--- title: "Slides: User Signal Dies at the Retrieval Boundary - Sonam Pankaj, StarlightSearch" category: "slides" video_id: "Jx4ZFEAq6bY" sourceLabels: ["Public YouTube video frames", "Public YouTube metadata"] --- # Slides: User Signal Dies at the Retrieval Boundary - Sonam Pankaj, StarlightSearch ## Source Video [User Signal Dies at the Retrieval Boundary - Sonam Pankaj, StarlightSearch](https://www.youtube.com/watch?v=Jx4ZFEAq6bY) ## Relationship To World's Fair 2026 These slides are extracted from a public AI Engineer YouTube video connected to World's Fair 2026. Speaker-matched clips are supporting context unless later confirmed as exact session recordings; official livestream recordings are day-level/event-level source material. ## Related Scheduled Sessions - No individual scheduled session mapping has been assigned yet; treat this as an event livestream deck. ## Extracted Slides ![[assets/slides/Jx4ZFEAq6bY/slide-001.jpg]] - Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/slides/Jx4ZFEAq6bY/slide-001.html) - AI slide classifier: `content_slide` confidence `0.99` - Text source: agent_vision. Slide text: > User Signal dies at the Retrieval Boundary. > Sonam Pankaj > CEO & Co-Founder StarlightSearch ![[assets/slides/Jx4ZFEAq6bY/slide-002.jpg]] - Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/slides/Jx4ZFEAq6bY/slide-002.html) - AI slide classifier: `content_slide` confidence `0.99` - Text source: agent_vision. Slide text: > What is an agent? > An agent is an LLM with agency: it reasons, invokes TOOLS to interact with the world, RETRIEVES from memory, and loops until the TASK is complete. ![[assets/slides/Jx4ZFEAq6bY/slide-003.jpg]] - Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/slides/Jx4ZFEAq6bY/slide-003.html) - AI slide classifier: `content_slide` confidence `0.99` - Text source: agent_vision. Slide text: > What is an agent? > An agent is an LLM with agency: it reasons, invokes TOOLS to interact with the world, RETRIEVES from memory, and loops until the TASK is complete. ![[assets/slides/Jx4ZFEAq6bY/slide-004.jpg]] - Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/slides/Jx4ZFEAq6bY/slide-004.html) - AI slide classifier: `content_slide` confidence `0.99` - Text source: agent_vision. Slide text: > Retrieval is Static > Context Stuffing ![[assets/slides/Jx4ZFEAq6bY/slide-005.jpg]] - Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/slides/Jx4ZFEAq6bY/slide-005.html) - AI slide classifier: `content_slide` confidence `0.98` - Text source: agent_vision. - OCR decision: ready — Diagram slide with multiple small text blocks and labels; OCR should extract the body copy more reliably than direct transcription. Slide text: > The missing layer between evals and action ![[assets/slides/Jx4ZFEAq6bY/slide-006.jpg]] - Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/slides/Jx4ZFEAq6bY/slide-006.html) - AI slide classifier: `content_slide` confidence `0.99` - Text source: agent_vision. Slide text: > The GAP > It has no access to why yesterday's runs passed or failed. The eval signal dies in a dashboard. > This is the missing layer: a system that consumes traces, absorbs eval outcomes, and converts both into retrievable guidance for future runs. ![[assets/slides/Jx4ZFEAq6bY/slide-007.jpg]] - Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/slides/Jx4ZFEAq6bY/slide-007.html) - AI slide classifier: `content_slide` confidence `0.97` - Text source: advanced OCR `rapidocr-live/bright-screen/opencv-adaptive`. - OCR decision: ready — Dense table and small text are better handled by OCR than manual transcription. Slide text: > In practice, most agent memory frameworks have focused on user continuity: preferences, profile facts, conversation history, and long-lived personalization. > chat experiences is not self-improving learning system for production agents. * > toddy What h storts Rttrlrvsl signa! L+hr ffom outcoemts? > Raw ctut history Recency To > Etacted lktr, protroco(mbodang imirity Ho > Ertity reutiorahips ovtr dme Graph traveryd · recency No > A14r Yerbu sell refvectiors Shmlarty to currert tast Partisty. refietlons capture lesons, but rerieval k not > atemtarx ttitty scores Iall·linhrd roftctomy with Sinilarity weighted by outcomalderred uirty ![[assets/slides/Jx4ZFEAq6bY/slide-008.jpg]] - Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/slides/Jx4ZFEAq6bY/slide-008.html) - AI slide classifier: `content_slide` confidence `0.99` - Text source: agent_vision. Slide text: > Memory as Reasoning ![[assets/slides/Jx4ZFEAq6bY/slide-009.jpg]] - Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/slides/Jx4ZFEAq6bY/slide-009.html) - AI slide classifier: `content_slide` confidence `0.93` - Text source: advanced OCR `rapidocr-live/bright-screen/contrast`. - OCR decision: ready — Browser screenshot with chart labels and small UI text is better suited to OCR. Slide text: > reflect > 米 > t²-bench > Wie compre beg mode on themirinrs tak inT-bench > erhancedperfarrsaroe. > GLMnenthinking OPTSA > AertRTX ![[assets/slides/Jx4ZFEAq6bY/slide-010.jpg]] - Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/slides/Jx4ZFEAq6bY/slide-010.html) - AI slide classifier: `content_slide` confidence `0.94` - Text source: advanced OCR `rapidocr-live/right-72/contrast`. - OCR decision: ready — Screenshot has small chart labels and benchmark numbers that OCR can read more reliably than vision transcription. Slide text: > Fettures Hewiworts nvgwtiors Doca Dog Detymlec > GRTSA > ncA ang Loiong-08 > SAD > 42. M.3. > 3.Y > EnowlesgeFnontir H CodirCan Uielony-ce CNTetk ![[assets/slides/Jx4ZFEAq6bY/slide-011.jpg]] - Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/slides/Jx4ZFEAq6bY/slide-011.html) - AI slide classifier: `content_slide` confidence `0.94` - Text source: advanced OCR `rapidocr-live/bright-screen/contrast`. - OCR decision: ready — Screenshot has small chart labels and benchmark numbers that OCR can read more reliably than vision transcription. Slide text: > ref/ect Howitwors ntogrtienn Gors > Refecton ebenehand Utieng-ce > RAO > 63.3 > P3 > Lieng-Cs ceTeua > AentRTX ![[assets/slides/Jx4ZFEAq6bY/slide-012.jpg]] - Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/slides/Jx4ZFEAq6bY/slide-012.html) - AI slide classifier: `content_slide` confidence `0.99` - Text source: agent_vision. Slide text: > Limitations > Cold Start > Utility Drift > Review quality > Lambda ![[assets/slides/Jx4ZFEAq6bY/slide-015.jpg]] - Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/slides/Jx4ZFEAq6bY/slide-015.html) - AI slide classifier: `content_slide` confidence `0.99` - Text source: agent_vision. Slide text: > 4. Agentbench > AgentRTX 61.3 > Mem0 58 > RAG 47 > Baseline 35.7 > AgentRTX 98 > DSPy 82 > Baseline 57 ![[assets/slides/Jx4ZFEAq6bY/slide-019.jpg]] - Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/slides/Jx4ZFEAq6bY/slide-019.html) - AI slide classifier: `content_slide` confidence `0.88` - Text source: advanced OCR `rapidocr-live/border-trim/contrast`. - OCR decision: ready — Dense product UI screenshot with many small text elements and interface labels that OCR can capture more reliably than manual transcription. Slide text: > reffect Memories 0/pg > PROJIC! Feomceviewedtraces.Sortbytimeora > Oveniw > Prcpects ） > cel > Revievs LEANNEETASE eerou fauebeupug > project Scops > iing CUeage sJafes4s-8fue-Siea-a97f-e29leSoo8rde 68/2026.4251FM FAILED > RETLEOTIGN > SCMUARY > The userreguestedagaming moueDut theagentfalod toretrone anyresits > bndoeppouppu broodersearchterm > COROTACTOM > agentRTx The sgent shcud hse triedabeosde seych_groouctstooltofindreie.antitems.ratherthnghinguparee ### Hidden Non-Slide Evidence - [`slide-013.jpg`](/assets/slides/Jx4ZFEAq6bY/slide-013.jpg) — `demo_video` confidence `0.98`; Code editor and terminal demo footage, not a presentation slide. - [`slide-014.jpg`](/assets/slides/Jx4ZFEAq6bY/slide-014.jpg) — `demo_video` confidence `0.98`; Browser product demo with webcam inset, not a slide. - [`slide-016.jpg`](/assets/slides/Jx4ZFEAq6bY/slide-016.jpg) — `demo_video` confidence `0.98`; Code editor demo footage with terminal prompt, not a slide. - [`slide-017.jpg`](/assets/slides/Jx4ZFEAq6bY/slide-017.jpg) — `demo_video` confidence `0.98`; Browser dashboard demo footage with webcam inset, not a slide. - [`slide-018.jpg`](/assets/slides/Jx4ZFEAq6bY/slide-018.jpg) — `demo_video` confidence `0.98`; Browser reflection/details demo footage with webcam inset, not a slide. Classification audit: `raw/sources/slide-ai-classification/slides/Jx4ZFEAq6bY/audit.json` ## Slide-Derived Subjects To Review Subject extraction uses video title, related session titles/descriptions, transcript context, and OCR text when available. OCR is best-effort and should be reviewed against the embedded slide images.