Tuesday, January 13, 2026

AI Diaries: This Week in the World of Artificial Intelligence (January 12, 2026)

AI Diaries: This Week in the World of Artificial Intelligence (January 12, 2026)

Artificial intelligence technologies, which have recently become much more widely used in daily life, continue to develop rapidly. Almost every day, remarkable advancements occur in this field, and significant milestones are being reached one by one. In our "AI Diaries" series, we continue to regularly document the development of this technological revolution, which has the potential to significantly change life worldwide in the coming period.

What Happened This Week in the World of Artificial Intelligence?

CES 2026, held last week in Las Vegas, also shaped the world of artificial intelligence this week. Companies like Nvidia, Razer, and LG showcased their latest AI-powered products at CES 2026. Parallel to these, as every week, new artificial intelligence tools entered our lives this week. Here are the prominent developments in the world of artificial intelligence this week:

Restrictions Imposed on Grok's Controversial Image Generation Feature

Elon Musk's AI company, xAI, made another highly controversial decision last week, allowing Grok to denude photos of real people. Due to this highly exploitable feature, thousands of people's photos were denuded without their permission on X in a short time. Naturally, this drew widespread reactions. Even a couple of countries, such as the UK and India, announced they might initiate legal proceedings.

xAI eventually backed down, and X imposed restrictions on this controversial feature. Now, only users with a paid X subscription can use Grok's image generation and editing feature. Of course, this applies to X. As you know, Grok also has a separate application. Users can continue to generate and edit images as they wish in this application. It remains unclear whether this is a permanent change or a temporary solution.

The Planet's Most Powerful AI Platform Unveiled: Nvidia Rubin

Nvidia unveiled its Rubin platform, developed for AI-focused data centers, and its new powering hardware at CES 2026. Offering up to 5 times performance increase compared to the company's existing Blackwell architecture, Rubin is positioned as Nvidia's most comprehensive and powerful AI platform developed to date.

The Rubin platform consists of a total of six separate chips that have come out of production and are currently in the testing phase at Nvidia laboratories. These components include the Rubin GPU, Vera CPU, NVLink 6 Switch, ConnectX-9 and BlueField-4 networking solutions, and the Spectrum-X 102.4T CPO silicon photonics infrastructure. The Rubin GPU comes with a dual-die design, each with independent processing and tensor cores. This GPU, specifically developed for AI workloads, offers 50 PFLOPs inference and 35 PFLOPs training performance in NVFP4 format. These values represent a 5-fold and 3.5-fold increase, respectively, compared to Blackwell. In terms of memory, the Rubin GPU transitions to HBM4 technology. This architecture provides 22 TB/s memory bandwidth per chip, a 2.8-fold increase over Blackwell. Additionally, 3.6 TB/s NVLink bandwidth is offered for each CPU, which is double that of the previous generation.

Humanoid Robots Dominated CES 2026

This year at CES was the year of humanoid robots. With Western companies on one side and their Chinese rivals on the other, numerous humanoid robots were showcased at CES this year. The most striking robot displayed at CES this year was Atlas, developed by Boston Dynamics. Atlas, which has been making headlines with its prototypes for years, was another humanoid robot exhibited at CES 2026. The production-ready version of Atlas was showcased at CES, while it was announced that Hyundai will produce thousands of these robots in the coming years.