Runway announced Gen-2, a video generation model. Where Gen-1, announced the month before, applied a different look to an existing video, Gen-2 could produce video from text alone, or from an image together with text. The output was a clip a few seconds long; reporters shown a demonstration described generation taking minutes and the motion as choppy. In March it was open only to selected testers through a Discord waiting list, and it reached anyone with a browser or an iPhone in June. Runway was started in 2018 by three people who met at NYU, and in 2022 it was among the co-developers of Stable Diffusion. Text-to-video already existed in research announcements at the time, but not in a form you could go and try.