This is real footage from 126 years ago.
What you are watching is the trottoir roulant, the moving sidewalk, built for the great World's Fair in Paris in 1900.
More than a century ago, three years before the Wright brothers would make the first airplane flight, the city built an electric street that carried you across itself while you simply stood there...
It ran in a loop of around three and a half kilometres, raised on a viaduct above the fairgrounds, with nine stations where you could step on and off.
And it had a clever design: two moving platforms side by side, one going at walking pace and one faster, so you could step onto the slow one first, then onto the quick one, and ride the whole circuit in about twenty-six minutes without taking a single step.
Nearly fifty million people came to that fair, and for most of them it was the first time they had ever moved through a place without taking a step.
The very first moving walkway had appeared seven years earlier, at the World's Fair in Chicago in 1893, built by the same designers. But the Paris version was longer, faster, and far more sophisticated, and it was here that the world truly fell in love with the idea.
It astonished people. The thought that the ground itself could carry you felt like magic, like something out of a dream of the future. They even called it the Rue de l'Avenir: the Street of the Future.
Thomas Edison sent a crew to film it, which is why we can still watch it today...
Best Models Per Use Case - The Un-Banned Version :)
back-end coding - GPT 5.5 xHigh
front-end coding Opus 4.8 xHigh
real-time - Grok 4.3
design - Opus 4.8
video - Seedance 2.0
images - GPT 5.2
real time voice - Flash live
simple coding - GLM 5.2
always-on - Kimi or MiniMax
small classifier - GPT 5.1 mini
every day chat - Gemini Flash 3.5
Combine models for best performance
Create an AGENTS.md file in your projects containing:
Project overview
Setup commands
Test commands
Coding standards
Security requirements
Deployment process
Files agents should never edit
Definition of done
Grok 4.5, based on our 1.5T V9 foundation model, with Cursor data added in supplemental training, is now in private beta at SpaceX & Tesla. Early evals show performance close to, perhaps exceeding Opus.
RL is continuing to significantly improve the model, and the Grok Build harness gets better every day.
Nice work by all those involved!
Completely trained from scratch new models will be released by @SpaceX every month this year.
Building high-quality evals is an increasingly important skill.
Especially if you're trying to land a job or get into AI, I'd recommend trying to benchmark models on a task/domain you care about.
If done well, you'll get the attention of any company training models.
@readswithravi Continuous learning is a lot easier when you have the financial freedom to invest time in it. Survival often comes before self-improvement