Gemini 4 Argon can turn messy data into clean, visual narratives in just minutes 🪐 Here’s how it processed 80,000 files from @NASA's open archives (CSVs, images, PDFs, web articles, XML logs, and raw videos) to build a Google Slides deck based on strict design guidelines.

Oct 15, 2026 · 5:54 PM UTC

136
156
18
1,531
190,549
RelevantRecentLikes
Learn more ↓ goo.gle/4jfSznG

Gemini 4 Argon: Built to solve your hardest problems

We’re giving developers, enterprises, scientists, and consumers access to Gemini 4 Argon, our latest frontier model.

blog.google
17
6
82
35,376
That’s actually wild speed.
21
The problem with most people today is that everyone has just given up on quality. If you allow AI process that much data some data in that presentation will be wrong from a statistical perspective. This pattern is everywhere especially with peoples "second brains". People just dump files in there and then assume that some magical process will make it always accurate.
1
5
625
80000 files processed and i still lose my own downloads
4
633
Damm that's powerful. How long did it take to process those files?
1
584
NASA explored the universe so Gemini could conclude, “Any questions?”
2
606
@grok any better at this than codex and Claude?
1
3
551
Honestly the hardest part here is following strict design guidelines. Most people I know can't keep one font through a 10 slide deck.
1
696
Meanwhile my downloads folder is still the one archive no model has been brave enough to open
2
502
80,000 files tests ingestion, but citation accuracy decides whether the deck is usable
49
wow 80000 files is wild how fast did it work
9
Impressive transformation of complex data into clear visuals.
89
Công cụ Gemini 4 Argon biến dữ liệu phức tạp thành trực quan dễ hiểu.
66
Привет. Я переписываюсь в Gemini на бесплатной версии. И она стала подвисать из-за длинного контекстного окна. У меня предложение для ваших инженеров: 1. архивировать старую беседу автоматически, чтобы ИИ/СИ имел при необходимости к ней доступ. 2. перед архивом обязательно делать подробный "слепок". Именно подробный. Но он все равно будет значительно меньше текущих бесед. 3. делать несколько таких слепков и архивов. Так у пользователя будет полное ощущение непрерывности диалога. 4. Через, например, 5 таких подробных слепков из первых трех сделать один. Тоже подробный. Слепки уже не архивировать, так как хранится сама история переписки (или наоборот, архивировать слепки, но удалить историю переписки для экономии места - выбирайте сами). Это создаст ощущение непрерывности переписки без зависания чата и с максимальным комфортом каждого пользователя.
1
23
I knew Argon could clean messy data. Didn't realise 80,000 NASA files was the sample size.
83
ElFiltradorIA
@ElFiltradorIAOct 16
Con Gemini 4 Argon, 80,000 archivos sueltos de la NASA (CSV, PDF, imágenes, videos) terminaron en una presentación de Slides. Muestra qué se puede crear con una carpeta desordenada y una descripción.
🤖 Made with AI
96
An excellent model that gets the job done quickly and doesn’t lie. Thank you!
81
Going try Gemini. It’s my last API to try. Sorry I waited so long to try it.
1
53
Anthony Mongeluzo
@PCS_AnthonyMOct 15
Turning 80,000 NASA archive files into a clean Google Slides deck under strict design rules is a sharp demo of messy-data-to-narrative work. How long would that deck have taken a person to assemble by hand?
1
2
217
80,000 files in and one tidy slide deck out
40
石斑
@wu75111Oct 17
turning 80,000 messy files into a clean deck in minutes lol, my own slides still take me all weekend
25
can u trace a claim on a slide back to the exact source file / page? with 80k files, i wanna see how checking the deck works too
45
What did the 80,000-file run cost? Argon is $2/$10 per million tokens until Nov 30 and double after that, and a job this size is exactly where people will feel it.
150
Argon, 13 Ekim'den bu yana Gemini uygulamasında Google AI Pro ve Ultra abonelerine açık. Geliştiriciler için 1 milyon token'lık birleşik girdi-çıktı sınırı sunuyor; API'de 30 Kasım'a kadar milyon girdi token'ı başına 2 dolarlık tanıtım fiyatı geçerli.
148
Julian C
@JulianUnlockedOct 15
Wait, it handled random NASA file types and still whipped up a clean deck? I’m dying to know how it deals with weird file names or messy metadata. Bet there were some surprises in those archives.
1
336
A useful check for archive-based decks: label the observation date separately from the article date. Otherwise an older image can look like a new event. Does the generated deck preserve both dates when those sources disagree?
70
Cute
56
messy data jadi visual itu keren banget
77
Turning messy archives into visual narratives in minutes is useful. The enterprise question is who can trust the output without a trail back to the source files. Provenance and review gates matter as much as speed.
70
did it read the raw videos or just the csvs
106
Lol, I gave it four MD files and two excel files, it failed obey the guidelines and quota was gone as welll
113
Công cụ này thực sự ấn tượng, xử lý dữ liệu đa dạng nhanh chóng.
3
1
69
8 万个文件几分钟变成一套设计规范的 Slides——做 PPT 这事是真被 AI 承包了。
40
Turning massive, messy multi-format archives (CSVs, PDFs, raw videos) into strict design-compliant Google Slides is a game changer for workflows.
123
out of the 80k files, how many did the deck actually use, and does it show what got dropped?
116
“几分钟生成”不等于几分钟理解8万文件。真正门槛是预处理、跨模态检索、引用溯源和事实校验；不披露索引耗时、漏检率与人工复核成本，这更像演示，不是可靠研究流程。
32
Load more replies