impressive output, but I’m increasingly skeptical that Google will ever release any of this work #
scroll up to read this animated comic about two trees talking to each other via FlowingData #
text-to-audio model, love the nonsensical speech #
another text-to-video model announced today that can handle long videos with multiple prompts, check out the two-minute example #
48K OS for the NES featuring “two core applications, the word processor, and the settings” via arstechnica.com #
impressive open-source web frontend for 20+ emulated game systems and support for custom feeds #
surprisingly good GPT-3 powered summaries of long YouTube videos #
finally played through this Instagram series-turned-videogame and it’s brilliant #
compete with others to get from one article to another in the shortest time using only article links #
after 25 screenshot rotations, the original image is smaller than an atom #
artist Dries Depoorter used AI on video from public webcams to figure out when geotagged Instagram photos were taken via @golan #