Gemini 4 Neon
Share

Post Content

 

 [[{“value”:”Gemini 4 Neon is Google’s new model for complex coding, knowledge work and science, and it sits between Gemini 4 Hydrogen and Gemini 4 Carbon. I had early access, so this video skips the benchmarks and shows real outputs: a 3D dragon written as code, an animated SVG peacock in one file, a walkable house from a 1959 floor plan, a swarm of agents building a voxel diorama, an on-call agent deployed on Google Cloud Run, three GPU experiments on a DGX Spark, and a neural network written with no training.

Official Blog: https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-neon/
Try Gemini 4 Neon in Google AI Studio: https://aistudio.google.com/prompts/new_chat?model=gemini-4-neon
Google’s developer guide (What’s new in Gemini 4 Neon): https://aistudio.google.com/docs/whats-new-gemini-4-neon
Keenan Crane’s SDF dragon experiment: https://x.com/keenanisalive/status/2108708523236602056
Growing Neural Cellular Automata (Distill, 2020): https://distill.pub/2020/growing-ca/
Weight-space programming (MIT CSAIL research post): https://rikiyat.github.io/blog/weight-space-programming/
My video on weight-space programming: https://youtu.be/wXJDtImT_-0

Disclosure: I recently joined Google, and these are my own views, and these experiments are based on my early access.

Timestamps:
0:00 Gemini 4 Neon
1:09 Antigravity and the four thinking levels
1:23 A 3D dragon written as one formula
2:40 An animated SVG peacock in one file
3:32 From a floor plan to a walkable house
4:23 A swarm of agents builds a voxel diorama
6:14 An on-call agent on Google Cloud Run
7:43 Breaking the app, then the rollback
9:03 Growing Neural Cellular Automata on a DGX Spark
9:48 Teaching a four-legged robot to run
10:27 Two galaxies colliding
11:09 A neural network written with no training
12:22 Picking the thinking level
12:48 Three rules for long tasks”}]] Read More Prompt Engineering 

#Promptengineering #AI

By ali

Leave a Reply