SITUATION EXPLAINED: Runway just released Praxis-1 with open weights, a robot model trained mostly on regular video instead of robot data.
• Runway, best known for AI video, first showed Praxis-1 in September and tested it with early partners before opening it up
• Its bet is that robot demonstration data is scarce, but video is "effectively limitless"
• In Runway's tests, pretraining on web video worked about as well as pretraining on real robot footage, with 16.1 cm vs 16.0 cm of placement error
• The same policy ran in a studio and a home kitchen without retraining
• Runway's benchmark results, it beats Black Forest Labs' FLUX 3 Action, NVIDIA’s Cosmos 3 Nano and Physical Intelligence’s π0.5
@sophiadew: "There's some really cool robots that are actually being built with the Praxis-1."
Today, we’re making Praxis-1 available to everyone. Our state-of-the-art world action model turns video pretraining into control for real robots, and it ships with open weights under Apache 2.0, ready to run and fine-tune on your own hardware.
Another example I think about a lot: roads were sized for horses and buggies, and today’s cars are sized for those roads. Most of what AI will change hasn’t been imagined yet - such crazy times.
Love this clip with @maggioant!!
New pod drop!! @RunwayML just turned 8 years and added $100M+ in ARR last quarter alone. I met Anastasis Germanidis (@agermanidis), co-CEO of Runway, who went from art school to building one of the most important video and world models of our generation.
We talk about:
- The Nietzsche-inspired game he tried to build at 15
- “We accidentally created one of the best models for robotics”
- 30 magic tools in 30 days
- Sora launch: a 2-hour existential crisis → a better model in 3 months
- Proof of work > credentials
- Running “a perfect simulation” of his co-CEO Cristóbal (@c_valenzuelab)
- “I’m always becoming someone else”
and so much more!
Timestamps
(00:00) Trailer
(01:21) Anastasis Germanidis, Co-Founder & Co-CEO, Runway
(02:27) Accidentally building one of the best robotics models
(07:18) How video models train robots and self-driving cars
(10:47) What a world model actually is
(16:19) Why folding clothes is still unsolved for AI
(17:37) Why AGI can never be reached
(22:03) A game built on Nietzsche's eternal recurrence
(30:35) Starting an AI company from art school
(39:47) Why staying in stealth is counterproductive
(42:03) 30 magic tools in 30 days
(55:55) Sora's launch and an hour of existential crisis
(57:02) Racing to beat Sora in three months
(1:01:56) Solaris and software generated in real time
(1:05:15) Why AI labs pay for fake versions of Amazon
(1:09:52) Why Sora struggled as a consumer app
(1:16:30) Hiring on proof of work
(1:19:44) Becoming co-CEO with Cristóbal Valenzuela (@c_valenzuelab)
(1:30:56) What he's still becoming
(1:32:52) Outro
The good and the beautiful go together
Recognizing the beautiful is a timeless way to identify the good
So if you want to convince people that what you’re building is good, make it beautiful
The Orange was the first AI video, and it is still my favorite.
Cool to consider that this is the first moving story of generated pixels, stitched together frame by frame by outputs from an image generator 🍊 Shoutout to @matamalaortiz for creating a new category.
Will Smith eating spaghetti was just 2-3 years ago, and for many people it was their first exposure to AI video. But if you look back a few more years, to around 2019, the progress looks even more impressive, and Will Smith eating spaghetti starts to look like high definition.
This is The Orange, our first short film made with Runway, back in 2019 by @matamalaortiz