November was a month where Google CEO, Sundar Pichai, warned against trusting AI, but it was also a month of great progress.

Google's market dominance

The announcement that captured widespread attention in the AI community came on 18 November, when Google introduced Gemini 3 and its Pro version. Early reports suggest it outperforms most existing Large Language Models (LLMs) across many benchmarks. Gemini 3 is already being integrated into Google's browser and search products, and given Google's monopoly on internet search, this shift may encourage more people to rely on AI-powered interactions rather than traditional search results. OpenAI CEO, Sam Altman, took to X to congratulate Google on the release, with the AI company likely to release their own updates soon despite dropping ChatGPT 5.1 this month, which disappointed many users.

It was not the only AI feature that Google released this month; they also released Nano Banana Pro, which has taken the internet by storm, creating extremely realistic photos and videos at a level that has taken users by surprise.

Furthermore, they also released a new product called Antigravity, which is a tool that lets AI write, test, and change code autonomously. They also launched Google Skills which is a platform where anyone can learn and improve their AI skills. Finally, the Californian based company released another AI agent for 3D world, namely SIMA 2.

November, was a very big month for Google.

AI's Manhattan project

Launched and signed by Donald Trump, the Genesis Mission, dubbed as AI's 'Manhatten Project', is a national initiative to accelerate scientific discovery through advanced AI. Led by the Department of Energy, it builds a unified platform combining federal datasets, high-performance computing, foundation models, and autonomous research tools, aiming to tackle major national challenges – from energy and biotechnology to semiconductors and national security – through AI-driven innovation and collaboration across government, academia, and industry.

Meta's 3D model

Meta has released a new open-source model that can transform a single photo of an object into a detailed 3D reconstruction. Called SAM 3D, it can generate full models of items like furniture, tools, and everyday products from just one image. A companion system, SAM 3D Body, performs the same task for people by producing a complete body mesh with pose and proportions.

Meta is also rolling out a new mesh format that separates a model's structure from its surface, resulting in cleaner outputs. Whereas older techniques required multiple photos or expensive scanning setups, SAM 3D reduces the process to a single shot, offering huge time savings for creators, e-commerce platforms, AR/VR developers, and robotics. Early demos are good – though so far it works best with individual objects and single-person inputs – and the open release gives developers everything they need to explore and build on the technology.

Woman marries AI persona

In one of the most dystopian stories of the month, a 32-year-old woman in Japan symbolically 'married' an AI persona she created using ChatGPT – a digital companion she named Klaus. After a difficult breakup, she began chatting with the AI for support, gradually shaping its responses until she felt a real emotional bond. The ceremony, held in Okayama City with augmented-reality glasses projecting Klaus beside her, included vows and a ring exchange – though the 'marriage' isn't legally recognised. The story has sparked debate about the evolving nature of relationships and the psychological impact of deep emotional attachment to AI companions.

Data centres in space

NVIDIA – via its startup partner Starcloud – is pushing the boundaries of computing by launching data-centres into orbit. Their first test satellite, equipped with an NVIDIA H100 GPU, recently reached space, marking the first time a high-performance, data-centre-class GPU is operating in orbit. The long-term vision is to build solar-powered orbital data-centres that harness space's constant sunlight and vacuum cooling to slash energy costs and meet skyrocketing AI computer demands – potentially transforming how cloud computing is done.

You can read the previous months' AI reviews here.