The receipts:
→ Brand + type from the brand-kit PDF, not a template.
→ My logo on the title slide + footer, as captioned.
→ Slides 6 + 7: native bar + pie charts from my .xlsx.
→ Every figure cited.
A real .pptx. Yours to fork.
Most AI deck tools give everyone the same deck.
I gave Deckova my files instead: a brand-kit PDF, my logo, an Excel.
14 slides on how Wikipedia works — on-brand, my logo placed as I asked, my spreadsheet as native charts. A real .pptx.
https://t.co/LyVL4vGs9i
#AI#PPTX
The receipts:
→ Slide 7: a real line chart — Journe vs Patek prices. Journe climbs; Patek spikes then falls.
→ Slides 12-14: 34 numbered citations.
→ Every slide: a morph transition.
→ No template: Garamond + gold + ink, picked from "Luxe".
A real .pptx. Yours to fork.
What people want from AI deck tools, rarely get:
→ Real .pptx, not PDF
→ Editable charts, not images
→ Cited sources, not guesses
→ A look that fits, not a template
→ Motion, not stills
14 slides on the quiet watch market. One paragraph in.
https://t.co/KVHGpIglHZ
#AI #PPTX
The receipts, slide by slide:
— Slide 2: a real PowerPoint bar chart of the signal's S/N ratio across the 72 seconds (6 → 14 → 26 → 30 → 19 → 5).
— Slide 6: a real native table — 5 hypotheses × 4 columns of analysis.
— Slides 8-12: 70 numbered citations, arXiv 2508.10657 + arXiv 2408.08513 inline.
— Every slide: a morph transition + a speaker-notes cue card.
A real .pptx file. Yours to fork.
What most AI-deck tools quietly skip:
→ Native .pptx (not PDF)
→ Real charts on real data
→ Real native tables
→ Citations with arXiv IDs
→ Morph transitions
→ Speaker-notes cue cards
→ Subject-fit imagery
13-slide deck on the Wow! Signal from one prompt. All seven inside.
→ https://t.co/KVHGpIglHZ
#AI #PPTX
Three decades of proof, from the deck:
— Gando Primary School (Burkina Faso, 2004): -6°C indoor cooling, no electricity. Architect: Francis Kéré.
— Wadi Hanifa Wetlands (Riyadh, 2010): 200,000 visitors per week, 120 km park. Moriyama & Teshima.
— Khudi Bari (Bangladesh, 2025): $450 per kit, 78 units, 3-day build. Marina Tabassum.
None of these will win a magazine cover. All of them will outlive one.
What lasts requires slowness.
There's a $1M architecture prize you've probably never heard of.
10× the value of the Pritzker medal. Full jury rotation every cycle. Nearly 10,000 projects from 100+ countries in its archive.
The Aga Khan Award. Fifty years old this year.
14-slide essay on it. Built in minutes on Deckova.
→ https://t.co/KVHGpIglHZ
#Architecture #AI
The recurring device across 1100+ chapters:
— Nami: Luffy destroys Arlong Park
— Robin: Luffy renounces the Marines at Enies Lobby
— Sanji: All Blue framed against Krieg's gluttony
— Franky: World Government for Tom's Workers
— Brook: Luffy bringing him home to Laboon
— Chopper: Luffy ridding Drum of Wapol
It's not coincidence. It's the join condition.
Every Straw Hat joins the crew only after Luffy punches out the symbol of their trauma.
Arlong's tower for Nami. Tequila Wolf for Robin. Tom's death for Franky. Brook's wait for Laboon — 50 years.
15-slide essay on Oda's emotional architecture. Built in a few minutes on Deckova.
→ https://t.co/KVHGpIglHZ
#OnePiece #AI
Things the deck did that I didn't ask it to:
— picked Rubik + DM Sans for the type pairing (both free Google Fonts; user can install in 30s)
— composed the 4-phase approach diagram natively (not a rasterised PNG)
— auto-credited the user upload on the references slide
— shipped 12 image objects (3 photos + 9 icons), every one swappable
The brief was: turn a brand overview into a pitch deck. One prompt.
https://t.co/KVHGpIglHZ
A brand-overview TXT in.
A real .pptx out.
No template, no web search, no extra context.
Editorial typography, native diagrams, four-card outcomes dashboard. 57 seconds of footage.
→ https://t.co/LyVL4vGs9i
#AI#PowerPoint
Slide 6 hit hardest:
To simulate one second of a low-resolution Earth at Planck scale, you'd need 14–100 million years of compute on the most efficient computer physics permits — a black hole.
That's just slide 6 of 10.
https://t.co/LyVL4vGs9i
"Are we living in a simulation?"
We typed this into @Deckova_AI. What came back: a researched 10-slide deck with Bostrom's trilemma, thermodynamic bounds, and Bayesian math.
A real PowerPoint file. Not a webview.
https://t.co/LyVL4vGs9i
#AI#SimulationHypothesis