Pete Florence on X: "Going Beyond World Models & VLAs" / X
To view keyboard shortcuts, press question mark View keyboard shortcuts Home Explore Notifications Chat Grok Premium Bookmarks Creator Studio Articles Profile More Post Tasha @TashaPais Article See new posts Conversation Pete Florence @peteflorence Going Beyond World Models & VLAs 43 220 1K 310K In GEN-1, approximately 99% of the parameters are trained from scratch. Previously, this might be considered wild. For Generalist, it’s a deliberate choice. It follows our strong conviction — pursued for two years — that when you have enough data, you can move faster at pushing the frontier by having complete control over the fundamental model. GEN-1 is not a fine-tuned vision-language model with robot actions bolted on, nor is it just a world model. It is a first-class-citizen, native foundation model for physical interaction. And there is growing evidence that if you have enough data and compute, training from scratch always wins. World models are having their moment in early 2026. VLAs had t
Pete Florence @peteflorence Going Beyond World Models & VLAs 2:51 PM · Apr 7, 2026 334K Views 46 0 4 6 162 0 1 6 2 1.1K 0 1 . 1 K 1.2K 0 1 . 2 K Read 46 replies
Explore this link on the map →related reading
- Essays · Gwern.netgwern.net
- RTFM: A Real-Time Frame Model | World Labsworldlabs.ai
- Replicate - Run AI with an APIreplicate.com
- Black Forest Labs - Frontier AI Labbfl.ai
- Lluminatejoelsimon.net
- Go smol or go home | Harm de Vriesharmdevries.com
- Large Language Model: world models or surface statistics?thegradient.pub
- June 2021 News · Gwern.netgwern.net
- How Does A Blind Model See The Earth? - by henryoutsidetext.substack.com
- [2411.02385] How Far is Video Generation from World Model: A Physical Law Perspectivearxiv.org
- The last six months in LLMs, illustrated by pelicans on bicyclessimonwillison.net
- GitHub - mazabou/awesome-neurofm: A curated list of awesome neuro-foundation models · GitHubgithub.com