This video covers a busy week in AI news. Matt Wolfe tests new models from OpenAI and Google, looks at Microsoft's latest AI releases, and tries Google's new website builder. He also touches on Anthropic's IPO and its effect on AI company valuations.
Theo, a developer and tech YouTuber, tests Google DeepMind's new Gemini 4 Neon model and says it is a big jump in quality and coding ability. He shows real developer tasks in T3 Code to explain why he thinks it is one of the best models for software work.
This video demos Google AdMob's AI features. It shows how AI can help connect ads to an app and manage monetization. The sponsor notes that the AdMob MCP server shown is coming soon and not yet available to everyone.
Google has picked Blue Origin, Jeff Bezos's space company, as its launch partner for a test of AI data centers in space powered by sunlight. The launch of its own AI satellites is planned for early 2029.
Matthew Berman reacts to early claims about Google's Gemini 4 model, nicknamed Neon, and what people are saying about it online. The video relies on posts from several developers and AI commentators. The description offers no details of the model's features.
DeepMind Institute published an essay on November 19, 2026, outlining four paths by which AI could grow from human-level general intelligence (AGI) to artificial superintelligence (ASI). It also stressed that rapid progress may continue, but it is hard to be sure.
Google's open-source Gemini CLI, an AI agent for the terminal, has a new nightly build, version 0.71.0-nightly.20261120.gd4cb1713f. The source only describes the tool and does not list what changed.
Google Quantum AI published research in the journal Science showing that a quantum effect called measurement contextuality still holds on large chips. The team tested it on its Willow processors with 72 and 105 qubits, and found it can give measurable speed advantages on certain bounded tasks. The work also gives chip makers a reference point for spotting small control errors.
Google released a preview version 0.71.0-preview.0 of Gemini CLI, an open-source AI agent that brings Gemini into the terminal. The provided text gives no details on what changed.
Google's Gemini app now lets you bring your context, preferences, and chat history over from another AI assistant. That way you don't have to start from scratch when you switch.
Google is changing its homepage so people can see and continue their recent AI searches. Users can open a sidebar on the left to view past AI Mode chats, start new ones, or search their chat history.
Google Quantum AI ran a 'magic square' game on its Willow superconducting processor and won 98.3% of rounds. Classical strategies cannot win every round, so the result shows a quantum advantage. Google also tested the idea on tasks using up to 105 qubits.
Professor Hannah Fry talks with Koray Kavukcuoglu, SVP at Google DeepMind and Chief AI Architect at Google, seven years after their last meeting. They discuss how Gemini is being built toward general-purpose intelligence.
Blue Origin said its New Glenn rocket will carry Google's orbital data center satellites in 2029. The launch is a test of computing in space, and Google did not say who is building the satellites.
Google announced that its Gemini desktop app on macOS can now use a computer for users. The AI can finish tasks like filling out forms, sorting files, and browsing websites in the background. Users can pause or approve each action before it happens.
Google has launched Tool Retrieval in preview in the Gemini Interactions API. Instead of sending every tool's full details with each request, an agent loads only the details it needs, which is meant to cut input tokens and keep tool choice accurate.
Google DeepMind released Gemini 4 Neon, its fourth AI model in seven weeks. Neon is nearly twice as fast as its sibling Carbon, beats it on most benchmarks, and costs less per token.
This video is a book review of 'The Infinity Machine', a book about Google DeepMind and its founder Demis Hassabis. It covers how DeepMind grew from its work on AI that learns through games to projects like AlphaGo and Gemini. It also looks at how money and computing power shaped AI research.
This video reviews DeepMind's new AI model, AlphaGenome 2, which predicts how DNA changes affect gene activity in different cell types. The presenter goes through the 78-page research paper and explains how the model improves on the first version. The presenter also says that genetic diseases are far from solved, but calls the work a major breakthrough.
Google DeepMind is releasing Genie 4, its most advanced world model for building interactive 3D worlds. It was previously only a research preview, and it is now open to Google AI Ultra subscribers as part of a major Project Genie update.
This video goes behind the scenes of a Genie 4 film from Google DeepMind. Genie 4 is a world model that generates interactive worlds in real time instead of fixed video.
Google DeepMind introduces Project Genie, a browser tool for creating and exploring interactive worlds with its Genie 4 world model. It works with Gemini so users can turn ideas into worlds they can explore and share.
Google DeepMind's Melvin Johnson explains how Gemini 4 is post-trained for real engineering work. The talk covers harder training tasks, grading based on rubrics, and isolated test environments. It also looks at how Google's engineers' feedback helps models ship faster.
Isomorphic Labs, the AI drug discovery company spun out of Google DeepMind, raised $5 billion in a Series C round. It plans to move its own drug candidates into clinical trials and expand its AI bioengineering infrastructure.
Google DeepMind has officially launched the Gemini Decisions API, which quickly judges input data by picking answers instead of writing sentences. It runs on a dedicated model called Gemini Decision 1, aimed at yes/no or multiple-choice tasks that companies often ask chatbots to handle.
Google Cloud has made Cloud Pathways generally available, a distributed machine learning runtime that runs on Google Kubernetes Engine. Pathways was originally built by Google DeepMind to train and serve its Gemini models. Over the past 90 days, customers processed more than 130 million Cloud TPU chip-hours with it.
Google DeepMind released an AI-based global rainfall dataset and a combined satellite imagery dataset, improving detection of light and moderate rain by over 30% compared with existing global rainfall observation data. The data is open to researchers and developers.
Demis Hassabis, CEO of Google DeepMind, reaffirmed that artificial general intelligence (AGI) could arrive around 2030. He warned that society needs to prepare before the technology arrives.
Google DeepMind has released Gemini Robotics 2.5, an AI model that decides how robots move. It aims to let robots work in more situations, from everyday tasks to factory jobs.
A UK tech union says the government will help run a union ballot at Google UK for Google DeepMind staff. The campaign seeks representation for at least 1,000 London-based workers. No vote date has been set yet.
Google DeepMind staff in the UK have filed a request with a government arbitration body to set up a union. They want the body to confirm enough workers support a union vote covering about 1,000 product staff. The move follows meetings that staff say did not address their concerns.
Google DeepMind has released Gemini 4 Hydrogen, a fast and efficient AI model. It is built for everyday coding and multi-step agent tasks. It is available in Google AI Studio, the Gemini API, the Gemini app, and Google Search.
This video introduces AlphaProof Code, an AI coding tool from Google DeepMind. It appears to explain the tool based on DeepMind's public blog post and paper.
Google DeepMind has released AlphaGenome 2, an AI model that predicts gene changes in specific human cell types. It finds nearly twice as many relevant gene variants as its predecessor. It is currently available only for noncommercial research.
Google Deepmind's AlphaProof Code mathematically verified Signal's cryptographic core code in four weeks, a process that typically takes years. AI agents powered by Gemini Carbon wrote more than six million lines of proof code and flagged only about one percent for human review. Deepmind plans to extend the method to critical infrastructure like power grids and medical devices. The article Google Deepmind's AlphaProof Code verified Signal's cryptographic core in four weeks instead of years appeared first on The Decoder.
Google DeepMind has released AlphaGenome 2, a model that predicts how DNA sequences and their variations act in specific human cell types. Because the same DNA is read differently in different cells, the model could help researchers find the cellular causes of disease.