Click any tag below to further narrow down your results
Links
Google is rolling out Project Genie, a browser tool for generating and exploring interactive worlds, to eligible Google AI Ultra subscribers. The big additions are Genie 4, which lets you change a world mid-session with new prompts and adds native audio, plus a community gallery where people can publish and remix worlds.
- Genie 4 lets you edit a running world by typing new prompts, such as shifting the weather or adding a character, and it adds real-time audio that adapts to the scene.
- Genie 3 gets sessions extended from 60 seconds to 3 minutes per generation, world chaining, and seeding from 360-degree panoramas.
- A new agent builds worlds from natural-language descriptions and can bring Gemini in to suggest next steps or trigger events, and a gallery lets users remix and publish worlds.
Genie 3 is a DeepMind research prototype that turns a text description into a photorealistic 3D environment you can explore in real time. The article frames it as a step toward training AI agents in varied, safe simulated worlds, though it still has clear limits on accuracy and duration.
- Genie 3 generates interactive environments at 720p and 20-24 frames per second, frame by frame from a text prompt and user actions.
- It keeps worlds consistent for several minutes and recalls changes from specific interactions for up to a minute, but can't sustain hours of interaction.
- Genie is grounded in Google Street View data, and "promptable world events" let users change weather or add objects mid-simulation to test how agents handle surprises.
The author fine-tuned a small 360M-parameter language model to simulate a terminal with no real shell behind it, predicting how the screen changes after each keypress. It can create files, run vim and tmux, and play snake, though it only sometimes remembers earlier commands.
- Instead of generating text token by token, the model predicts screen updates, with a COPY operation for scrolling that cuts a full redraw from roughly 3,000 tokens to about three tokens per row.
- Training on its own mistakes (a DAgger-style repair approach) raised PresetBench from 24/47 to 38/47 at int4.
- Decoding rows in parallel brought full-screen redraws from up to 12 seconds down to something usable, making near-realtime performance possible on consumer hardware.
- Memory is still unreliable: the model read back correct values in only 9 of 13 file tests, and combining commands in new ways often breaks.