The week, in one sitting

Must know

1. GPT-6 arrives in ChatGPT, and it can draw the screenModels · A chat box that was a transcript can now be an interface the model writes, on free accounts as well as paid.
2. EmbeddingGemma 2 searches text, images, audio, and video on the deviceModels · Private photos, voice memos, and code can be searched without uploading them.

Worth watching

1. Odyssey-3 is a world model you can act insideAgents & Automation · A learned world that can train a robot policy would matter. That claim is the lab’s, and this is not a robot you can deploy.

Also happened

1. Carbon-A proposes 566 million gene candidates across 22,617 speciesResearch · Most software teams can note it and move on. The number is a pile of candidates, not confirmed genes.

The interface is part of the model

GPT-6’s Intelligent UI treats layout, controls, and small tools as something the model writes. A version number matters less than whether people start doing the task inside that generated screen.

Retrieval is leaving the data center

EmbeddingGemma 2 is small enough to embed a private photo library, a voice memo, or a codebase on the device that holds them. That changes the default for search over sensitive material.

Agents need a world to practice in

Odyssey-3 predicts how a scene changes when something acts. The claim to watch is that a learned world can train policies for arms, humanoids, and cars. A convincing video is not that claim.

Specialist models are writing into the reference

Carbon-A is a defined prediction job, not a general assistant. Hugging Face ran it across GenBank and published coordinates and confidence scores. Most software teams can note it and move on.