https://dejan.ai/ https://dejan.ai/blog/ https://dejan.ai/blog/brave-ai-search/ 2026-07-13 https://dejan.ai/blog/average-paradox/ 2026-07-12 https://dejan.ai/blog/ecommerce-optimization-engine/ 2026-07-09 https://dejan.ai/blog/coe/ 2026-07-08 https://dejan.ai/blog/grounding-citation-mention/ 2026-07-07 https://dejan.ai/blog/cascading-hallucination/ 2026-07-07 https://dejan.ai/blog/linkjebert/ 2026-07-05 https://dejan.ai/blog/chrome-context-gemini/ 2026-07-05 https://dejan.ai/blog/generative-self-retrieval/ 2026-06-27 https://dejan.ai/blog/primary-bias/ 2026-06-21 https://dejan.ai/blog/grounding-snippets/ 2026-06-21 https://dejan.ai/blog/selection-rate-optimization/ 2026-06-21 https://dejan.ai/blog/relevance-engineering/ 2026-06-21 https://dejan.ai/blog/ai-seo/ 2026-06-21 https://dejan.ai/blog/geo/ 2026-06-21 https://dejan.ai/blog/aeo/ 2026-06-21 https://dejan.ai/blog/aio/ 2026-06-21 https://dejan.ai/blog/the-open-knowledge-format-okf/ 2026-06-21 https://dejan.ai/blog/ai-visibility/ 2026-06-21 https://dejan.ai/blog/teaching-a-model-to-reason-before-it-learns-to-talk/ 2026-06-19 https://dejan.ai/blog/how-search-grounding-biased-an-llm-against-youtube/ 2026-06-18 https://dejan.ai/blog/grounding/ 2026-06-13 https://dejan.ai/blog/emotions-gemma/ 2026-05-17 https://dejan.ai/blog/googles-still-doesnt-see-your-live-page/ 2026-05-07 https://dejan.ai/blog/gemma-4-brand-authority-map/ 2026-04-04 https://dejan.ai/blog/google-shopping-classifier/ 2026-04-03 https://dejan.ai/blog/brands/ 2026-03-28 https://dejan.ai/blog/turboquant/ 2026-03-25 https://dejan.ai/blog/latent-entities/ 2026-03-22 https://dejan.ai/blog/fanout-query-analysis/ 2026-03-20 https://dejan.ai/blog/reverse-prompting/ 2026-03-18 https://dejan.ai/blog/rufus/ 2026-03-15 https://dejan.ai/blog/query-length-vs-volume/ 2026-03-12 https://dejan.ai/blog/search-grounding-is-transient/ 2026-03-06 https://dejan.ai/blog/sro-grounding-snippets/ 2026-03-01 https://dejan.ai/blog/what-extraction-method-is-google-using-to-build-grounding-snippets/ 2026-02-24 https://dejan.ai/blog/implicit-queries-in-ai-search/ 2026-02-24 https://dejan.ai/blog/sorry-google-i-was-wrong/ 2026-02-18 https://dejan.ai/blog/ai-spam/ 2026-02-18 https://dejan.ai/blog/webmcp/ 2026-02-10 https://dejan.ai/blog/bias-and-prejudice-in-ai-search/ 2025-12-30 https://dejan.ai/blog/most-people-dont-read/ 2025-12-30 https://dejan.ai/blog/googles-trajectory-2026-and-beyond/ 2025-12-25 https://dejan.ai/blog/googles-ranking-signals/ 2025-12-24 https://dejan.ai/blog/how-big-are-googles-grounding-chunks/ 2025-12-20 https://dejan.ai/blog/googles-ai-schema/ 2025-12-20 https://dejan.ai/blog/dvl/ 2025-12-18 https://dejan.ai/blog/grounding-snippet-extraction-tool/ 2025-12-15 https://dejan.ai/blog/how-long-are-web-pages/ 2025-12-14 https://dejan.ai/blog/google-ai-search-update-completely-new-grounding-format/ 2025-12-13 https://dejan.ai/blog/ai-mode-content-search-index/ 2025-12-13 https://dejan.ai/blog/how-user-prompts-shape-your-content-visibility-in-ai-search/ 2025-12-13 https://dejan.ai/blog/report-ai-workplace/ 2025-12-10 https://dejan.ai/blog/how-do-people-use-ai-assistants/ 2025-12-05 https://dejan.ai/blog/ricursive/ 2025-12-03 https://dejan.ai/blog/better-clustering/ 2025-11-28 https://dejan.ai/blog/advanced-prompting-techniques/ 2025-11-27 https://dejan.ai/blog/ai-bots/ 2025-11-26 https://dejan.ai/blog/gemini-3-hallucinates-fan-out-queries/ 2025-11-22 https://dejan.ai/blog/ai-seo-deep-dive-tom-critchlow-dan-petrovic/ 2025-11-19 https://dejan.ai/blog/openais-sparse-circuits-breakthrough-and-what-it-means-for-ai-seo/ 2025-11-14 https://dejan.ai/blog/how-gpt-sees-the-web/ 2025-11-14 https://dejan.ai/blog/blockrank/ 2025-11-10 https://dejan.ai/blog/in-ai-seo-10-is-the-new-1/ 2025-11-09 https://dejan.ai/blog/ai-search-filter/ 2025-11-08 https://dejan.ai/blog/browsing-vs-content-fetcher/ 2025-11-08 https://dejan.ai/blog/purchase-intent/ 2025-10-15 https://dejan.ai/blog/claude-system-internals/ 2025-10-09 https://dejan.ai/blog/caps/ 2025-09-30 https://dejan.ai/blog/ai-search-citation-mining/ 2025-09-27 https://dejan.ai/blog/ai-reveal/ 2025-09-27 https://dejan.ai/blog/timesfm-icf/ 2025-09-26 https://dejan.ai/blog/chrome-screen-ai-protos/ 2025-09-23 https://dejan.ai/blog/rexbert/ 2025-09-23 https://dejan.ai/blog/annotated-page-content-apc/ 2025-09-22 https://dejan.ai/blog/deconstructing-domdistiller-how-chromes-reader-mode-algorithm-impacts-technical-seo/ 2025-09-22 https://dejan.ai/blog/llm-is-a-presentation-layer-in-ai-search/ 2025-09-21 https://dejan.ai/blog/gemini-app-tools-a-technical-overview/ 2025-09-14 https://dejan.ai/blog/embeddinggemma/ 2025-09-05 https://dejan.ai/blog/sr/ 2025-09-04 https://dejan.ai/blog/ai-boom/ 2025-09-01 https://dejan.ai/blog/ai-overviews-dialogflow-agent/ 2025-08-31 https://dejan.ai/blog/fan-out-query-search-volume-prediction-using-deep-learning/ 2025-08-30 https://dejan.ai/blog/comment-bots/ 2025-08-28 https://dejan.ai/blog/chrome-help-me-write/ 2025-08-27 https://dejan.ai/blog/tree-walker/ 2025-08-24 https://dejan.ai/blog/does-schema-help-with-ai/ 2025-08-23 https://dejan.ai/blog/your-website-is-about-to-start-talking-are-you-ready-for-this/ 2025-08-21 https://dejan.ai/blog/inside-chromes-semantic-engine-a-technical-analysis-of-history-embeddings/ 2025-08-21 https://dejan.ai/blog/what-does-an-seo-do-in-the-ai-age/ 2025-08-19 https://dejan.ai/blog/understanding-and-control/ 2025-08-17 https://dejan.ai/blog/people-call-them-ai-thats-it/ 2025-08-16 https://dejan.ai/blog/gpt-5-made-seo-irreplaceable/ 2025-08-10 https://dejan.ai/blog/googles-query-fan-out-system-a-technical-overview/ 2025-08-09 https://dejan.ai/blog/gpt-5-system-prompt/ 2025-08-08 https://dejan.ai/blog/journalism-is-dead-say-hello-to-gournalism/ 2025-08-06 https://dejan.ai/blog/human-friendly-content-is-ai-friendly-content/ 2025-07-21 https://dejan.ai/blog/analysis-of-gemini-embed-task-based-dimensionality-deltas/ 2025-07-16 https://dejan.ai/blog/otsu/ 2025-07-09 https://dejan.ai/blog/prompt-engineers-guide-to-gemini-schemas/ 2025-07-02 https://dejan.ai/blog/top-10-most-recent-papers-by-muvera-authors/ 2025-06-30 https://dejan.ai/blog/gemma-embed/ 2025-06-28 https://dejan.ai/blog/training-a-query-fan-out-model/ 2025-06-24 https://dejan.ai/blog/cosine-similarity-or-dot-product/ 2025-06-19 https://dejan.ai/blog/universal-query-classifier/ 2025-06-13 https://dejan.ai/blog/geo-to-seo/ 2025-06-09 https://dejan.ai/blog/vector-embedding-optimization/ 2025-06-06 https://dejan.ai/blog/gemini-toknizer/ 2025-06-05 https://dejan.ai/blog/theres-a-small-army-of-on-device-models-coming-to-chrome/ 2025-06-05 https://dejan.ai/blog/ai-mode-site-search/ 2025-06-04 https://dejan.ai/blog/multi-step-research-agent/ 2025-06-04 https://dejan.ai/blog/query-fan-out-prompt/ 2025-06-04 https://dejan.ai/blog/from-hallucinations-to-clicks/ 2025-06-02 https://dejan.ai/blog/what-is-geo/ 2025-06-02 https://dejan.ai/blog/ai-mode-page-indexing/ 2025-05-30 https://dejan.ai/blog/ai-mode-is-not-live-web/ 2025-05-29 https://dejan.ai/blog/how-ai-mode-selects-snippets/ 2025-05-28 https://dejan.ai/blog/ai-mode-internals/ 2025-05-28 https://dejan.ai/blog/the-future-of-google/ 2025-05-28 https://dejan.ai/blog/gpt-file_search-tool/ 2025-05-27 https://dejan.ai/blog/live-blog-hacking-gemini-embeddings/ 2025-05-24 https://dejan.ai/blog/googles-new-url-context-tool/ 2025-05-21 https://dejan.ai/blog/llm-search-volume/ 2025-05-19 https://dejan.ai/blog/gemini-grounding/ 2025-05-08 https://dejan.ai/blog/google-lens-modes/ 2025-05-08 https://dejan.ai/blog/content-substance-classification/ 2025-04-23 https://dejan.ai/blog/chromes-new-embedding-model/ 2025-04-19 https://dejan.ai/blog/ai-content-detection/ 2025-04-17 https://dejan.ai/blog/generate-then-ground/ 2025-04-17 https://dejan.ai/blog/grounding-classifier/ 2025-04-02 https://dejan.ai/blog/advanced-interpretability-techniques-for-tracing-llm-activations/ 2025-03-31 https://dejan.ai/blog/temperature-parameter-for-controlling-ai-randomness/ 2025-03-30 https://dejan.ai/blog/probability-threshold-for-top-p-nucleus-sampling/ 2025-03-30 https://dejan.ai/blog/how-google-decides-when-to-use-gemini-grounding-for-user-queries/ 2025-03-29 https://dejan.ai/blog/cross-model-circuit-analysis-gemini-vs-gemma-comparison-framework/ 2025-03-29 https://dejan.ai/blog/neural-circuit-analysis-framework-for-brand-mention-optimization/ 2025-03-29 https://dejan.ai/blog/strategic-brand-positioning-in-llms-a-methodological-framework-for-prompt-engineering-and-model-behavior-analysis/ 2025-03-29 https://dejan.ai/blog/alexnet-the-deep-learning-breakthrough-that-reshaped-googles-ai-strategy/ 2025-03-21 https://dejan.ai/blog/the-next-chapter-of-search-get-ready-to-influence-the-robots/ 2025-03-19 https://dejan.ai/blog/hacking-gemini/ 2025-03-14 https://dejan.ai/blog/beyond-rank-tracking-analyzing-brand-perceptions-through-language-model-association-networks/ 2025-02-27 https://dejan.ai/blog/teaching-ai-models-to-be-better-search-engines-a-new-approach-to-training-data/ 2025-02-13 https://dejan.ai/blog/self-supervised-quantized-representation-for-kg-llm-integration/ 2025-02-06 https://dejan.ai/blog/what-does-gemini-think-about-your-brand/ 2025-01-29 https://dejan.ai/blog/googles-privacy-sandbox-navigating-the-cookieless-future/ 2025-01-14 https://dejan.ai/blog/why-deep-learning-works/ 2024-12-26 https://dejan.ai/blog/introducing-veczip-embedding-compression-algorithm/ 2024-12-12 https://dejan.ai/blog/site-engagement-metrics/ 2024-11-29 https://dejan.ai/blog/transitions/ 2024-11-27 https://dejan.ai/blog/random-numbers/ 2024-11-13 https://dejan.ai/blog/chrome-ai-models/ 2024-10-30 https://dejan.ai/blog/attention-is-all-you-need/ 2024-10-13 https://dejan.ai/blog/the-state-of-ai/ 2024-10-13 https://dejan.ai/blog/ilo/ 2024-10-05 https://dejan.ai/blog/resource-efficient-binary-vector-embeddings-with-matryoshka-representation-learning/ 2024-09-05 https://dejan.ai/blog/query-intent-via-retrieval-augmentation-and-model-distillation/ 2024-09-05 https://dejan.ai/blog/search-query-quality-classifier/ 2024-08-31 https://dejan.ai/blog/how-gemini-selects-results/ 2024-08-26 https://dejan.ai/blog/gemini-system-prompt/ 2024-08-26 https://dejan.ai/concepts/ https://dejan.ai/concepts/a2ui/ 2026-07-06 https://dejan.ai/concepts/aeo-answer-engine-optimization/ 2026-06-21 https://dejan.ai/concepts/ai-agent/ 2026-07-06 https://dejan.ai/concepts/ai-assistant/ 2026-07-06 https://dejan.ai/concepts/ai-content-detection/ 2026-07-06 https://dejan.ai/concepts/ai-crawlers/ 2026-07-06 https://dejan.ai/concepts/ai-mode/ 2026-07-06 https://dejan.ai/concepts/ai-overviews/ 2026-07-06 https://dejan.ai/concepts/ai-rank/ 2026-07-06 https://dejan.ai/concepts/ai-seo/ 2026-06-21 https://dejan.ai/concepts/ai-visibility/ 2026-06-21 https://dejan.ai/concepts/aio-ai-optimization/ 2026-06-21 https://dejan.ai/concepts/ap2/ 2026-07-06 https://dejan.ai/concepts/activation-steering/ 2026-07-06 https://dejan.ai/concepts/agentic-harness/ 2026-07-06 https://dejan.ai/concepts/annotated-page-content-apc/ 2026-07-05 https://dejan.ai/concepts/associative-embeddedness/ 2026-07-06 https://dejan.ai/concepts/attention/ 2026-07-06 https://dejan.ai/concepts/binary-classification/ 2026-07-06 https://dejan.ai/concepts/binary-vector-embeddings/ 2026-07-06 https://dejan.ai/concepts/blockrank/ 2026-07-06 https://dejan.ai/concepts/brand-association-network/ 2026-07-06 https://dejan.ai/concepts/brand-authority/ 2026-07-06 https://dejan.ai/concepts/caps/ 2026-07-06 https://dejan.ai/concepts/checkpoint/ 2026-07-06 https://dejan.ai/concepts/chrome-history-embeddings/ 2026-07-06 https://dejan.ai/concepts/citation-mining/ 2026-07-06 https://dejan.ai/concepts/class-imbalance/ 2026-07-06 https://dejan.ai/concepts/classifier/ 2026-07-06 https://dejan.ai/concepts/content-fetcher/ 2026-07-06 https://dejan.ai/concepts/content-node/ 2026-07-05 https://dejan.ai/concepts/content-substance-classification/ 2026-07-06 https://dejan.ai/concepts/context-engineering/ 2026-07-06 https://dejan.ai/concepts/context-window/ 2026-07-06 https://dejan.ai/concepts/cosine-similarity/ 2026-07-06 https://dejan.ai/concepts/document-grounding/ 2026-07-06 https://dejan.ai/concepts/documentchunker/ 2026-07-06 https://dejan.ai/concepts/domdistiller/ 2026-07-06 https://dejan.ai/concepts/dynamic-retrieval/ 2026-07-06 https://dejan.ai/concepts/early-stopping/ 2026-07-06 https://dejan.ai/concepts/embeddinggemma/ 2026-07-06 https://dejan.ai/concepts/epoch/ 2026-07-06 https://dejan.ai/concepts/f1-score/ 2026-07-06 https://dejan.ai/concepts/fine-tuning/ 2026-07-06 https://dejan.ai/concepts/focal-loss/ 2026-07-06 https://dejan.ai/concepts/frequency/ 2026-06-21 https://dejan.ai/concepts/geo-generative-engine-optimization/ 2026-06-21 https://dejan.ai/concepts/generate-then-ground/ 2026-07-06 https://dejan.ai/concepts/generative-self-retrieval/ 2026-06-27 https://dejan.ai/concepts/grounding/ 2026-06-21 https://dejan.ai/concepts/grounding-bias/ 2026-06-21 https://dejan.ai/concepts/grounding-chunk/ 2026-07-06 https://dejan.ai/concepts/grounding-classifier/ 2026-07-06 https://dejan.ai/concepts/grounding-snippet/ 2026-06-21 https://dejan.ai/concepts/hyperparameters/ 2026-07-06 https://dejan.ai/concepts/implicit-queries/ 2026-07-06 https://dejan.ai/concepts/in-context-ranking/ 2026-07-06 https://dejan.ai/concepts/inference/ 2026-07-06 https://dejan.ai/concepts/llm-search-volume-prediction/ 2026-07-06 https://dejan.ai/concepts/language-model/ 2026-07-06 https://dejan.ai/concepts/large-language-model/ 2026-07-06 https://dejan.ai/concepts/latent-entities/ 2026-07-06 https://dejan.ai/concepts/lora/ 2026-07-06 https://dejan.ai/concepts/longformcontextresult/ 2026-07-05 https://dejan.ai/concepts/loss-function/ 2026-07-06 https://dejan.ai/concepts/muvera/ 2026-07-06 https://dejan.ai/concepts/manifold-hypothesis/ 2026-07-06 https://dejan.ai/concepts/matryoshka-representation-learning/ 2026-07-06 https://dejan.ai/concepts/mechanistic-interpretability/ 2026-07-06 https://dejan.ai/concepts/mixture-of-experts/ 2026-07-06 https://dejan.ai/concepts/model-context-protocol/ 2026-07-06 https://dejan.ai/concepts/model-distillation/ 2026-07-06 https://dejan.ai/concepts/model-probing/ 2026-07-06 https://dejan.ai/concepts/multi-label-classification/ 2026-07-06 https://dejan.ai/concepts/multimodal-model/ 2026-07-06 https://dejan.ai/concepts/needle-in-a-haystack/ 2026-07-07 https://dejan.ai/concepts/node-id/ 2026-07-05 https://dejan.ai/concepts/number-of-citations/ 2026-06-21 https://dejan.ai/concepts/number-of-mentions/ 2026-06-21 https://dejan.ai/concepts/on-device-models/ 2026-07-06 https://dejan.ai/concepts/open-knowledge-format/ 2026-06-21 https://dejan.ai/concepts/otsus-thresholding/ 2026-07-06 https://dejan.ai/concepts/overfitting/ 2026-07-06 https://dejan.ai/concepts/pre-training/ 2026-07-06 https://dejan.ai/concepts/precision/ 2026-07-06 https://dejan.ai/concepts/primary-bias/ 2026-06-21 https://dejan.ai/concepts/quadratic-attention-cost/ 2026-07-07 https://dejan.ai/concepts/query-classification/ 2026-07-06 https://dejan.ai/concepts/query-deserves-grounding/ 2026-07-06 https://dejan.ai/concepts/query-fanout/ 2026-06-21 https://dejan.ai/concepts/query-well-formedness/ 2026-07-06 https://dejan.ai/concepts/rank/ 2026-06-21 https://dejan.ai/concepts/recall/ 2026-07-06 https://dejan.ai/concepts/relevance-engineering/ 2026-06-21 https://dejan.ai/concepts/retrieval-augmented-generation/ 2026-07-06 https://dejan.ai/concepts/reverse-prompting/ 2026-07-06 https://dejan.ai/concepts/rufus/ 2026-07-06 https://dejan.ai/concepts/secondary-bias/ 2026-06-21 https://dejan.ai/concepts/selection-rate/ 2026-06-21 https://dejan.ai/concepts/selection-rate-optimization/ 2026-06-21 https://dejan.ai/concepts/semantic-compression/ 2026-07-06 https://dejan.ai/concepts/share-of-citations/ 2026-06-21 https://dejan.ai/concepts/share-of-mentions/ 2026-06-21 https://dejan.ai/concepts/share-of-voice/ 2026-06-21 https://dejan.ai/concepts/site-engagement-score/ 2026-07-06 https://dejan.ai/concepts/site-search-in-ai-mode/ 2026-07-06 https://dejan.ai/concepts/synthetic-training-data/ 2026-07-06 https://dejan.ai/concepts/temperature/ 2026-07-06 https://dejan.ai/concepts/token/ 2026-07-06 https://dejan.ai/concepts/token-probability/ 2026-07-06 https://dejan.ai/concepts/tokenization/ 2026-07-06 https://dejan.ai/concepts/tokenizer/ 2026-07-06 https://dejan.ai/concepts/top-p-sampling/ 2026-07-06 https://dejan.ai/concepts/training-data/ 2026-07-06 https://dejan.ai/concepts/transfer-learning/ 2026-07-06 https://dejan.ai/concepts/tree-walker/ 2026-07-06 https://dejan.ai/concepts/url-context-tool/ 2026-07-06 https://dejan.ai/concepts/vector-embedding/ 2026-07-06 https://dejan.ai/concepts/vector-embedding-optimization/ 2026-07-06 https://dejan.ai/concepts/vector-search/ 2026-07-06 https://dejan.ai/concepts/web-search-grounding/ 2026-07-06 https://dejan.ai/concepts/webmcp/ 2026-07-06 https://dejan.ai/ai-rank-privacy-policy/ 2026-06-23 https://dejan.ai/ai-seo/ 2026-07-06 https://dejan.ai/ai-visibility/ 2026-06-29 https://dejan.ai/auxy-manual/ 2026-06-23 https://dejan.ai/best-ai-seo-agencies/ 2026-07-15 https://dejan.ai/brand-relevance/ 2026-06-30 https://dejan.ai/canvas-demo/ 2026-06-24 https://dejan.ai/careers/ 2026-07-13 https://dejan.ai/dan-petrovic/ 2026-07-01 https://dejan.ai/googles-ai-models/ 2026-06-23 https://dejan.ai/ilo-user-manual/ 2026-06-23 https://dejan.ai/ir-oddy/ 2026-06-24 https://dejan.ai/llms/ 2026-07-13 https://dejan.ai/ml-resources/ 2026-07-13 https://dejan.ai/model-training-evaluation-metrics/ 2026-07-13 https://dejan.ai/models/ 2026-06-30 https://dejan.ai/optimizer/ 2026-06-24 https://dejan.ai/podcast/ 2026-06-23 https://dejan.ai/sro/ 2026-07-01 https://dejan.ai/test/ 2026-06-26 https://dejan.ai/tools/ 2026-07-03 https://dejan.ai/vertex-ai-sku-reference/ 2026-06-23 https://dejan.ai/concepts/brand-authority/listen/ 2026-07-06 https://dejan.ai/blog/how-search-grounding-biased-an-llm-against-youtube/listen/ 2026-06-18 https://dejan.ai/concepts/ai-overviews/listen/ 2026-07-06 https://dejan.ai/concepts/ai-mode/listen/ 2026-07-06 https://dejan.ai/blog/gemma-4-brand-authority-map/listen/ 2026-04-04 https://dejan.ai/blog/google-shopping-classifier/listen/ 2026-04-03 https://dejan.ai/blog/googles-still-doesnt-see-your-live-page/listen/ 2026-05-07 https://dejan.ai/concepts/rank/listen/ 2026-06-21 https://dejan.ai/blog/turboquant/listen/ 2026-03-25 https://dejan.ai/concepts/chrome-history-embeddings/listen/ 2026-07-06 https://dejan.ai/blog/reverse-prompting/listen/ 2026-03-18 https://dejan.ai/blog/rufus/listen/ 2026-03-15 https://dejan.ai/concepts/citation-mining/listen/ 2026-07-06 https://dejan.ai/blog/does-schema-help-with-ai/listen/ 2025-08-23 https://dejan.ai/blog/what-extraction-method-is-google-using-to-build-grounding-snippets/listen/ 2026-02-24 https://dejan.ai/blog/sro-grounding-snippets/listen/ 2026-03-01 https://dejan.ai/concepts/classifier/listen/ 2026-07-06 https://dejan.ai/blog/sorry-google-i-was-wrong/listen/ 2026-02-18 https://dejan.ai/blog/bias-and-prejudice-in-ai-search/listen/ 2025-12-30 https://dejan.ai/blog/webmcp/listen/ 2026-02-10 https://dejan.ai/blog/ai-spam/listen/ 2026-02-18 https://dejan.ai/concepts/epoch/listen/ 2026-07-06 https://dejan.ai/blog/googles-ai-schema/listen/ 2025-12-20 https://dejan.ai/blog/how-big-are-googles-grounding-chunks/listen/ 2025-12-20 https://dejan.ai/blog/grounding-snippet-extraction-tool/listen/ 2025-12-15 https://dejan.ai/concepts/generate-then-ground/listen/ 2026-07-06 https://dejan.ai/blog/google-ai-search-update-completely-new-grounding-format/listen/ 2025-12-13 https://dejan.ai/concepts/ai-rank/listen/ 2026-07-06 https://dejan.ai/concepts/caps/listen/ 2026-07-06 https://dejan.ai/blog/better-clustering/listen/ 2025-11-28 https://dejan.ai/blog/ricursive/listen/ 2025-12-03 https://dejan.ai/blog/how-do-people-use-ai-assistants/listen/ 2025-12-05 https://dejan.ai/concepts/generative-self-retrieval/listen/ 2026-06-27 https://dejan.ai/blog/ai-seo-deep-dive-tom-critchlow-dan-petrovic/listen/ 2025-11-19 https://dejan.ai/blog/openais-sparse-circuits-breakthrough-and-what-it-means-for-ai-seo/listen/ 2025-11-14 https://dejan.ai/blog/gemini-3-hallucinates-fan-out-queries/listen/ 2025-11-22 https://dejan.ai/concepts/geo-generative-engine-optimization/listen/ 2026-06-21 https://dejan.ai/blog/browsing-vs-content-fetcher/listen/ 2025-11-08 https://dejan.ai/blog/purchase-intent/listen/ 2025-10-15 https://dejan.ai/concepts/associative-embeddedness/listen/ 2026-07-06 https://dejan.ai/blog/ai-search-citation-mining/listen/ 2025-09-27 https://dejan.ai/blog/claude-system-internals/listen/ 2025-10-09 https://dejan.ai/concepts/blockrank/listen/ 2026-07-06 https://dejan.ai/blog/ai-reveal/listen/ 2025-09-27 https://dejan.ai/blog/chrome-screen-ai-protos/listen/ 2025-09-23 https://dejan.ai/blog/timesfm-icf/listen/ 2025-09-26 https://dejan.ai/blog/rexbert/listen/ 2025-09-23 https://dejan.ai/blog/deconstructing-domdistiller-how-chromes-reader-mode-algorithm-impacts-technical-seo/listen/ 2025-09-22 https://dejan.ai/blog/embeddinggemma/listen/ 2025-09-05 https://dejan.ai/concepts/grounding-chunk/listen/ 2026-07-06 https://dejan.ai/blog/gemini-app-tools-a-technical-overview/listen/ 2025-09-14 https://dejan.ai/blog/ai-overviews-dialogflow-agent/listen/ 2025-08-31 https://dejan.ai/concepts/content-fetcher/listen/ 2026-07-06 https://dejan.ai/blog/chrome-help-me-write/listen/ 2025-08-27 https://dejan.ai/blog/fan-out-query-search-volume-prediction-using-deep-learning/listen/ 2025-08-30 https://dejan.ai/concepts/hyperparameters/listen/ 2026-07-06 https://dejan.ai/blog/understanding-and-control/listen/ 2025-08-17 https://dejan.ai/blog/people-call-them-ai-thats-it/listen/ 2025-08-16 https://dejan.ai/blog/gpt-5-system-prompt/listen/ 2025-08-08 https://dejan.ai/blog/journalism-is-dead-say-hello-to-gournalism/listen/ 2025-08-06 https://dejan.ai/blog/analysis-of-gemini-embed-task-based-dimensionality-deltas/listen/ 2025-07-16 https://dejan.ai/blog/googles-query-fan-out-system-a-technical-overview/listen/ 2025-08-09 https://dejan.ai/concepts/implicit-queries/listen/ 2026-07-06 https://dejan.ai/blog/prompt-engineers-guide-to-gemini-schemas/listen/ 2025-07-02 https://dejan.ai/blog/top-10-most-recent-papers-by-muvera-authors/listen/ 2025-06-30 https://dejan.ai/blog/gemma-embed/listen/ 2025-06-28 https://dejan.ai/blog/cosine-similarity-or-dot-product/listen/ 2025-06-19 https://dejan.ai/blog/geo-to-seo/listen/ 2025-06-09 https://dejan.ai/blog/training-a-query-fan-out-model/listen/ 2025-06-24 https://dejan.ai/blog/universal-query-classifier/listen/ 2025-06-13 https://dejan.ai/blog/vector-embedding-optimization/listen/ 2025-06-06 https://dejan.ai/blog/theres-a-small-army-of-on-device-models-coming-to-chrome/listen/ 2025-06-05 https://dejan.ai/blog/ai-mode-site-search/listen/ 2025-06-04 https://dejan.ai/blog/query-fan-out-prompt/listen/ 2025-06-04 https://dejan.ai/blog/multi-step-research-agent/listen/ 2025-06-04 https://dejan.ai/blog/from-hallucinations-to-clicks/listen/ 2025-06-02 https://dejan.ai/concepts/document-grounding/listen/ 2026-07-06 https://dejan.ai/blog/ai-mode-internals/listen/ 2025-05-28 https://dejan.ai/blog/ai-mode-page-indexing/listen/ 2025-05-30 https://dejan.ai/blog/what-is-geo/listen/ 2025-06-02 https://dejan.ai/blog/the-future-of-google/listen/ 2025-05-28 https://dejan.ai/blog/gpt-file_search-tool/listen/ 2025-05-27 https://dejan.ai/concepts/inference/listen/ 2026-07-06 https://dejan.ai/blog/googles-new-url-context-tool/listen/ 2025-05-21 https://dejan.ai/blog/google-lens-modes/listen/ 2025-05-08 https://dejan.ai/blog/gemini-grounding/listen/ 2025-05-08 https://dejan.ai/concepts/llm-search-volume-prediction/listen/ 2026-07-06 https://dejan.ai/concepts/longformcontextresult/listen/ 2026-07-05 https://dejan.ai/blog/generate-then-ground/listen/ 2025-04-17 https://dejan.ai/blog/grounding-classifier/listen/ 2025-04-02 https://dejan.ai/blog/how-google-decides-when-to-use-gemini-grounding-for-user-queries/listen/ 2025-03-29 https://dejan.ai/blog/temperature-parameter-for-controlling-ai-randomness/listen/ 2025-03-30 https://dejan.ai/concepts/loss-function/listen/ 2026-07-06 https://dejan.ai/blog/probability-threshold-for-top-p-nucleus-sampling/listen/ 2025-03-30 https://dejan.ai/blog/advanced-interpretability-techniques-for-tracing-llm-activations/listen/ 2025-03-31 https://dejan.ai/blog/alexnet-the-deep-learning-breakthrough-that-reshaped-googles-ai-strategy/listen/ 2025-03-21 https://dejan.ai/concepts/cosine-similarity/listen/ 2026-07-06 https://dejan.ai/concepts/domdistiller/listen/ 2026-07-06 https://dejan.ai/blog/hacking-gemini/listen/ 2025-03-14 https://dejan.ai/blog/teaching-ai-models-to-be-better-search-engines-a-new-approach-to-training-data/listen/ 2025-02-13 https://dejan.ai/concepts/dynamic-retrieval/listen/ 2026-07-06 https://dejan.ai/blog/googles-privacy-sandbox-navigating-the-cookieless-future/listen/ 2025-01-14 https://dejan.ai/blog/introducing-veczip-embedding-compression-algorithm/listen/ 2024-12-12 https://dejan.ai/concepts/content-substance-classification/listen/ 2026-07-06 https://dejan.ai/blog/random-numbers/listen/ 2024-11-13 https://dejan.ai/blog/chrome-ai-models/listen/ 2024-10-30 https://dejan.ai/blog/resource-efficient-binary-vector-embeddings-with-matryoshka-representation-learning/listen/ 2024-09-05 https://dejan.ai/concepts/multimodal-model/listen/ 2026-07-06 https://dejan.ai/blog/gemini-system-prompt/listen/ 2024-08-26 https://dejan.ai/blog/query-intent-via-retrieval-augmentation-and-model-distillation/listen/ 2024-09-05 https://dejan.ai/concepts/on-device-models/listen/ 2026-07-06 https://dejan.ai/concepts/open-knowledge-format/listen/ 2026-06-21 https://dejan.ai/concepts/grounding-classifier/listen/ 2026-07-06 https://dejan.ai/blog/teaching-a-model-to-reason-before-it-learns-to-talk/listen/ 2026-06-19 https://dejan.ai/blog/latent-entities/listen/ 2026-03-22 https://dejan.ai/blog/brands/listen/ 2026-03-28 https://dejan.ai/blog/query-length-vs-volume/listen/ 2026-03-12 https://dejan.ai/concepts/annotated-page-content-apc/listen/ 2026-07-05 https://dejan.ai/concepts/grounding/listen/ 2026-06-21 https://dejan.ai/blog/advanced-prompting-techniques/listen/ 2025-11-27 https://dejan.ai/blog/googles-trajectory-2026-and-beyond/listen/ 2025-12-25 https://dejan.ai/concepts/embeddinggemma/listen/ 2026-07-06 https://dejan.ai/blog/caps/listen/ 2025-09-30 https://dejan.ai/blog/ai-search-filter/listen/ 2025-11-08 https://dejan.ai/blog/ai-boom/listen/ 2025-09-01 https://dejan.ai/blog/how-gpt-sees-the-web/listen/ 2025-11-14 https://dejan.ai/concepts/muvera/listen/ 2026-07-06 https://dejan.ai/concepts/otsus-thresholding/listen/ 2026-07-06 https://dejan.ai/blog/human-friendly-content-is-ai-friendly-content/listen/ 2025-07-21 https://dejan.ai/blog/what-does-an-seo-do-in-the-ai-age/listen/ 2025-08-19 https://dejan.ai/blog/llm-search-volume/listen/ 2025-05-19 https://dejan.ai/blog/your-website-is-about-to-start-talking-are-you-ready-for-this/listen/ 2025-08-21 https://dejan.ai/blog/tree-walker/listen/ 2025-08-24 https://dejan.ai/concepts/webmcp/listen/ 2026-07-06 https://dejan.ai/blog/how-gemini-selects-results/listen/ 2024-08-26 https://dejan.ai/blog/site-engagement-metrics/listen/ 2024-11-29 https://dejan.ai/concepts/matryoshka-representation-learning/listen/ 2026-07-06 https://dejan.ai/concepts/ai-seo/listen/ 2026-06-21 https://dejan.ai/blog/beyond-rank-tracking-analyzing-brand-perceptions-through-language-model-association-networks/listen/ 2025-02-27 https://dejan.ai/blog/gemini-toknizer/listen/ 2025-06-05 https://dejan.ai/blog/content-substance-classification/listen/ 2025-04-23 https://dejan.ai/concepts/reverse-prompting/listen/ 2026-07-06 https://dejan.ai/blog/how-ai-mode-selects-snippets/listen/ 2025-05-28 https://dejan.ai/blog/search-query-quality-classifier/listen/ 2024-08-31 https://dejan.ai/blog/the-next-chapter-of-search-get-ready-to-influence-the-robots/listen/ 2025-03-19 https://dejan.ai/blog/what-does-gemini-think-about-your-brand/listen/ 2025-01-29 https://dejan.ai/blog/generative-self-retrieval/listen/ 2026-06-27 https://dejan.ai/concepts/frequency/listen/ 2026-06-21 https://dejan.ai/concepts/grounding-bias/listen/ 2026-06-21 https://dejan.ai/concepts/content-node/listen/ 2026-07-05 https://dejan.ai/concepts/in-context-ranking/listen/ 2026-07-06 https://dejan.ai/concepts/node-id/listen/ 2026-07-05 https://dejan.ai/concepts/number-of-citations/listen/ 2026-06-21 https://dejan.ai/concepts/number-of-mentions/listen/ 2026-06-21 https://dejan.ai/concepts/grounding-snippet/listen/ 2026-06-21 https://dejan.ai/concepts/query-fanout/listen/ 2026-06-21 https://dejan.ai/concepts/ap2/listen/ 2026-07-06 https://dejan.ai/concepts/secondary-bias/listen/ 2026-06-21 https://dejan.ai/concepts/selection-rate-optimization/listen/ 2026-06-21 https://dejan.ai/concepts/share-of-citations/listen/ 2026-06-21 https://dejan.ai/concepts/share-of-voice/listen/ 2026-06-21 https://dejan.ai/concepts/selection-rate/listen/ 2026-06-21 https://dejan.ai/concepts/latent-entities/listen/ 2026-07-06 https://dejan.ai/concepts/primary-bias/listen/ 2026-06-21 https://dejan.ai/concepts/relevance-engineering/listen/ 2026-06-21 https://dejan.ai/concepts/share-of-mentions/listen/ 2026-06-21 https://dejan.ai/concepts/url-context-tool/listen/ 2026-07-06 https://dejan.ai/concepts/web-search-grounding/listen/ 2026-07-06 https://dejan.ai/blog/linkjebert/listen/ 2026-07-05 https://dejan.ai/concepts/tree-walker/listen/ 2026-07-06 https://dejan.ai/concepts/tokenization/listen/ 2026-07-06 https://dejan.ai/concepts/model-context-protocol/listen/ 2026-07-06 https://dejan.ai/concepts/temperature/listen/ 2026-07-06 https://dejan.ai/concepts/mechanistic-interpretability/listen/ 2026-07-06 https://dejan.ai/concepts/query-deserves-grounding/listen/ 2026-07-06 https://dejan.ai/concepts/vector-embedding/listen/ 2026-07-06 https://dejan.ai/concepts/vector-search/listen/ 2026-07-06 https://dejan.ai/concepts/rufus/listen/ 2026-07-06 https://dejan.ai/concepts/token-probability/listen/ 2026-07-06 https://dejan.ai/concepts/vector-embedding-optimization/listen/ 2026-07-06 https://dejan.ai/concepts/ai-agent/listen/ 2026-07-06 https://dejan.ai/concepts/documentchunker/listen/ 2026-07-06 https://dejan.ai/concepts/ai-crawlers/listen/ 2026-07-06 https://dejan.ai/concepts/semantic-compression/listen/ 2026-07-06 https://dejan.ai/concepts/a2ui/listen/ 2026-07-06 https://dejan.ai/concepts/retrieval-augmented-generation/listen/ 2026-07-06 https://dejan.ai/concepts/site-search-in-ai-mode/listen/ 2026-07-06 https://dejan.ai/concepts/query-classification/listen/ 2026-07-06 https://dejan.ai/concepts/top-p-sampling/listen/ 2026-07-06 https://dejan.ai/concepts/brand-association-network/listen/ 2026-07-06 https://dejan.ai/concepts/training-data/listen/ 2026-07-06 https://dejan.ai/blog/aeo/listen/ 2026-06-21 https://dejan.ai/concepts/attention/listen/ 2026-07-06 https://dejan.ai/concepts/context-engineering/listen/ 2026-07-06 https://dejan.ai/blog/ai-bots/listen/ 2025-11-26 https://dejan.ai/concepts/activation-steering/listen/ 2026-07-06 https://dejan.ai/blog/ai-content-detection/listen/ 2025-04-17 https://dejan.ai/concepts/query-well-formedness/listen/ 2026-07-06 https://dejan.ai/concepts/model-distillation/listen/ 2026-07-06 https://dejan.ai/blog/ai-mode-content-search-index/listen/ 2025-12-13 https://dejan.ai/concepts/model-probing/listen/ 2026-07-06 https://dejan.ai/concepts/manifold-hypothesis/listen/ 2026-07-06 https://dejan.ai/concepts/synthetic-training-data/listen/ 2026-07-06 https://dejan.ai/concepts/recall/listen/ 2026-07-06 https://dejan.ai/blog/ai-mode-is-not-live-web/listen/ 2025-05-29 https://dejan.ai/concepts/class-imbalance/listen/ 2026-07-06 https://dejan.ai/concepts/site-engagement-score/listen/ 2026-07-06 https://dejan.ai/concepts/checkpoint/listen/ 2026-07-06 https://dejan.ai/concepts/f1-score/listen/ 2026-07-06 https://dejan.ai/blog/ai-visibility/listen/ 2026-06-21 https://dejan.ai/concepts/early-stopping/listen/ 2026-07-06 https://dejan.ai/concepts/fine-tuning/listen/ 2026-07-06 https://dejan.ai/blog/annotated-page-content-apc/listen/ 2025-09-22 https://dejan.ai/concepts/focal-loss/listen/ 2026-07-06 https://dejan.ai/concepts/precision/listen/ 2026-07-06 https://dejan.ai/blog/chrome-context-gemini/listen/ 2026-07-05 https://dejan.ai/concepts/multi-label-classification/listen/ 2026-07-06 https://dejan.ai/concepts/pre-training/listen/ 2026-07-06 https://dejan.ai/concepts/lora/listen/ 2026-07-06 https://dejan.ai/concepts/overfitting/listen/ 2026-07-06 https://dejan.ai/concepts/tokenizer/listen/ 2026-07-06 https://dejan.ai/blog/chromes-new-embedding-model/listen/ 2025-04-19 https://dejan.ai/blog/comment-bots/listen/ 2025-08-28 https://dejan.ai/concepts/context-window/listen/ 2026-07-06 https://dejan.ai/concepts/transfer-learning/listen/ 2026-07-06 https://dejan.ai/concepts/language-model/listen/ 2026-07-06 https://dejan.ai/concepts/large-language-model/listen/ 2026-07-06 https://dejan.ai/concepts/token/listen/ 2026-07-06 https://dejan.ai/concepts/mixture-of-experts/listen/ 2026-07-06 https://dejan.ai/blog/aio/listen/ 2026-06-21 https://dejan.ai/blog/cross-model-circuit-analysis-gemini-vs-gemma-comparison-framework/listen/ 2025-03-29 https://dejan.ai/blog/dvl/listen/ 2025-12-18 https://dejan.ai/blog/emotions-gemma/listen/ 2026-05-17 https://dejan.ai/blog/fanout-query-analysis/listen/ 2026-03-20 https://dejan.ai/blog/selection-rate-optimization/listen/ 2026-06-21 https://dejan.ai/blog/ai-seo/listen/ 2026-06-21 https://dejan.ai/blog/grounding-snippets/listen/ 2026-06-21 https://dejan.ai/blog/relevance-engineering/listen/ 2026-06-21 https://dejan.ai/blog/geo/listen/ 2026-06-21 https://dejan.ai/concepts/quadratic-attention-cost/listen/ 2026-07-07 https://dejan.ai/blog/googles-ranking-signals/listen/ 2025-12-24 https://dejan.ai/blog/gpt-5-made-seo-irreplaceable/listen/ 2025-08-10 https://dejan.ai/concepts/needle-in-a-haystack/listen/ 2026-07-07 https://dejan.ai/blog/cascading-hallucination/listen/ 2026-07-07 https://dejan.ai/blog/grounding/listen/ 2026-06-13 https://dejan.ai/blog/grounding-citation-mention/listen/ 2026-07-07 https://dejan.ai/blog/how-long-are-web-pages/listen/ 2025-12-14 https://dejan.ai/blog/coe/listen/ 2026-07-08 https://dejan.ai/blog/how-user-prompts-shape-your-content-visibility-in-ai-search/listen/ 2025-12-13 https://dejan.ai/blog/implicit-queries-in-ai-search/listen/ 2026-02-24 https://dejan.ai/blog/in-ai-seo-10-is-the-new-1/listen/ 2025-11-09 https://dejan.ai/blog/inside-chromes-semantic-engine-a-technical-analysis-of-history-embeddings/listen/ 2025-08-21 https://dejan.ai/blog/live-blog-hacking-gemini-embeddings/listen/ 2025-05-24 https://dejan.ai/blog/llm-is-a-presentation-layer-in-ai-search/listen/ 2025-09-21 https://dejan.ai/blog/most-people-dont-read/listen/ 2025-12-30 https://dejan.ai/blog/neural-circuit-analysis-framework-for-brand-mention-optimization/listen/ 2025-03-29 https://dejan.ai/blog/otsu/listen/ 2025-07-09 https://dejan.ai/blog/primary-bias/listen/ 2026-06-21 https://dejan.ai/concepts/aeo-answer-engine-optimization/listen/ 2026-06-21 https://dejan.ai/concepts/agentic-harness/listen/ 2026-07-06 https://dejan.ai/concepts/ai-assistant/listen/ 2026-07-06 https://dejan.ai/concepts/ai-content-detection/listen/ 2026-07-06 https://dejan.ai/concepts/ai-visibility/listen/ 2026-06-21 https://dejan.ai/concepts/aio-ai-optimization/listen/ 2026-06-21 https://dejan.ai/blog/report-ai-workplace/listen/ 2025-12-10 https://dejan.ai/blog/search-grounding-is-transient/listen/ 2026-03-06 https://dejan.ai/blog/sr/listen/ 2025-09-04 https://dejan.ai/blog/strategic-brand-positioning-in-llms-a-methodological-framework-for-prompt-engineering-and-model-behavior-analysis/listen/ 2025-03-29 https://dejan.ai/blog/blockrank/listen/ 2025-11-10 https://dejan.ai/blog/the-open-knowledge-format-okf/listen/ 2026-06-21 https://dejan.ai/blog/transitions/listen/ 2024-11-27 https://dejan.ai/blog/why-deep-learning-works/listen/ 2024-12-26 https://dejan.ai/blog/average-paradox/listen/ 2026-07-12 https://dejan.ai/blog/ecommerce-optimization-engine/listen/ 2026-07-09 https://dejan.ai/blog/brave-ai-search/listen/ 2026-07-13 https://dejan.ai/concepts/binary-classification/listen/ 2026-07-06 https://dejan.ai/concepts/binary-vector-embeddings/listen/ 2026-07-06 https://dejan.ai/concepts/brand-authority/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/brand-authority.jpg Brand Authority How much cognitive space a brand occupies in a model's memory — how readily and prominently it recalls the brand unprompted. https://dejan.ai/media/video/concepts/brand-authority.mp4 77 https://dejan.ai/blog/how-search-grounding-biased-an-llm-against-youtube/watch/ 2026-06-18 https://dejan.ai/media/video/posts/how-search-grounding-biased-an-llm-against-youtube.jpg How Search Grounding Biased an LLM Against YouTube An analysis of how Claude's webinar platform recommendations were influenced by affiliate-driven content, and a correction regarding YouTube's live features. https://dejan.ai/media/video/posts/how-search-grounding-biased-an-llm-against-youtube.mp4 89 https://dejan.ai/concepts/ai-overviews/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/ai-overviews.jpg AI Overviews Google's AI-generated answer summaries atop search results, which leaked signals suggest are built on its Dialogflow agentic framework. https://dejan.ai/media/video/concepts/ai-overviews.mp4 77 https://dejan.ai/concepts/ai-mode/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/ai-mode.jpg AI Mode Google's AI-powered search experience — essentially Gemini with search, calculator, time, location and Python tools — answering queries conversationally. https://dejan.ai/media/video/concepts/ai-mode.mp4 91 https://dejan.ai/blog/googles-still-doesnt-see-your-live-page/watch/ 2026-05-07 https://dejan.ai/media/video/posts/googles-still-doesnt-see-your-live-page.jpg Google’s (still) doesn’t see your live page. Australian AI SEO agency specialising in brand visibility optimisation for global brands and e-commerce websites using machine learning and data science. https://dejan.ai/media/video/posts/googles-still-doesnt-see-your-live-page.mp4 53 https://dejan.ai/concepts/rank/watch/ 2026-06-21 https://dejan.ai/media/video/concepts/rank.jpg Rank How prominently your citations and mentions appear within an AI answer. https://dejan.ai/media/video/concepts/rank.mp4 51 https://dejan.ai/concepts/chrome-history-embeddings/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/chrome-history-embeddings.jpg Chrome History Embeddings Chrome's local system that turns pages you visit into vectors, enabling natural-language search of your own browsing history on-device. https://dejan.ai/media/video/concepts/chrome-history-embeddings.mp4 72 https://dejan.ai/blog/reverse-prompting/watch/ 2026-03-18 https://dejan.ai/media/video/posts/reverse-prompting.jpg Reverse Prompting: Reconstructing Prompts from AI-Generated Text A fine-tuned Gemma 3 270M model reconstructs the most likely prompts from AI-generated responses using synthetic data and contrastive search configurations. https://dejan.ai/media/video/posts/reverse-prompting.mp4 82 https://dejan.ai/blog/rufus/watch/ 2026-03-15 https://dejan.ai/media/video/posts/rufus.jpg Rufus – Under the Hood. What Drives Amazon’s AI Shopping Assistant? An overview of the technical architecture behind Amazon's Rufus, covering its query planning, RAG-based retrieval, custom LLM models, and streaming response. https://dejan.ai/media/video/posts/rufus.mp4 92 https://dejan.ai/concepts/citation-mining/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/citation-mining.jpg Citation Mining Systematically running many prompts across AI platforms to record which domains and URLs get cited — mapping the AI citation landscape for a topic. https://dejan.ai/media/video/concepts/citation-mining.mp4 73 https://dejan.ai/blog/what-extraction-method-is-google-using-to-build-grounding-snippets/watch/ 2026-02-24 https://dejan.ai/media/video/posts/what-extraction-method-is-google-using-to-build-grounding-snippets.jpg What extraction method is Google using to build grounding snippets? An analysis of Google's Gemini grounding pipeline, examining how extractive summarization selects query-focused sentences to build grounding context from web sources. https://dejan.ai/media/video/posts/what-extraction-method-is-google-using-to-build-grounding-snippets.mp4 97 https://dejan.ai/blog/sro-grounding-snippets/watch/ 2026-03-01 https://dejan.ai/media/video/posts/sro-grounding-snippets.jpg SRO & Grounding Snippets Selection Rate Optimization (SRO) is a new discipline focused on visibility in AI-powered search by measuring how often content is selected for grounding. https://dejan.ai/media/video/posts/sro-grounding-snippets.mp4 89 https://dejan.ai/concepts/classifier/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/classifier.jpg Classifier A model that assigns one or more labels to an input — the fundamental pattern behind query intent models, spam detectors, sentiment models, and content detection. https://dejan.ai/media/video/concepts/classifier.mp4 91 https://dejan.ai/concepts/epoch/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/epoch.jpg Epoch One complete pass through the entire training dataset — the basic unit of training progress, used to track how many times the model has seen each example. https://dejan.ai/media/video/concepts/epoch.mp4 93 https://dejan.ai/blog/how-big-are-googles-grounding-chunks/watch/ 2025-12-20 https://dejan.ai/media/video/posts/how-big-are-googles-grounding-chunks.jpg How big are Google’s grounding chunks? Analysis of how Google selects content to ground Gemini-powered AI shows a fixed 2,000-word budget per query, where relevance rank determines word share. https://dejan.ai/media/video/posts/how-big-are-googles-grounding-chunks.mp4 77 https://dejan.ai/concepts/generate-then-ground/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/generate-then-ground.jpg Generate-then-Ground Google's approach of drafting an answer first, then searching for sources to attribute it — which we argue is more fragile than retrieving first. https://dejan.ai/media/video/concepts/generate-then-ground.mp4 80 https://dejan.ai/concepts/ai-rank/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/ai-rank.jpg AI Rank Our AI visibility and rank-tracking framework that probes language models to map how they perceive and associate brands. https://dejan.ai/media/video/concepts/ai-rank.mp4 75 https://dejan.ai/concepts/caps/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/caps.jpg CAPS Our proposed Content Attribution Payment Scheme — a framework to fairly attribute and pay creators when AI systems use their content. https://dejan.ai/media/video/concepts/caps.mp4 82 https://dejan.ai/blog/how-do-people-use-ai-assistants/watch/ 2025-12-05 https://dejan.ai/media/video/posts/how-do-people-use-ai-assistants.jpg How do people use AI assistants? An analysis of 3.9 million AI chat sessions reveals that most interactions are short, non-commercial, and involve users seeking help with writing, learning, or coding. https://dejan.ai/media/video/posts/how-do-people-use-ai-assistants.mp4 89 https://dejan.ai/concepts/generative-self-retrieval/watch/ 2026-06-27 https://dejan.ai/media/video/concepts/generative-self-retrieval.jpg Generative self-retrieval A reasoning model surfacing facts from its own parameters by generating them as text, where those generated facts then condition and improve its final answer. https://dejan.ai/media/video/concepts/generative-self-retrieval.mp4 84 https://dejan.ai/blog/gemini-3-hallucinates-fan-out-queries/watch/ 2025-11-22 https://dejan.ai/media/video/posts/gemini-3-hallucinates-fan-out-queries.jpg Gemini 3 hallucinates fan-out queries An analysis of Gemini 3 API responses reveals the model fabricating search queries to justify its answers, demonstrating persistent hallucination behaviors. https://dejan.ai/media/video/posts/gemini-3-hallucinates-fan-out-queries.mp4 82 https://dejan.ai/concepts/geo-generative-engine-optimization/watch/ 2026-06-21 https://dejan.ai/media/video/concepts/geo-generative-engine-optimization.jpg GEO (Generative Engine Optimization) Generative Engine Optimization. A name for optimising to appear in generative AI engines. https://dejan.ai/media/video/concepts/geo-generative-engine-optimization.mp4 71 https://dejan.ai/concepts/associative-embeddedness/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/associative-embeddedness.jpg Associative Embeddedness A measure of how deeply a brand is woven into a model's memory, from how often and how early the model recalls it across many runs. https://dejan.ai/media/video/concepts/associative-embeddedness.mp4 84 https://dejan.ai/concepts/blockrank/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/blockrank.jpg BlockRank A method that makes LLM in-context ranking scale to large candidate sets by exploiting the block-sparse structure of attention. https://dejan.ai/media/video/concepts/blockrank.mp4 97 https://dejan.ai/concepts/grounding-chunk/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/grounding-chunk.jpg Grounding Chunk The slice of a page's content an AI system actually uses to support a sentence, drawn from a fixed per-query word budget. https://dejan.ai/media/video/concepts/grounding-chunk.mp4 83 https://dejan.ai/concepts/content-fetcher/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/content-fetcher.jpg Content Fetcher One of AI Mode's two page-reading tools — it retrieves a batch of structured sources at once, versus browsing a single URL. https://dejan.ai/media/video/concepts/content-fetcher.mp4 67 https://dejan.ai/blog/fan-out-query-search-volume-prediction-using-deep-learning/watch/ 2025-08-30 https://dejan.ai/media/video/posts/fan-out-query-search-volume-prediction-using-deep-learning.jpg Fan-Out Query Search Volume Prediction Using Deep Learning A deep learning approach using a Query Demand Estimator to automatically predict search volume ranges for long-tail queries generated by a fan-out model. https://dejan.ai/media/video/posts/fan-out-query-search-volume-prediction-using-deep-learning.mp4 83 https://dejan.ai/concepts/hyperparameters/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/hyperparameters.jpg Hyperparameters Settings configured before training begins that control how a model learns — learning rate, batch size, epoch count, dropout rate — as distinct from the weights the model learns itself. https://dejan.ai/media/video/concepts/hyperparameters.mp4 85 https://dejan.ai/blog/understanding-and-control/watch/ 2025-08-17 https://dejan.ai/media/video/posts/understanding-and-control.jpg Understanding and Control AI optimization relies on mechanistic interpretability to understand internal neural computations and model steering to actively control model behavior. https://dejan.ai/media/video/posts/understanding-and-control.mp4 86 https://dejan.ai/blog/people-call-them-ai-thats-it/watch/ 2025-08-16 https://dejan.ai/media/video/posts/people-call-them-ai-thats-it.jpg People call them AI. That’s it. Social media poll results from 864 votes show that while AI is the dominant label for tools like ChatGPT and Claude, users remain divided on preferred terms. https://dejan.ai/media/video/posts/people-call-them-ai-thats-it.mp4 86 https://dejan.ai/blog/googles-query-fan-out-system-a-technical-overview/watch/ 2025-08-09 https://dejan.ai/media/video/posts/googles-query-fan-out-system-a-technical-overview.jpg Google’s Query Fan-Out System – A Technical Overview This article describes a system that replicates Google's query fan-out approach by using generative neural networks to automatically create intelligent search variants. https://dejan.ai/media/video/posts/googles-query-fan-out-system-a-technical-overview.mp4 99 https://dejan.ai/concepts/implicit-queries/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/implicit-queries.jpg Implicit Queries Searches an AI system formulates and runs on your behalf without you typing them, based on your on-device context and profile. https://dejan.ai/media/video/concepts/implicit-queries.mp4 77 https://dejan.ai/blog/training-a-query-fan-out-model/watch/ 2025-06-24 https://dejan.ai/media/video/posts/training-a-query-fan-out-model.jpg Training a Query Fan-Out Model Google generates high-quality query reformulations by traversing the mathematical latent space between queries and documents to train the qsT5 model. https://dejan.ai/media/video/posts/training-a-query-fan-out-model.mp4 104 https://dejan.ai/blog/universal-query-classifier/watch/ 2025-06-13 https://dejan.ai/media/video/posts/universal-query-classifier.jpg Universal Query Classifier A zero-shot, multi-label search query classifier that maps queries to any user-provided label taxonomy without the need for retraining or bespoke models. https://dejan.ai/media/video/posts/universal-query-classifier.mp4 79 https://dejan.ai/concepts/document-grounding/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/document-grounding.jpg Document Grounding Grounding an AI answer in the content of a specific web page you supply, rather than in results from an open web search. https://dejan.ai/media/video/concepts/document-grounding.mp4 68 https://dejan.ai/blog/ai-mode-page-indexing/watch/ 2025-05-30 https://dejan.ai/media/video/posts/ai-mode-page-indexing.jpg AI Mode & Page Indexing Tests indicate Google's AI Mode uses a proprietary content store rather than the live web, as it fails to fetch indexed pages that are otherwise ranking. https://dejan.ai/media/video/posts/ai-mode-page-indexing.mp4 75 https://dejan.ai/blog/what-is-geo/watch/ 2025-06-02 https://dejan.ai/media/video/posts/what-is-geo.jpg What is GEO? Generative Engine Optimisation (GEO) is a term used to describe SEO for AI assistants and generative search engines, often based on a single research paper. https://dejan.ai/media/video/posts/what-is-geo.mp4 77 https://dejan.ai/concepts/inference/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/inference.jpg Inference Running a trained model on new inputs to produce predictions — the production mode of a model, as distinct from the training phase where weights are learned. https://dejan.ai/media/video/concepts/inference.mp4 104 https://dejan.ai/concepts/llm-search-volume-prediction/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/llm-search-volume-prediction.jpg LLM Search Volume Prediction Using a language model to estimate a query's monthly search volume — directionally useful for sizing topics, but not precise. https://dejan.ai/media/video/concepts/llm-search-volume-prediction.mp4 84 https://dejan.ai/concepts/longformcontextresult/watch/ 2026-07-05 https://dejan.ai/media/video/concepts/longformcontextresult.jpg LongFormContextResult The data wrapper Chrome uses to deliver extracted page content to Gemini's context window. It separates external source material from conversation, ensuring the model grounds its answers in the page rather than its training data. Each block carries a session-wide index used for citation mapping. https://dejan.ai/media/video/concepts/longformcontextresult.mp4 81 https://dejan.ai/concepts/loss-function/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/loss-function.jpg Loss Function The mathematical measure of how wrong a model's predictions are — the quantity training minimises, and the choice of which shapes everything about how the model learns. https://dejan.ai/media/video/concepts/loss-function.mp4 97 https://dejan.ai/concepts/cosine-similarity/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/cosine-similarity.jpg Cosine Similarity A measure of how alike two vectors are by the angle between them; for normalised vectors it's identical to a dot product. https://dejan.ai/media/video/concepts/cosine-similarity.mp4 66 https://dejan.ai/concepts/domdistiller/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/domdistiller.jpg DomDistiller Chrome's Reader Mode engine, whose content-extraction heuristics are a public proxy for how machines separate main content from boilerplate. https://dejan.ai/media/video/concepts/domdistiller.mp4 79 https://dejan.ai/concepts/dynamic-retrieval/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/dynamic-retrieval.jpg Dynamic Retrieval Google's mechanism for deciding, per query, whether Gemini runs a live search — governed by a default grounding threshold of 0.3. https://dejan.ai/media/video/concepts/dynamic-retrieval.mp4 72 https://dejan.ai/concepts/content-substance-classification/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/content-substance-classification.jpg Content Substance Classification Detecting whether content actually says something or just sounds like it does — separating genuine substance from eloquent fluff. https://dejan.ai/media/video/concepts/content-substance-classification.mp4 87 https://dejan.ai/concepts/multimodal-model/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/multimodal-model.jpg Multimodal Model A model that processes and reasons across more than one type of input — text, images, audio, video — within a single unified architecture rather than routing each modality to a separate specialist. https://dejan.ai/media/video/concepts/multimodal-model.mp4 94 https://dejan.ai/concepts/on-device-models/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/on-device-models.jpg On-Device Models Machine-learning models that run locally in software like Chrome — fast, private, and shaping how your content is read before it reaches a server. https://dejan.ai/media/video/concepts/on-device-models.mp4 77 https://dejan.ai/concepts/open-knowledge-format/watch/ 2026-06-21 https://dejan.ai/media/video/concepts/open-knowledge-format.jpg Open Knowledge Format An open, vendor-neutral format for packaging the knowledge AI systems need, as plain markdown files any model or agent can read. https://dejan.ai/media/video/concepts/open-knowledge-format.mp4 41 https://dejan.ai/concepts/grounding-classifier/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/grounding-classifier.jpg Grounding Classifier Our model that predicts whether a query "deserves grounding" — whether an AI system will run a live web search to answer it. https://dejan.ai/media/video/concepts/grounding-classifier.mp4 77 https://dejan.ai/blog/brands/watch/ 2026-03-28 https://dejan.ai/media/video/posts/brands.jpg AI Brand Authority Index: Ranking 2.9 Million Brands by Associative Embeddedness in Gemini’s Memory This research presents a methodology for quantifying brand authority in large language model memory using Personalized PageRank and directed association graphs. https://dejan.ai/media/video/posts/brands.mp4 88 https://dejan.ai/blog/query-length-vs-volume/watch/ 2026-03-12 https://dejan.ai/media/video/posts/query-length-vs-volume.jpg Is Query Length a Reliable Predictor of Search Volume? An analysis of 39.6 million Amazon search queries reveals that query length is an unreliable predictor of search volume compared to semantic content. https://dejan.ai/media/video/posts/query-length-vs-volume.mp4 112 https://dejan.ai/concepts/annotated-page-content-apc/watch/ 2026-07-05 https://dejan.ai/media/video/concepts/annotated-page-content-apc.jpg Annotated Page Content (APC) The structured, machine-readable representation Chrome builds from a page's rendering tree when a tab is shared with Gemini. It captures every visible element — text, links, images, forms, tables — as a tree of content nodes, each tagged with geometry, styling, interaction data, and a unique node ID https://dejan.ai/media/video/concepts/annotated-page-content-apc.mp4 85 https://dejan.ai/concepts/grounding/watch/ 2026-06-21 https://dejan.ai/media/video/concepts/grounding.jpg Grounding The process by which an AI system answers with current information: it runs a search, retrieves pages, reads short extracts, and writes its answer from them rather than from memory. https://dejan.ai/media/video/concepts/grounding.mp4 44 https://dejan.ai/blog/advanced-prompting-techniques/watch/ 2025-11-27 https://dejan.ai/media/video/posts/advanced-prompting-techniques.jpg Advanced Prompting Techniques for AI SEO Explore prompt engineering techniques for SEO, including zero-shot, few-shot, role, and chain-of-thought prompting to improve content and automate tasks. https://dejan.ai/media/video/posts/advanced-prompting-techniques.mp4 94 https://dejan.ai/blog/googles-trajectory-2026-and-beyond/watch/ 2025-12-25 https://dejan.ai/media/video/posts/googles-trajectory-2026-and-beyond.jpg Google’s Trajectory: 2026 and Beyond Google's shift toward agentic AI involves Gemini robotics, A2UI for secure interfaces, and the AP2 protocol for autonomous agent payments and commerce. https://dejan.ai/media/video/posts/googles-trajectory-2026-and-beyond.mp4 102 https://dejan.ai/concepts/embeddinggemma/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/embeddinggemma.jpg EmbeddingGemma Google's compact open embedding model, a miniaturised relative of Gemini, that turns text into meaning-carrying vectors for search understanding. https://dejan.ai/media/video/concepts/embeddinggemma.mp4 74 https://dejan.ai/blog/caps/watch/ 2025-09-30 https://dejan.ai/media/video/posts/caps.jpg CAPS: A Content Attribution Payment Scheme for the AI Era The collapse of the web's economic model due to AI is addressed through the Content Attribution Payment Scheme, a framework for micropayments and grounding. https://dejan.ai/media/video/posts/caps.mp4 82 https://dejan.ai/blog/ai-search-filter/watch/ 2025-11-08 https://dejan.ai/media/video/posts/ai-search-filter.jpg How much of your content survives the AI Search filter? An analysis of the Google grounding process, detailing how user prompts and source snippets are processed by models and measuring citation coverage rates. https://dejan.ai/media/video/posts/ai-search-filter.mp4 91 https://dejan.ai/blog/ai-boom/watch/ 2025-09-01 https://dejan.ai/media/video/posts/ai-boom.jpg The Latent History of AI Boom An exploration of how the transition from RNNs to transformers and the discovery of double descent enabled the scaling of large language models like GPT. https://dejan.ai/media/video/posts/ai-boom.mp4 97 https://dejan.ai/blog/how-gpt-sees-the-web/watch/ 2025-11-14 https://dejan.ai/media/video/posts/how-gpt-sees-the-web.jpg How GPT Sees the Web A technical walkthrough of how GPT handles web search, including snippets, expansions, context size settings, and the sliding window mechanism for retrieval. https://dejan.ai/media/video/posts/how-gpt-sees-the-web.mp4 89 https://dejan.ai/concepts/muvera/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/muvera.jpg MUVERA A Google method for multi-vector retrieval that compresses many per-token vectors into a single fixed-dimensional encoding for fast search. https://dejan.ai/media/video/concepts/muvera.mp4 88 https://dejan.ai/concepts/otsus-thresholding/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/otsus-thresholding.jpg Otsu's Thresholding (for Queries) An unsupervised method borrowed from computer vision that sets optimal per-label confidence cut-offs in query classification by maximizing the variance between positive and negative score clusters. https://dejan.ai/media/video/concepts/otsus-thresholding.mp4 69 https://dejan.ai/blog/human-friendly-content-is-ai-friendly-content/watch/ 2025-07-21 https://dejan.ai/media/video/posts/human-friendly-content-is-ai-friendly-content.jpg Human Friendly Content is AI Friendly Content Explore the parallels between human and AI attention mechanisms and learn how to optimize content for both through scannable structures and hierarchy. https://dejan.ai/media/video/posts/human-friendly-content-is-ai-friendly-content.mp4 80 https://dejan.ai/blog/what-does-an-seo-do-in-the-ai-age/watch/ 2025-08-19 https://dejan.ai/media/video/posts/what-does-an-seo-do-in-the-ai-age.jpg What does an SEO do in the AI age? Modern search engines use a hybrid structure consisting of a strategic Agentic Layer for decision-making and an Interpretative Layer for generative synthesis. https://dejan.ai/media/video/posts/what-does-an-seo-do-in-the-ai-age.mp4 84 https://dejan.ai/blog/llm-search-volume/watch/ 2025-05-19 https://dejan.ai/media/video/posts/llm-search-volume.jpg LLM-Based Search Volume Prediction An analysis comparing Google Gemini's keyword volume predictions against actual Google Search Console data reveals weak-to-moderate correlation and limited accuracy. https://dejan.ai/media/video/posts/llm-search-volume.mp4 78 https://dejan.ai/blog/your-website-is-about-to-start-talking-are-you-ready-for-this/watch/ 2025-08-21 https://dejan.ai/media/video/posts/your-website-is-about-to-start-talking-are-you-ready-for-this.jpg Your website is about to start talking. Are you ready for this? Explore how Chrome's built-in Gemini Nano model uses semantic HTML and the accessibility tree to enable private, on-device AI conversations on websites. https://dejan.ai/media/video/posts/your-website-is-about-to-start-talking-are-you-ready-for-this.mp4 96 https://dejan.ai/blog/tree-walker/watch/ 2025-08-24 https://dejan.ai/media/video/posts/tree-walker.jpg Introducing Tree Walker Tree Walker is an analysis tool designed to deconstruct how AI models like Gemini perceive brands by uncovering word uncertainty and probabilistic language paths. https://dejan.ai/media/video/posts/tree-walker.mp4 81 https://dejan.ai/concepts/webmcp/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/webmcp.jpg WebMCP A proposed web standard that lets sites expose structured tools to AI agents, so agents invoke defined actions instead of screen-scraping the page. https://dejan.ai/media/video/concepts/webmcp.mp4 77 https://dejan.ai/blog/how-gemini-selects-results/watch/ 2024-08-26 https://dejan.ai/media/video/posts/how-gemini-selects-results.jpg How Gemini Selects Results An explanation of how internal algorithms use relevance scoring, recency bias, user intent, and stochasticity to retrieve and present information. https://dejan.ai/media/video/posts/how-gemini-selects-results.mp4 48 https://dejan.ai/blog/site-engagement-metrics/watch/ 2024-11-29 https://dejan.ai/media/video/posts/site-engagement-metrics.jpg Site Engagement Metrics The Google Site Engagement Metrics Framework in Chromium tracks user interactions, engagement scores, and browsing behavior using UMA histograms. https://dejan.ai/media/video/posts/site-engagement-metrics.mp4 73 https://dejan.ai/concepts/matryoshka-representation-learning/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/matryoshka-representation-learning.jpg Matryoshka Representation Learning A training technique that nests coarse-to-fine information inside one embedding, so it can be truncated to fewer dimensions with little loss. https://dejan.ai/media/video/concepts/matryoshka-representation-learning.mp4 77 https://dejan.ai/concepts/ai-seo/watch/ 2026-06-21 https://dejan.ai/media/video/concepts/ai-seo.jpg AI SEO A common name for AI-centric SEO: the practice of working toward AI Visibility. https://dejan.ai/media/video/concepts/ai-seo.mp4 57 https://dejan.ai/blog/beyond-rank-tracking-analyzing-brand-perceptions-through-language-model-association-networks/watch/ 2025-02-27 https://dejan.ai/media/video/posts/beyond-rank-tracking-analyzing-brand-perceptions-through-language-model-associat.jpg Beyond Rank Tracking: Analyzing Brand Perceptions Through Language Model Association Networks The DEJAN methodology uses large language models to analyze brand perception and semantic associations, moving beyond traditional keyword rank tracking. https://dejan.ai/media/video/posts/beyond-rank-tracking-analyzing-brand-perceptions-through-language-model-associat.mp4 74 https://dejan.ai/blog/gemini-toknizer/watch/ 2025-06-05 https://dejan.ai/media/video/posts/gemini-toknizer.jpg Dissecting Gemini’s Tokenizer and Token Scores Explore how Google’s Gemini processes text using subword tokenization. Use this tool to inspect SentencePiece log-likelihood scores for common and rare tokens. https://dejan.ai/media/video/posts/gemini-toknizer.mp4 77 https://dejan.ai/blog/content-substance-classification/watch/ 2025-04-23 https://dejan.ai/media/video/posts/content-substance-classification.jpg Content Substance Classification Cyberfluff is a novel approach for detecting low-quality web content using curriculum-driven contrastive pretraining to distinguish fluff from substance. https://dejan.ai/media/video/posts/content-substance-classification.mp4 85 https://dejan.ai/concepts/reverse-prompting/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/reverse-prompting.jpg Reverse Prompting Reconstructing the most likely prompt that generated a piece of AI output — used for detection, content auditing, and understanding how models respond to specific framings. https://dejan.ai/media/video/concepts/reverse-prompting.mp4 91 https://dejan.ai/blog/how-ai-mode-selects-snippets/watch/ 2025-05-28 https://dejan.ai/media/video/posts/how-ai-mode-selects-snippets.jpg How AI Mode Selects Snippets An analysis of how Google selects content for AI Mode snippets, identifying patterns in value propositions, HTML structure, and semantic selection criteria. https://dejan.ai/media/video/posts/how-ai-mode-selects-snippets.mp4 76 https://dejan.ai/blog/search-query-quality-classifier/watch/ 2024-08-31 https://dejan.ai/media/video/posts/search-query-quality-classifier.jpg Search Query Quality Classifier A search query classifier using ALBERT architecture to identify well-formed queries with 80% accuracy, improving upon Google's LSTM-based model by 10%. https://dejan.ai/media/video/posts/search-query-quality-classifier.mp4 63 https://dejan.ai/blog/the-next-chapter-of-search-get-ready-to-influence-the-robots/watch/ 2025-03-19 https://dejan.ai/media/video/posts/the-next-chapter-of-search-get-ready-to-influence-the-robots.jpg The Next Chapter of Search: Get Ready to Influence the Robots Explore the evolving landscape of SEO, focusing on how AI, conversational search, and Large Language Models are changing brand representation and visibility. https://dejan.ai/media/video/posts/the-next-chapter-of-search-get-ready-to-influence-the-robots.mp4 87 https://dejan.ai/blog/what-does-gemini-think-about-your-brand/watch/ 2025-01-29 https://dejan.ai/media/video/posts/what-does-gemini-think-about-your-brand.jpg What does Gemini think about your brand? Chrome Dev includes a quantized Gemini model for tasks like scam prevention. This analysis examines its on-device execution and reverse-engineered prompts. https://dejan.ai/media/video/posts/what-does-gemini-think-about-your-brand.mp4 71 https://dejan.ai/blog/generative-self-retrieval/watch/ 2026-06-27 https://dejan.ai/media/video/posts/generative-self-retrieval.jpg Generative Self-Retrieval: How AI Models Recall Brand Facts From Memory When an AI answers about your brand from memory, generative self-retrieval decides whether it recalls you correctly or invents a plausible wrong answer. https://dejan.ai/media/video/posts/generative-self-retrieval.mp4 84 https://dejan.ai/concepts/frequency/watch/ 2026-06-21 https://dejan.ai/media/video/concepts/frequency.jpg Frequency How often your citations and mentions occur across AI queries and over time. https://dejan.ai/media/video/concepts/frequency.mp4 44 https://dejan.ai/concepts/grounding-bias/watch/ 2026-06-21 https://dejan.ai/media/video/concepts/grounding-bias.jpg Grounding Bias A form of secondary bias: how much an AI model defers to, or discounts, retrieved sources once they are grounded into the context, rather than leaning on what it already believed. https://dejan.ai/media/video/concepts/grounding-bias.mp4 72 https://dejan.ai/concepts/content-node/watch/ 2026-07-05 https://dejan.ai/media/video/concepts/content-node.jpg Content Node A single element in Chrome's page representation. Each node carries a type (one of 21, from headings and links to form controls and dialogs), bounding box coordinates, text styling, accessibility metadata, and an ID that the AI uses to reference or interact with that specific element. https://dejan.ai/media/video/concepts/content-node.mp4 81 https://dejan.ai/concepts/in-context-ranking/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/in-context-ranking.jpg In-Context Ranking Re-ranking candidate documents for a query by feeding them all into an LLM's context and letting it pick the most relevant. https://dejan.ai/media/video/concepts/in-context-ranking.mp4 65 https://dejan.ai/concepts/node-id/watch/ 2026-07-05 https://dejan.ai/media/video/concepts/node-id.jpg Node ID A sequential identifier assigned to each content node during extraction via depth-first traversal. These IDs appear in the structured Markdown output as {#ID} references, enabling the AI to target specific elements on the page for reading, citing, or interaction. https://dejan.ai/media/video/concepts/node-id.mp4 96 https://dejan.ai/concepts/number-of-citations/watch/ 2026-06-21 https://dejan.ai/media/video/concepts/number-of-citations.jpg Number of Citations The absolute count of times your pages are cited in AI answers. https://dejan.ai/media/video/concepts/number-of-citations.mp4 42 https://dejan.ai/concepts/number-of-mentions/watch/ 2026-06-21 https://dejan.ai/media/video/concepts/number-of-mentions.jpg Number of Mentions The absolute count of times your brand is named in AI answers. https://dejan.ai/media/video/concepts/number-of-mentions.mp4 35 https://dejan.ai/concepts/grounding-snippet/watch/ 2026-06-21 https://dejan.ai/media/video/concepts/grounding-snippet.jpg Grounding Snippet A short, extractive selection of sentences pulled from a page and supplied to an AI model as the evidence it grounds an answer on; the atomic unit of visibility in AI search. https://dejan.ai/media/video/concepts/grounding-snippet.mp4 59 https://dejan.ai/concepts/query-fanout/watch/ 2026-06-21 https://dejan.ai/media/video/concepts/query-fanout.jpg Query Fanout The step where an AI system splits one prompt into several single-intent sub-queries, each retrieving its own sources; a page can be grounded for one angle of a question and absent for another. https://dejan.ai/media/video/concepts/query-fanout.mp4 48 https://dejan.ai/concepts/ap2/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/ap2.jpg AP2 (Agent Payments Protocol) A payment-agnostic open standard that lets autonomous AI agents authorize and complete transactions securely via cryptographically signed "mandates," removing manual checkout. https://dejan.ai/media/video/concepts/ap2.mp4 68 https://dejan.ai/concepts/secondary-bias/watch/ 2026-06-21 https://dejan.ai/media/video/concepts/secondary-bias.jpg Secondary Bias The post-retrieval layer of selection: how an AI model treats, weights, and is swayed by content once it has been retrieved. Unlike primary bias, it is addressable now. https://dejan.ai/media/video/concepts/secondary-bias.mp4 62 https://dejan.ai/concepts/selection-rate-optimization/watch/ 2026-06-21 https://dejan.ai/media/video/concepts/selection-rate-optimization.jpg Selection Rate Optimization The AI-search counterpart to CTR optimization: improving how often AI systems select your content as the grounding source for their generated answers. https://dejan.ai/media/video/concepts/selection-rate-optimization.mp4 48 https://dejan.ai/concepts/share-of-citations/watch/ 2026-06-21 https://dejan.ai/media/video/concepts/share-of-citations.jpg Share of Citations The proportion of cited sources in AI answers that are yours. https://dejan.ai/media/video/concepts/share-of-citations.mp4 54 https://dejan.ai/concepts/share-of-voice/watch/ 2026-06-21 https://dejan.ai/media/video/concepts/share-of-voice.jpg Share of Voice Your overall slice of the AI answer space for a topic, relative to competitors. https://dejan.ai/media/video/concepts/share-of-voice.mp4 46 https://dejan.ai/concepts/selection-rate/watch/ 2026-06-21 https://dejan.ai/media/video/concepts/selection-rate.jpg Selection Rate The rate at which an AI system chooses a given source from the available grounding candidates when composing an answer; the AI-search equivalent of click-through rate. https://dejan.ai/media/video/concepts/selection-rate.mp4 54 https://dejan.ai/concepts/latent-entities/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/latent-entities.jpg Latent Entities A critical piece of information a title deliberately withholds — the subject, reason, process or outcome — to force a click. https://dejan.ai/media/video/concepts/latent-entities.mp4 81 https://dejan.ai/concepts/primary-bias/watch/ 2026-06-21 https://dejan.ai/media/video/concepts/primary-bias.jpg Primary Bias An AI model's ungrounded confidence in an entity, formed during training and present before any retrieval; in AI search it is the largest single factor in whether a source is selected. https://dejan.ai/media/video/concepts/primary-bias.mp4 67 https://dejan.ai/concepts/relevance-engineering/watch/ 2026-06-21 https://dejan.ai/media/video/concepts/relevance-engineering.jpg Relevance Engineering Engineering a page's semantic relevance to a query with embeddings and vector math, treating visibility as an engineering problem rather than keyword optimization. https://dejan.ai/media/video/concepts/relevance-engineering.mp4 62 https://dejan.ai/concepts/share-of-mentions/watch/ 2026-06-21 https://dejan.ai/media/video/concepts/share-of-mentions.jpg Share of Mentions The proportion of brand mentions in AI answers that are yours. https://dejan.ai/media/video/concepts/share-of-mentions.mp4 74 https://dejan.ai/concepts/url-context-tool/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/url-context-tool.jpg URL Context Tool Google's Gemini capability that fetches and reads the content of a specific supplied URL, enabling document grounding. https://dejan.ai/media/video/concepts/url-context-tool.mp4 78 https://dejan.ai/concepts/web-search-grounding/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/web-search-grounding.jpg Web Search Grounding The process where an AI model runs a live web search, reads the results, and weaves some of those pages into its answer. https://dejan.ai/media/video/concepts/web-search-grounding.mp4 82 https://dejan.ai/blog/linkjebert/watch/ 2026-07-05 https://dejan.ai/media/video/posts/linkjebert.jpg Introducing LinkjeBERT a Dutch Language Model LinkjeBERT is a Dutch language model trained on structured Markdown to predict where humans will place links in text using token-level confidence levels. https://dejan.ai/media/video/posts/linkjebert.mp4 99 https://dejan.ai/concepts/tree-walker/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/tree-walker.jpg Tree Walker Our brand-analysis tool that reveals how Gemini talks about a brand by walking the probabilistic paths of the model's own language. https://dejan.ai/media/video/concepts/tree-walker.mp4 84 https://dejan.ai/concepts/tokenization/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/tokenization.jpg Tokenization Breaking text into subword tokens a model can process; each token carries a frequency bias from how often it appeared in training. https://dejan.ai/media/video/concepts/tokenization.mp4 75 https://dejan.ai/concepts/model-context-protocol/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/model-context-protocol.jpg Model Context Protocol An open protocol that gives AI models a standard way to connect to external tools and data sources; WebMCP brings the idea to websites. https://dejan.ai/media/video/concepts/model-context-protocol.mp4 86 https://dejan.ai/concepts/temperature/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/temperature.jpg Temperature A setting that rescales next-word probabilities to control randomness — low values make output focused, high values more diverse. https://dejan.ai/media/video/concepts/temperature.mp4 90 https://dejan.ai/concepts/mechanistic-interpretability/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/mechanistic-interpretability.jpg Mechanistic Interpretability Techniques for tracing what happens inside a model — logging activations and patching them — to see how and where it forms an output. https://dejan.ai/media/video/concepts/mechanistic-interpretability.mp4 80 https://dejan.ai/concepts/query-deserves-grounding/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/query-deserves-grounding.jpg Query Deserves Grounding Google's internal decision about whether a query needs a live web search; below its threshold, the model answers from memory. https://dejan.ai/media/video/concepts/query-deserves-grounding.mp4 71 https://dejan.ai/concepts/vector-embedding/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/vector-embedding.jpg Vector Embedding A dense numeric representation of text that captures meaning and intent, so machines can compare content by similarity rather than keywords. https://dejan.ai/media/video/concepts/vector-embedding.mp4 66 https://dejan.ai/concepts/vector-search/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/vector-search.jpg Vector Search Finding the most relevant content by comparing meaning-carrying vectors, ranking by nearest neighbour rather than keyword overlap. https://dejan.ai/media/video/concepts/vector-search.mp4 75 https://dejan.ai/concepts/rufus/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/rufus.jpg Rufus Amazon's AI shopping assistant — a multi-component RAG system that plans a query, retrieves across Amazon's own sources, and streams an answer. https://dejan.ai/media/video/concepts/rufus.mp4 83 https://dejan.ai/concepts/token-probability/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/token-probability.jpg Token Probability The likelihood a model assigns to each candidate next token; low probabilities flag where the model is uncertain about your brand or topic. https://dejan.ai/media/video/concepts/token-probability.mp4 71 https://dejan.ai/concepts/vector-embedding-optimization/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/vector-embedding-optimization.jpg Vector Embedding Optimization Choosing embedding dimension and quantisation to balance speed, storage and accuracy — often gaining efficiency with little loss. https://dejan.ai/media/video/concepts/vector-embedding-optimization.mp4 82 https://dejan.ai/concepts/ai-agent/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/ai-agent.jpg AI Agent An AI system that plans and executes multi-step tasks autonomously — researching, calling tools, and acting toward a goal rather than just answering. https://dejan.ai/media/video/concepts/ai-agent.mp4 81 https://dejan.ai/concepts/documentchunker/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/documentchunker.jpg DocumentChunker Chrome's internal Blink algorithm that recursively walks a page's DOM layout tree to group text nodes into semantically cohesive passages of up to 200 words. https://dejan.ai/media/video/concepts/documentchunker.mp4 62 https://dejan.ai/concepts/ai-crawlers/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/ai-crawlers.jpg AI Crawlers The bots AI systems use — training-data scrapers versus agentic assistants — that site owners must decide whether to allow or block. https://dejan.ai/media/video/concepts/ai-crawlers.mp4 78 https://dejan.ai/concepts/semantic-compression/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/semantic-compression.jpg Semantic Compression Writing dense, self-contained passages so that when an AI system extracts an isolated grounding chunk, the core context like the brand or product name travels with it. https://dejan.ai/media/video/concepts/semantic-compression.mp4 62 https://dejan.ai/concepts/a2ui/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/a2ui.jpg A2UI (Agent-to-User Interface) An open web standard that lets AI agents generate rich, interactive interfaces as declarative data, rendered through a client app's own trusted components instead of raw code. https://dejan.ai/media/video/concepts/a2ui.mp4 75 https://dejan.ai/concepts/retrieval-augmented-generation/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/retrieval-augmented-generation.jpg Retrieval-Augmented Generation A framework that retrieves relevant passages first, then feeds query plus evidence to the model so it answers from real text. https://dejan.ai/media/video/concepts/retrieval-augmented-generation.mp4 77 https://dejan.ai/concepts/site-search-in-ai-mode/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/site-search-in-ai-mode.jpg Site Search in AI Mode Vertex AI's website-search capability that lets an owner run an AI-Mode-style search over their own content, confirming the Prepare→Retrieve→Signal→Serve flow. https://dejan.ai/media/video/concepts/site-search-in-ai-mode.mp4 74 https://dejan.ai/concepts/query-classification/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/query-classification.jpg Query Classification Assigning search queries to labels — like intent categories — so they can be routed, mapped to content, and reported at scale. https://dejan.ai/media/video/concepts/query-classification.mp4 83 https://dejan.ai/concepts/top-p-sampling/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/top-p-sampling.jpg Top-p Sampling A decoding method that samples the next word from the smallest set of top candidates whose probabilities sum to a threshold p. https://dejan.ai/media/video/concepts/top-p-sampling.mp4 88 https://dejan.ai/concepts/brand-association-network/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/brand-association-network.jpg Brand Association Network The graph of entities, concepts, sentiments, and competitor brands a language model associates with a given brand, surfaced through systematic probing and tracked over time. https://dejan.ai/media/video/concepts/brand-association-network.mp4 103 https://dejan.ai/concepts/training-data/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/training-data.jpg Training Data The labelled or unlabelled examples a model learns from — the single biggest determinant of what a model knows, what it believes, and which brands and entities it associates with which topics. https://dejan.ai/media/video/concepts/training-data.mp4 103 https://dejan.ai/blog/aeo/watch/ 2026-06-21 https://dejan.ai/media/video/posts/aeo.jpg AEO (Answer Engine Optimization) AEO (Answer Engine Optimization) is optimizing to be the answer that AI assistants and answer engines give to a question. https://dejan.ai/media/video/posts/aeo.mp4 64 https://dejan.ai/concepts/attention/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/attention.jpg Attention The transformer mechanism that lets a model weigh how much every token relates to every other, in parallel — the core of modern LLMs. https://dejan.ai/media/video/concepts/attention.mp4 82 https://dejan.ai/concepts/context-engineering/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/context-engineering.jpg Context Engineering Deliberately constructing the semantic environment of a prompt to activate specific representational circuits within a model — moving beyond keyword targeting toward architecture-aware content design. https://dejan.ai/media/video/concepts/context-engineering.mp4 84 https://dejan.ai/blog/ai-bots/watch/ 2025-11-26 https://dejan.ai/media/video/posts/ai-bots.jpg To block or not to block? Bot is the question. An overview of AI bots, distinguishing between training data scrapers used for LLM development and agentic bots designed for autonomous, goal-oriented tasks. https://dejan.ai/media/video/posts/ai-bots.mp4 82 https://dejan.ai/concepts/activation-steering/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/activation-steering.jpg Activation Steering Injecting a computed direction vector into a model's residual stream at inference time to bias its output toward a target behaviour — without any weight updates or retraining. https://dejan.ai/media/video/concepts/activation-steering.mp4 106 https://dejan.ai/blog/ai-content-detection/watch/ 2025-04-17 https://dejan.ai/media/video/posts/ai-content-detection.jpg AI Content Detection DEJAN-LM is an AI content detection model trained on 20 million sentences, using a combined deep learning and heuristic approach to identify advanced AI text. https://dejan.ai/media/video/posts/ai-content-detection.mp4 79 https://dejan.ai/concepts/query-well-formedness/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/query-well-formedness.jpg Query Well-formedness The property of a search query being grammatically natural and unambiguous — a signal used to distinguish full-sentence questions from fragmentary keyword strings, with implications for query expansion and intent classification. https://dejan.ai/media/video/concepts/query-well-formedness.mp4 84 https://dejan.ai/concepts/model-distillation/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/model-distillation.jpg Model Distillation Training a smaller student model to replicate the behaviour of a larger teacher model, producing compact models that punch above their parameter count. https://dejan.ai/media/video/concepts/model-distillation.mp4 86 https://dejan.ai/blog/ai-mode-content-search-index/watch/ 2025-12-13 https://dejan.ai/media/video/posts/ai-mode-content-search-index.jpg AI Mode, Content & Search Index Tests suggest Google’s AI Mode uses a proprietary content store rather than retrieving live web content from the search index during the query fan out process. https://dejan.ai/media/video/posts/ai-mode-content-search-index.mp4 63 https://dejan.ai/concepts/model-probing/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/model-probing.jpg Model Probing Systematically querying a language model with structured prompts to map its latent biases, brand associations, and internal knowledge graphs without relying on traditional rank trackers. https://dejan.ai/media/video/concepts/model-probing.mp4 82 https://dejan.ai/concepts/manifold-hypothesis/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/manifold-hypothesis.jpg Manifold Hypothesis The principle that natural data lies on a low-dimensional manifold within high-dimensional space — the theoretical foundation for why vector embeddings capture meaning rather than noise. https://dejan.ai/media/video/concepts/manifold-hypothesis.mp4 67 https://dejan.ai/concepts/synthetic-training-data/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/synthetic-training-data.jpg Synthetic Training Data AI-generated datasets used to train or fine-tune language models; in AI SEO, a potential early-stage strategy for influencing which brands and associations future models learn. https://dejan.ai/media/video/concepts/synthetic-training-data.mp4 84 https://dejan.ai/concepts/recall/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/recall.jpg Recall Of all the actual positive cases in the data, the fraction the model correctly identified — a measure of how much the model misses. https://dejan.ai/media/video/concepts/recall.mp4 76 https://dejan.ai/blog/ai-mode-is-not-live-web/watch/ 2025-05-29 https://dejan.ai/media/video/posts/ai-mode-is-not-live-web.jpg AI Mode is Not Live Web An experiment testing Google's AI Mode suggests it may rely on Google's existing index or cached web data rather than performing live HTTP requests for all URLs. https://dejan.ai/media/video/posts/ai-mode-is-not-live-web.mp4 85 https://dejan.ai/concepts/class-imbalance/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/class-imbalance.jpg Class Imbalance When one class in a training dataset vastly outnumbers another, causing a naive model to ignore the minority class — requires deliberate handling through weighted loss, resampling, or specialised loss functions. https://dejan.ai/media/video/concepts/class-imbalance.mp4 84 https://dejan.ai/concepts/site-engagement-score/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/site-engagement-score.jpg Site Engagement Score Chrome's internal 0–100 per-domain user engagement metric, built from navigation events, input interactions, and media consumption, decaying over time when a site is not visited. https://dejan.ai/media/video/concepts/site-engagement-score.mp4 92 https://dejan.ai/concepts/checkpoint/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/checkpoint.jpg Checkpoint A saved snapshot of a model's weights at a point during training, enabling recovery, comparison across training stages, and selection of the best-performing version. https://dejan.ai/media/video/concepts/checkpoint.mp4 89 https://dejan.ai/concepts/f1-score/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/f1-score.jpg F1 Score The harmonic mean of precision and recall — the standard single-number summary of classifier performance when both false positives and false negatives matter. https://dejan.ai/media/video/concepts/f1-score.mp4 95 https://dejan.ai/blog/ai-visibility/watch/ 2026-06-21 https://dejan.ai/media/video/posts/ai-visibility.jpg AI Visibility AI Visibility is the outcome of AI-centric SEO, focusing on being cited and recommended by AI systems through brand mentions and source citations in answers. https://dejan.ai/media/video/posts/ai-visibility.mp4 80 https://dejan.ai/concepts/early-stopping/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/early-stopping.jpg Early Stopping Halting training when validation performance stops improving rather than running for a fixed number of epochs — the primary defence against overfitting in practice. https://dejan.ai/media/video/concepts/early-stopping.mp4 94 https://dejan.ai/concepts/fine-tuning/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/fine-tuning.jpg Fine-tuning Continuing training of a pre-trained model on a smaller, task-specific dataset to specialise its behaviour — the standard way DEJAN adapts base models into production classifiers. https://dejan.ai/media/video/concepts/fine-tuning.mp4 97 https://dejan.ai/blog/annotated-page-content-apc/watch/ 2025-09-22 https://dejan.ai/media/video/posts/annotated-page-content-apc.jpg Annotated Page Content (APC) Annotated Page Content (APC) is a structured protobuf representation of a webpage's layout and content, designed for actionable and efficient downstream use. https://dejan.ai/media/video/posts/annotated-page-content-apc.mp4 78 https://dejan.ai/concepts/focal-loss/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/focal-loss.jpg Focal Loss A modified cross-entropy loss that down-weights easy, well-classified examples so training focuses on hard cases — the standard solution when positive examples are rare. https://dejan.ai/media/video/concepts/focal-loss.mp4 91 https://dejan.ai/concepts/precision/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/precision.jpg Precision Of all the times a model predicted positive, the fraction it was actually right — a measure of how trustworthy positive predictions are. https://dejan.ai/media/video/concepts/precision.mp4 89 https://dejan.ai/blog/chrome-context-gemini/watch/ 2026-07-05 https://dejan.ai/media/video/posts/chrome-context-gemini.jpg Google Uses Chrome to Supply Context to Gemini Chat An analysis of Chrome's Page Content Agent, which builds a structured tree of page elements to provide Gemini with grounded context for reading and interaction. https://dejan.ai/media/video/posts/chrome-context-gemini.mp4 93 https://dejan.ai/concepts/multi-label-classification/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/multi-label-classification.jpg Multi-label Classification A classification task where each input can be assigned more than one label simultaneously — a query can be both informational and local; a sentence can be both positive and sarcastic. https://dejan.ai/media/video/concepts/multi-label-classification.mp4 106 https://dejan.ai/concepts/pre-training/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/pre-training.jpg Pre-training Training a model from scratch on a large general corpus to build broad language understanding before any task-specific work begins. https://dejan.ai/media/video/concepts/pre-training.mp4 94 https://dejan.ai/concepts/lora/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/lora.jpg LoRA Low-Rank Adaptation — a parameter-efficient fine-tuning technique that trains small adapter matrices instead of updating a model's full weights, dramatically reducing memory and compute requirements. https://dejan.ai/media/video/concepts/lora.mp4 103 https://dejan.ai/concepts/overfitting/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/overfitting.jpg Overfitting When a model memorises training examples rather than learning generalisable patterns — it scores well on training data but poorly on new inputs it hasn't seen. https://dejan.ai/media/video/concepts/overfitting.mp4 93 https://dejan.ai/concepts/tokenizer/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/tokenizer.jpg Tokenizer The component that converts raw text into the sequence of tokens a model actually processes — fixed at pre-training time, and different for every model family. https://dejan.ai/media/video/concepts/tokenizer.mp4 103 https://dejan.ai/blog/chromes-new-embedding-model/watch/ 2025-04-19 https://dejan.ai/media/video/posts/chromes-new-embedding-model.jpg Chrome’s New Embedding Model: Smaller, Faster, Same Quality Chrome's latest update features a new text embedding model that is 57% smaller than its predecessor, using int8 quantization to maintain search quality. https://dejan.ai/media/video/posts/chromes-new-embedding-model.mp4 79 https://dejan.ai/blog/comment-bots/watch/ 2025-08-28 https://dejan.ai/media/video/posts/comment-bots.jpg Comprehensive Guide to Identifying AI Comment Bots Identify AI-generated comments through statistical analysis of sentiment, formulaic linguistic patterns, repetitive vocabulary, and a lack of human imperfection. https://dejan.ai/media/video/posts/comment-bots.mp4 87 https://dejan.ai/concepts/context-window/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/context-window.jpg Context Window The maximum number of tokens a model can read at once — the total working memory available for a prompt, grounding documents, conversation history, and the generated response combined. https://dejan.ai/media/video/concepts/context-window.mp4 90 https://dejan.ai/concepts/transfer-learning/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/transfer-learning.jpg Transfer Learning Applying knowledge learned on one task or dataset to a different but related task — the principle that makes fine-tuning a pre-trained model far more efficient than training from scratch. https://dejan.ai/media/video/concepts/transfer-learning.mp4 97 https://dejan.ai/concepts/language-model/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/language-model.jpg Language Model A model trained to understand and generate text by learning the statistical patterns of language — the foundational class of model underpinning search, AI assistants, and every DEJAN classifier. https://dejan.ai/media/video/concepts/language-model.mp4 83 https://dejan.ai/concepts/large-language-model/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/large-language-model.jpg Large Language Model A language model at sufficient scale — billions of parameters, trained on web-scale text — to exhibit emergent capabilities like reasoning, instruction following, and open-ended generation. Gemini, GPT, Claude, and Llama are all LLMs. https://dejan.ai/media/video/concepts/large-language-model.mp4 87 https://dejan.ai/concepts/token/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/token.jpg Token The atomic unit a language model reads and generates — a subword chunk, not a whole word — that determines how text is measured, priced, and processed. https://dejan.ai/media/video/concepts/token.mp4 94 https://dejan.ai/concepts/mixture-of-experts/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/mixture-of-experts.jpg Mixture of Experts An architecture where only a subset of specialised sub-networks (experts) activate for each token, allowing a model to have enormous total capacity while spending only a fraction of that compute on any single input. https://dejan.ai/media/video/concepts/mixture-of-experts.mp4 105 https://dejan.ai/blog/aio/watch/ 2026-06-21 https://dejan.ai/media/video/posts/aio.jpg AIO (AI Optimization) AIO (AI Optimization) is shaping your content and presence so AI systems favor them in generated responses — one route to AI Visibility. https://dejan.ai/media/video/posts/aio.mp4 64 https://dejan.ai/blog/cross-model-circuit-analysis-gemini-vs-gemma-comparison-framework/watch/ 2025-03-29 https://dejan.ai/media/video/posts/cross-model-circuit-analysis-gemini-vs-gemma-comparison-framework.jpg Cross-Model Circuit Analysis: Gemini vs. Gemma Comparison Framework A framework for comparative circuit analysis between Google's Gemini and Gemma models to identify how different architectures represent brand information. https://dejan.ai/media/video/posts/cross-model-circuit-analysis-gemini-vs-gemma-comparison-framework.mp4 87 https://dejan.ai/blog/dvl/watch/ 2025-12-18 https://dejan.ai/media/video/posts/dvl.jpg Dynamic Visual Layouts Dynamic visual layout (DVL) is a generative user interface where layouts are created on demand to suit specific queries, shifting the focus from SEO to information. https://dejan.ai/media/video/posts/dvl.mp4 98 https://dejan.ai/blog/emotions-gemma/watch/ 2026-05-17 https://dejan.ai/media/video/posts/emotions-gemma.jpg Emotion Geometry of Google’s AI Models A replication study of Anthropic’s emotion research on Google’s Gemma 4 31B model, finding that internal emotion representations organize along a valence axis. https://dejan.ai/media/video/posts/emotions-gemma.mp4 101 https://dejan.ai/blog/fanout-query-analysis/watch/ 2026-03-20 https://dejan.ai/media/video/posts/fanout-query-analysis.jpg Fanout Query Analysis An analysis of 365,920 fanout queries from Google, OpenAI, and Amazon reveals how different AI models generate internal search queries for web grounding. https://dejan.ai/media/video/posts/fanout-query-analysis.mp4 82 https://dejan.ai/blog/selection-rate-optimization/watch/ 2026-06-21 https://dejan.ai/media/video/posts/selection-rate-optimization.jpg Selection Rate Optimization Selection Rate Optimization is the AI-search counterpart to click-through-rate optimization: improving how often AI systems choose your content as the source they ground and cite their answers on. https://dejan.ai/media/video/posts/selection-rate-optimization.mp4 88 https://dejan.ai/blog/ai-seo/watch/ 2026-06-21 https://dejan.ai/media/video/posts/ai-seo.jpg AI SEO AI SEO is optimizing your content and brand so AI systems surface them when answering questions — the broad practice behind AI Visibility. https://dejan.ai/media/video/posts/ai-seo.mp4 72 https://dejan.ai/blog/grounding-snippets/watch/ 2026-06-21 https://dejan.ai/media/video/posts/grounding-snippets.jpg Grounding Snippets Grounding snippets are the short, extractive sentences AI systems pull from your page to build their answers: the atomic unit of AI-search visibility, where most of your page never makes the cut. https://dejan.ai/media/video/posts/grounding-snippets.mp4 89 https://dejan.ai/blog/relevance-engineering/watch/ 2026-06-21 https://dejan.ai/media/video/posts/relevance-engineering.jpg Relevance Engineering Relevance Engineering is the practice of building a page's semantic relevance to a query with embeddings and vector math, treating search visibility as an engineering problem rather than keyword optimization. https://dejan.ai/media/video/posts/relevance-engineering.mp4 89 https://dejan.ai/blog/geo/watch/ 2026-06-21 https://dejan.ai/media/video/posts/geo.jpg GEO (Generative Engine Optimization) GEO (Generative Engine Optimization) is optimizing for LLM-powered search and chat so your brand and pages appear in generated answers. https://dejan.ai/media/video/posts/geo.mp4 58 https://dejan.ai/concepts/quadratic-attention-cost/watch/ 2026-07-07 https://dejan.ai/media/video/concepts/quadratic-attention-cost.jpg Quadratic Attention Cost The property that standard transformer attention grows with the square of the input length — doubling the context roughly quadruples the compute and memory — which is why long context windows are expensive and why grounding budgets exist. https://dejan.ai/media/video/concepts/quadratic-attention-cost.mp4 91 https://dejan.ai/blog/googles-ranking-signals/watch/ 2025-12-24 https://dejan.ai/media/video/posts/googles-ranking-signals.jpg Google’s Ranking Signals Overview of search ranking factors including popularity signals, PCTR models, semantic relevance, keyword matching, freshness, and various search modes. https://dejan.ai/media/video/posts/googles-ranking-signals.mp4 92 https://dejan.ai/blog/gpt-5-made-seo-irreplaceable/watch/ 2025-08-10 https://dejan.ai/media/video/posts/gpt-5-made-seo-irreplaceable.jpg GPT-5 Made SEO Irreplaceable OpenAI is shifting its model design to prioritize reasoning and intelligence over memorized world knowledge, relying on tools and retrieval for information. https://dejan.ai/media/video/posts/gpt-5-made-seo-irreplaceable.mp4 98 https://dejan.ai/concepts/needle-in-a-haystack/watch/ 2026-07-07 https://dejan.ai/media/video/concepts/needle-in-a-haystack.jpg Needle in a Haystack A stress test that plants a single fact deep inside a long context and checks whether the model can retrieve it — the standard way to measure whether a large context window is genuinely usable, or just advertised. https://dejan.ai/media/video/concepts/needle-in-a-haystack.mp4 95 https://dejan.ai/blog/cascading-hallucination/watch/ 2026-07-07 https://dejan.ai/media/video/posts/cascading-hallucination.jpg Cascading Hallucination A novel failure mode in multi-agent AI systems: a capable model trusts its sub-agents to deliver grounded facts, but they hallucinate, and the orchestrator launders the fiction into a confident answer without ever questioning it. https://dejan.ai/media/video/posts/cascading-hallucination.mp4 116 https://dejan.ai/blog/grounding/watch/ 2026-06-13 https://dejan.ai/media/video/posts/grounding.jpg How AI Search Grounding Actually Works: Google vs OpenAI vs Anthropic An analysis of how Google, OpenAI, and Anthropic handle web grounding, comparing their search processes, citation rates, and how they process page content. https://dejan.ai/media/video/posts/grounding.mp4 100 https://dejan.ai/blog/grounding-citation-mention/watch/ 2026-07-07 https://dejan.ai/media/video/posts/grounding-citation-mention.jpg Grounding Source ≠ Citation ≠ Mention Being a cited grounding source does not mean your brand was mentioned in the answer, and being mentioned does not mean you were cited. They are correlated but distinct events produced at different stages of the AI answer pipeline — and conflating them is the source of most AI-visibility measurement https://dejan.ai/media/video/posts/grounding-citation-mention.mp4 106 https://dejan.ai/blog/how-long-are-web-pages/watch/ 2025-12-14 https://dejan.ai/media/video/posts/how-long-are-web-pages.jpg How Long Are Web Pages? An analysis of 44,684 web pages reveals a median content length of 3,201 tokens and an average of 10,403 tokens, highlighting implications for AI systems. https://dejan.ai/media/video/posts/how-long-are-web-pages.mp4 74 https://dejan.ai/blog/how-user-prompts-shape-your-content-visibility-in-ai-search/watch/ 2025-12-13 https://dejan.ai/media/video/posts/how-user-prompts-shape-your-content-visibility-in-ai-search.jpg How user prompts shape your content visibility in AI search. An analysis of how AI search rankers use semantic alignment to surface different content zones within a single article based on query specificity and intent. https://dejan.ai/media/video/posts/how-user-prompts-shape-your-content-visibility-in-ai-search.mp4 84 https://dejan.ai/blog/implicit-queries-in-ai-search/watch/ 2026-02-24 https://dejan.ai/media/video/posts/implicit-queries-in-ai-search.jpg Implicit Queries in AI Search An analysis of Google patent US11769017B1, detailing a system that uses context and implied input engines to proactively generate and push AI summaries. https://dejan.ai/media/video/posts/implicit-queries-in-ai-search.mp4 94 https://dejan.ai/blog/in-ai-seo-10-is-the-new-1/watch/ 2025-11-09 https://dejan.ai/media/video/posts/in-ai-seo-10-is-the-new-1.jpg In AI SEO #10 is the new #1 An empirical study analyzing how Google's AI Mode uses text snippets from multiple sources, finding that snippets are more prompt-aligned than full web pages. https://dejan.ai/media/video/posts/in-ai-seo-10-is-the-new-1.mp4 86 https://dejan.ai/blog/inside-chromes-semantic-engine-a-technical-analysis-of-history-embeddings/watch/ 2025-08-21 https://dejan.ai/media/video/posts/inside-chromes-semantic-engine-a-technical-analysis-of-history-embeddings.jpg Inside Chrome’s Semantic Engine: A Technical Analysis of History Embeddings Technical analysis of Chrome's history embeddings system, detailing the DocumentChunker algorithm, passage extraction, and the 1540-dimensional vector pipeline. https://dejan.ai/media/video/posts/inside-chromes-semantic-engine-a-technical-analysis-of-history-embeddings.mp4 91 https://dejan.ai/blog/live-blog-hacking-gemini-embeddings/watch/ 2025-05-24 https://dejan.ai/media/video/posts/live-blog-hacking-gemini-embeddings.jpg Live Blog: Hacking Gemini Embeddings An experimental study reproducing the vec2vec research paper by attempting to translate and align Gemini and MxbAI embedding spaces using unsupervised methods. https://dejan.ai/media/video/posts/live-blog-hacking-gemini-embeddings.mp4 87 https://dejan.ai/blog/llm-is-a-presentation-layer-in-ai-search/watch/ 2025-09-21 https://dejan.ai/media/video/posts/llm-is-a-presentation-layer-in-ai-search.jpg LLM is a Presentation Layer in AI Search Large language models act as a presentation layer on top of classic information retrieval. They rely on crawling, indexing, and ranking to prevent hallucinations. https://dejan.ai/media/video/posts/llm-is-a-presentation-layer-in-ai-search.mp4 102 https://dejan.ai/blog/most-people-dont-read/watch/ 2025-12-30 https://dejan.ai/media/video/posts/most-people-dont-read.jpg Most People Don’t Read A qualitative study comparing self-reported reading habits against actual user behavior, tracking mouse movements, scroll patterns, and time on page. https://dejan.ai/media/video/posts/most-people-dont-read.mp4 80 https://dejan.ai/blog/neural-circuit-analysis-framework-for-brand-mention-optimization/watch/ 2025-03-29 https://dejan.ai/media/video/posts/neural-circuit-analysis-framework-for-brand-mention-optimization.jpg Neural Circuit Analysis Framework for Brand Mention Optimization This framework uses open-weight models like Gemma 3 Instruct to perform mechanistic brand positioning through direct neural circuit and activation analysis. https://dejan.ai/media/video/posts/neural-circuit-analysis-framework-for-brand-mention-optimization.mp4 96 https://dejan.ai/blog/otsu/watch/ 2025-07-09 https://dejan.ai/media/video/posts/otsu.jpg Dynamic per-label thresholds for large-scale search query classification with Otsu’s method Explore how to use Otsu's algorithm to solve the problem of inconsistent confidence thresholds in search-query intent classifiers using dynamic, per-label tuning. https://dejan.ai/media/video/posts/otsu.mp4 86 https://dejan.ai/blog/primary-bias/watch/ 2026-06-21 https://dejan.ai/media/video/posts/primary-bias.jpg Primary Bias Primary bias is what an AI model already believes about your brand before it searches: an ungrounded confidence baked into training that becomes the biggest factor in whether your content is selected in AI answers. https://dejan.ai/media/video/posts/primary-bias.mp4 99 https://dejan.ai/concepts/aeo-answer-engine-optimization/watch/ 2026-06-21 https://dejan.ai/media/video/concepts/aeo-answer-engine-optimization.jpg AEO (Answer Engine Optimization) Answer Engine Optimization. A name for optimising to appear in AI answer engines. https://dejan.ai/media/video/concepts/aeo-answer-engine-optimization.mp4 55 https://dejan.ai/concepts/agentic-harness/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/agentic-harness.jpg Agentic Harness The orchestration layer around a language model that equips it with tools, memory, and control flow — turning a text predictor into an agent that can plan, act, and loop. https://dejan.ai/media/video/concepts/agentic-harness.mp4 93 https://dejan.ai/concepts/ai-assistant/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/ai-assistant.jpg AI Assistant A conversational AI like ChatGPT, Gemini or Claude used for short, varied tasks — usually one question and one answer, not search-style queries. https://dejan.ai/media/video/concepts/ai-assistant.mp4 87 https://dejan.ai/concepts/ai-content-detection/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/ai-content-detection.jpg AI Content Detection Classifying whether text was written by AI; as models improve, detection needs constant fine-tuning and hybrid deep-learning-plus-heuristic methods. https://dejan.ai/media/video/concepts/ai-content-detection.mp4 81 https://dejan.ai/concepts/ai-visibility/watch/ 2026-06-21 https://dejan.ai/media/video/concepts/ai-visibility.jpg AI Visibility The broad outcome of AI-centric SEO: being seen, cited, and recommended by AI systems when they answer questions. https://dejan.ai/media/video/concepts/ai-visibility.mp4 67 https://dejan.ai/concepts/aio-ai-optimization/watch/ 2026-06-21 https://dejan.ai/media/video/concepts/aio-ai-optimization.jpg AIO (AI Optimization) AI Optimization. Another name for the practice of improving AI Visibility. https://dejan.ai/media/video/concepts/aio-ai-optimization.mp4 44 https://dejan.ai/blog/report-ai-workplace/watch/ 2025-12-10 https://dejan.ai/media/video/posts/report-ai-workplace.jpg Report: How People Use AI at Work An analysis of qualitative interviews with 1,250 professionals exploring how the general workforce, creatives, and scientists integrate AI into their work. https://dejan.ai/media/video/posts/report-ai-workplace.mp4 100 https://dejan.ai/blog/search-grounding-is-transient/watch/ 2026-03-06 https://dejan.ai/media/video/posts/search-grounding-is-transient.jpg Search Grounding is Transient Google’s AI search and Gemini use a single-turn transient architecture that purges raw web snippets from working memory immediately after a response is sent. https://dejan.ai/media/video/posts/search-grounding-is-transient.mp4 89 https://dejan.ai/blog/sr/watch/ 2025-09-04 https://dejan.ai/media/video/posts/sr.jpg Primary Bias on Selection Rate in AI Search Selection Rate measures how often AI systems select specific items from grounding results. It explores primary bias, model relevance, and the Tree Walker algo. https://dejan.ai/media/video/posts/sr.mp4 94 https://dejan.ai/blog/strategic-brand-positioning-in-llms-a-methodological-framework-for-prompt-engineering-and-model-behavior-analysis/watch/ 2025-03-29 https://dejan.ai/media/video/posts/strategic-brand-positioning-in-llms-a-methodological-framework-for-prompt-engine.jpg Strategic Brand Positioning in LLMs: A Methodological Framework for Prompt Engineering and Model Behavior Analysis This paper presents a methodological framework for analyzing and optimizing brand mentions in large language models through systematic prompt probing and analysis. https://dejan.ai/media/video/posts/strategic-brand-positioning-in-llms-a-methodological-framework-for-prompt-engine.mp4 93 https://dejan.ai/blog/blockrank/watch/ 2025-11-10 https://dejan.ai/media/video/posts/blockrank.jpg BlockRank: A Faster, Smarter Way to Rank Documents with LLMs BlockRank is a novel method for in-context ranking that uses structured sparse attention and contrastive training to improve LLM efficiency and accuracy. https://dejan.ai/media/video/posts/blockrank.mp4 88 https://dejan.ai/blog/the-open-knowledge-format-okf/watch/ 2026-06-21 https://dejan.ai/media/video/posts/the-open-knowledge-format-okf.jpg The Open Knowledge Format (OKF) Google Cloud's Open Knowledge Format is an open, vendor-neutral way to package the context AI systems need, as plain markdown files any model or agent can read. https://dejan.ai/media/video/posts/the-open-knowledge-format-okf.mp4 80 https://dejan.ai/blog/transitions/watch/ 2024-11-27 https://dejan.ai/media/video/posts/transitions.jpg Beyond Links: Understanding Page Transitions in Chrome Explore Chrome page transition types and qualifiers to understand user intent, navigation pathways, and the SEO implications of different browser behaviors. https://dejan.ai/media/video/posts/transitions.mp4 92 https://dejan.ai/blog/why-deep-learning-works/watch/ 2024-12-26 https://dejan.ai/media/video/posts/why-deep-learning-works.jpg Why deep learning works. An excerpt from François Chollet’s Deep Learning with Python exploring the manifold hypothesis and how structured information enables deep learning to work. https://dejan.ai/media/video/posts/why-deep-learning-works.mp4 93 https://dejan.ai/blog/brave-ai-search/watch/ 2026-07-13 https://dejan.ai/media/video/posts/brave-ai-search.jpg How Brave's AI Search Works DEJAN AI reveals the system instructions for Brave Ask, detailing how the AI grounding and generation pipeline processes web evidence to produce answers. https://dejan.ai/media/video/posts/brave-ai-search.mp4 101 https://dejan.ai/concepts/binary-classification/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/binary-classification.jpg Binary Classification A model task with exactly two possible output labels — yes/no, spam/not-spam, AI/human — the simplest and most common form of supervised classification. https://dejan.ai/media/video/concepts/binary-classification.mp4 101 https://dejan.ai/concepts/binary-vector-embeddings/watch/ 2026-07-06 https://dejan.ai/media/video/concepts/binary-vector-embeddings.jpg Binary Vector Embeddings Compressing float embeddings to one bit per dimension, reducing storage by ~32× with surprisingly small quality loss — useful when speed and scale matter more than marginal precision. https://dejan.ai/media/video/concepts/binary-vector-embeddings.mp4 99