confirmation.” Operator-style functionality baked into Search and Gemini app. Gemini 4 Likely Late 2026 The pattern is clear:…
-
Google’s Ranking Signals
Popularity Popularity signals are derived from user interactions based on ingested user events. The more the users interact with a document, the stronger the boosts are. These data requirements check the overall readiness of your events to generate the popularity signals. This is regardless of the specific search app that you choose. Predicted CTR model…
-
How big are Google’s grounding chunks?
Note: Highlighted bits of this article indicate the parts used to ground Gemini with article title as prompt. Our prior analysis showed that Google doesn’t use your full page content when grounding its Gemini-powered AI systems. Now we have substantially more data to share, specifically around how much content gets selected and what determines that…
-
Google’s AI Uses Schema?
Article updated thanks to a sharp observation from Lukasz Rogala who makes my claim less certain and putting us back in the “needs more evidence category”. There’s some evidence Google uses structured data to ground Gemini in its AI search. If true this is good news for AI SEO people and vindication for schema advocates…
-
Grounding Snippet Extraction Tool
You can rank #1 and still be invisible to AI search. That’s the uncomfortable truth of the AI Mode era. Google’s AI doesn’t just look at your page, it extracts specific sentences, evaluates them against the query, and decides whether your content deserves to ground its answer. The rest of your carefully crafted copy? Find…