--- title: "Reconstructed Slides: Realtime Data Connectivity for AI: Tanmai Gopal" category: "slides" video_id: "himhGiWJXjo" sourceLabels: ["Cropped public YouTube video frames", "Local OpenCV slide-region detection", "Local RapidOCR"] --- # Reconstructed Slides: Realtime Data Connectivity for AI: Tanmai Gopal ## Source Video [Realtime Data Connectivity for AI: Tanmai Gopal](https://www.youtube.com/watch?v=himhGiWJXjo) ## Method This deck is reconstructed from the existing video frame captures by detecting likely slide regions with OpenCV, cropping/upscaling those regions, deduplicating similar crops, and OCRing the cropped slide images locally. It is a cleaner companion to the full-stage frame deck. ## Reconstructed Slides ![[assets/reconstructed-slides/himhGiWJXjo/slide-002.jpg]] - Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/reconstructed/himhGiWJXjo/slide-002.html) - AI slide classifier: `title_card` confidence `0.97` - Text source: agent_vision. Slide text: > Embrace the AI overlords > Give them the data they need. ![[assets/reconstructed-slides/himhGiWJXjo/slide-004.jpg]] - Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/reconstructed/himhGiWJXjo/slide-004.html) - AI slide classifier: `title_card` confidence `0.98` - Text source: agent_vision. Slide text: > Why is the deal with ACME stuck in stage 3? ![[assets/reconstructed-slides/himhGiWJXjo/slide-005.jpg]] - Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/reconstructed/himhGiWJXjo/slide-005.html) - AI slide classifier: `title_card` confidence `0.99` - Text source: agent_vision. Slide text: > We solved this. > Make your data & business logic available as a tool for your LLM. ![[assets/reconstructed-slides/himhGiWJXjo/slide-007.jpg]] - Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/reconstructed/himhGiWJXjo/slide-007.html) - AI slide classifier: `demo_video` confidence `0.94` - Text source: advanced OCR `rapidocr-live/bright-screen/contrast`. - OCR decision: ready — Dense terminal/demo screenshot with small code and UI text; OCR is likely better than manual transcription. Slide text: > coco-aasc:af-engine(esin)s try runchat-d postgrescsrosIGRES-s public > taaPacha Chatmts > Press double Entar after you are done entering your query. > User:help ee vriu > AIE > Wond'sFai ALmbda Microsoft smol ai ![[assets/reconstructed-slides/himhGiWJXjo/slide-008.jpg]] - Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/reconstructed/himhGiWJXjo/slide-008.html) - AI slide classifier: `demo_video` confidence `0.95` - Text source: advanced OCR `rapidocr-live/bright-screen/contrast`. - OCR decision: ready — Dense SQL/code output screenshot with small multi-line text and query results; OCR is better suited here. Slide text: > executor.cutput(ho recentrentals foundfor top custoaer') > executor.output（fTep custoner:[custonernane]（foustoner_eaail])VnRecent sovles:[*,.join（recent.movles)') rocent.novies=[aovietitle]fornovie inrecent.rentals] > AIE merId CROER BY COUNT(reotal.io) DESC LINIT1) SOL Oury: SELEcY flrstnano,last.nae,eall FRO publlc.custo#er WERE custo#r_id(SELECI customer_id FRO publlc.rental GouP 8Y custo > SOLRagponse:[frst_A#e':*ELEANOR,*lasLnaRo:HUNT','e#aiI:LEANOR,HUNTsakSlacustomr.rg'] > SL Quory: SElECT f.title FRo public.rental r3oiN public.1nvontory i ON r.iventory.idi.inventory-ld Joi public.filn f C i.filn.id.rental-date DESC LIHITS fileLid EREr.cstomer_id*(SELECI custoerid FRO public.rental GROuP BY custooer-id CRDER BY COuNT(rental.i) DESC LIET 1) ORDER BYr > SOReSpO: Etitle':RACER EGG'],tItle:NURNY CREATURES'},'Itle':HELLFIGHTERS SIERRA',['itle:IZARO COLDBLOOOED'},[it 1:'DINOSAURSECRETARY] > Dacy Plan Exacytion Result: NT（ELEANOR.NuNTsaklscustomer.org) MURKY CREATURES,HELLFIGHTERS SIERRA,WTZARO COLCBLGOOED,DINOSAUR SECRETARY > Wonr'sFai Lambda Microsoft smol ai ![[assets/reconstructed-slides/himhGiWJXjo/slide-009.jpg]] - Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/reconstructed/himhGiWJXjo/slide-009.html) - AI slide classifier: `diagram` confidence `0.99` - Text source: agent_vision. Slide text: > LLMs don't know what your API is. > But LLMs know what SQL is. > What if everything was SQL? ![[assets/reconstructed-slides/himhGiWJXjo/slide-010.jpg]] - Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/reconstructed/himhGiWJXjo/slide-010.html) - AI slide classifier: `diagram` confidence `0.98` - Text source: agent_vision. Slide text: > Property of the data, property of the session. > WHY. IS. THIS. SO. COMPLEX. ![[assets/reconstructed-slides/himhGiWJXjo/slide-011.jpg]] - Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/reconstructed/himhGiWJXjo/slide-011.html) - AI slide classifier: `title_card` confidence `0.99` - Text source: agent_vision. Slide text: > #3. Get LLMs to figure out the plan of how to access data. ![[assets/reconstructed-slides/himhGiWJXjo/slide-012.jpg]] - Recreated text/layout view: [open HTML recreation](/assets/slide-recreations/reconstructed/himhGiWJXjo/slide-012.html) - AI slide classifier: `title_card` confidence `0.99` - Text source: agent_vision. Slide text: > Write a python program to figure out a plan to retrieve multiple pieces of data required. ### Hidden Non-Slide Evidence - [`slide-001.jpg`](/assets/reconstructed-slides/himhGiWJXjo/slide-001.jpg) — `sponsor_logo` confidence `0.99`; Sponsor/logo slide, not presentation content. - [`slide-003.jpg`](/assets/reconstructed-slides/himhGiWJXjo/slide-003.jpg) — `speaker_stage` confidence `0.99`; Speaker on stage with podium and backdrop, not a slide. - [`slide-006.jpg`](/assets/reconstructed-slides/himhGiWJXjo/slide-006.jpg) — `demo_video` confidence `0.98`; Camera-captured terminal demo footage, not a presentation slide. - [`slide-013.jpg`](/assets/reconstructed-slides/himhGiWJXjo/slide-013.jpg) — `sponsor_logo` confidence `0.97`; Logo/sponsor-style frame with no presentation content or readable slide argument. Classification audit: `raw/sources/slide-ai-classification/reconstructed/himhGiWJXjo/audit.json`