Vidit Goel

GenerativeML and 3D @Snap | @IITKGP '21 | Computer Vision, Deep learning

[Vidit Goel @ViditGoel7]

Check our latest updates and improved model for PAIR Diffusion: A Comprehensive Multimodal Object-Level Image Editor🚀🚀

Project page: https://t.co/6aIxv2MAy2
ArXiv: https://t.co/ShNOGF7Ntz

We show that 👇👇

[AK @_akhaliq]

PAIR-Diffusion: Object-Level Image Editing with Structure-and-Appearance Paired Diffusion Models @Gradio demo is out on @huggingface Spaces

demo: https://t.co/hR63tPMl5i

[X Games @XGames]

The best jumps for the best show, built with Bobcat.

[Vidit Goel @ViditGoel7]

Claude has now become true software engineer. It can now create a bug, then report it, fix it and ask for credits.

[Vidit Goel @ViditGoel7]

@zuhayeer @gowthami_s Too low, atleast for specialized ML roles

[Vidit Goel @ViditGoel7]

@gowthami_s @zuhayeer +1

[Vidit Goel @ViditGoel7]

Great addition, just tried searching for the best restaurants in NYC, definitely prefer Gemini output.

[Josh Woodward @joshwoodward]

We're making it easier for @GeminiApp to work across Google. Three weeks ago, it was Google's Shopping Graph and the 50 billion product listings there.

Today, it's Gemini 🤝 Google Maps!

[Vidit Goel @ViditGoel7]

@yash2kant @igilitschenski @UofTCompSci @Snapchat Congratulations 🎉

[Vidit Goel @ViditGoel7]

With generative video models rising, people ask: do we still need 3D/4D? If you want cheap, realtime, interactive experiences — the answer is yes.
We just launched Animated Selfie Attachments in Lens Studio 5.17: generate interactive on-device experiences from a text prompt.

[Sergey Korolev @libfun_sk]

Very happy to bring generative 4D to the hands of our users!

This was a great collab with Snap Research and would not have been possible without @wang12_gordon and @ashmrz10’s paper that was published this year on NeurIPS!

https://t.co/RlGwmzNGea

[Vidit Goel @ViditGoel7]

With generative video models rising, people ask: do we still need 3D/4D?
If you want cheap, realtime, interactive experiences — the answer is yes.
We just launched Animated Selfie Attachments in Lens Studio 5.17: generate interactive on-device experiences from a text prompt.

[Vidit Goel @ViditGoel7]

@deedydas Super cool! How do you think it will affect companies like @ideogram_ai ?

[Vidit Goel @ViditGoel7]

@kshitijgokul What if there was a metal piece inside the coke can for testing

[Vidit Goel @ViditGoel7]

@enggirlfriend snapchat?

[Vidit Goel @ViditGoel7]

Possibly, we can only store high level information like semantics and also some 3D representation in efficient manner and use generative models to decode them to high quality 3D representation of the world where our system can take actions.

[Vidit Goel @ViditGoel7]

Interesting read. Further, I recently noticed that though the brain only takes 20W of energy, it can have 1-2 petabytes of storage!
Moving forward I think we should relax some constraints on long term memory a world model can store.

[Andrew Davison @AjdDavison]

Representation representation representation
#SpatialAI

See the SLAM Handbook Chapter 18 for my views!
https://t.co/EdTa9zcl5F

[Vidit Goel @ViditGoel7]

Even with this large storage representation is highly important. If we record 30 fps video in 512 x 512 resolution we will only be able to record ~1-2 year of data even if we use whole of storage of brain 1-2 PB. Generative models might be useful for compressing data efficiently

[Humphrey Shi @humphrey_shi]

This morning on the way to school, my 8-year-old daughter and I talked about fame and impact. We started with MrBeast and internet celebrities—whose work she knows well—but then I introduced Einstein, whose discoveries shaped the technology we use every day.

Her big question: ‘Do I have to change the world, Dad?’

My answer: you don’t have to aim for that. We just want you to be happy. Do the right thing, be kind, and follow what you love—the rest will follow naturally.

In today’s noisy world of rapid AI and technological breakthroughs and constant announcements, external pressures and expectations can make it feel like only big, visible achievements matter. But doing the right thing—and pursuing what you love, the things that keep you up at night and get you out of bed in the morning—even without instant recognition, is the only path to true happiness and serenity.

A reminder I share with my children, my students—and myself.

[Vidit Goel @ViditGoel7]

RT @humphrey_shi: Multi-agent coding systems (e.g., Claude Code) are sweeping the world like a storm this summer. The success rests on a si…

[Vidit Goel @ViditGoel7]

Hi all, I will be at CVPR in Nashville from 10-15 June. Lets meet!
Also drop by our paper Wonderland: Navigating 3D Scenes from a Single Image
https://t.co/Qe43iqeNrJ
When - Friday morning session
Where - ExHall D Poster #59
#CVPR25

[Vidit Goel @ViditGoel7]

talk by @jon_barron . Completely agree, further if we move towards spatial computing 3D would be definitely needed but again a long term bet.

Full talk: https://t.co/7tUzEKrA3K

[Ali Hassani @AliHassaniJr]

Wondering what's happening with NATTEN in 2025?
Check out Generalized Neighborhood Attention!

Spoiler: NATTEN gets a new stride parameter, we made a simulator for all your analytical studies, AND a Blackwell kernel!

Keep reading for more...

(1 / 5)

[Vidit Goel @ViditGoel7]

@harshit_sikchi @scottniekum @yayitsamyzhang @marcgbellemare @yukez @PeterStone_TX Congratulations!!