Baffled Londoners have been sharing photos of a strange Christmas mural that appeared in Kingston upon Thames and instantly drew attention for all the wrong reasons. The artwork stretched nearly 100 feet and was supposed to create a warm festive backdrop for the riverside restaurants, but close up shots revealed warped faces, tangled limbs and animals that looked like they were stitched together, turning the scene into something closer to a nightmare than a holiday celebration. People on Bluesky and Reddit shared zoomed images showing distorted expressions, melted features and even a snowman with an eye on its cheek, while a second mural nearby showed the same odd errors and raised questions about how such an obviously flawed piece was approved and printed at this scale. The local council clarified that it had no involvement in the installation, saying the landlord was responsible for the display, and the mural was quietly taken down the next morning after the backlash spread across social platforms. Follow us (👉@artificialintelligenceee) for everything latest from the AI world. Source: Standard News

This week’s updates show how quickly the field is shifting, with major companies rolling out new tools, models and platforms that push deeper into coding, search, infrastructure and creative work. The pace of releases makes it clear that the industry is moving beyond single-model drops and into full product ecosystems, with companies upgrading everything from development environments to multimodal systems and collaborative features. Taken together, this week reflects a landscape where progress is accelerating on every front, and the companies that tie these pieces together fastest will shape how the next generation of tools actually gets built and used. Follow us (👉@artificialintelligenceee) for everything latest from the AI world.

Yann LeCun’s recent comments show how sharply the field is splitting between people who believe scaling existing systems is enough and those who think the current direction cannot reach the next level of intelligence. His critique is landing louder because it comes from someone who shaped core ideas that nearly every modern system relies on, giving his warnings weight inside a community that usually treats scale as the default answer. This debate is becoming a defining moment for the industry, and whichever path proves right will determine how the next generation of systems learn, reason, and interact with the world. 👉 Follow @uncover.ai to stay updated with the latest AI news Source: The Wall Street Journal

Newly leaked internal documents reviewed by TechCrunch reveal that Microsoft has already received $865.9 million from OpenAI through a 20% revenue-share deal covering the first three quarters of 2025. This payout is far lower than the multibillion-dollar revenue figures the company has often implied publicly. The documents also show that OpenAI’s inference costs reached an estimated $12.4 billion from early 2024 through Q3 2025, underscoring how expensive it is to run large AI models at scale. These costs are significantly higher than what many analysts previously assumed. The leak raises questions about the sustainability of OpenAI’s business model and whether its rapid growth is being fueled more by massive compute spending than actual profitability. It also highlights how dependent OpenAI remains on Microsoft’s infrastructure. If accurate, the numbers suggest that OpenAI’s financial reality is much more fragile than its public image. Investors and partners may now push for transparency around true revenue, margins, and long-term viability. Follow us (👉@artificialintelligenceee) for everything latest from the AI world. Source: TechCrunch

David Sacks, Greg Brockman (OpenAI president), Elon Musk, and Jensen Huang hanging out at the White House dinner on Tuesday.

Elon Musk publicly offered to help Apple upgrade Siri by integrating xAI’s Grok model, after reposting a suggestion that Apple should “fix Siri” and replace it with Grok 4.1. His reply was short but direct: “I’m down.” The proposal comes at a time when Apple is under pressure to improve Siri’s intelligence and compete with rapidly advancing AI assistants from Google, OpenAI, and now xAI. Siri’s limitations have been heavily criticized for years. Grok, built by Musk’s xAI, is designed to provide more conversational and real-time responses with access to live internet data. Musk claims future versions of Grok will outperform GPT-4, especially for reasoning and humor. There is no official partnership yet, but even a public suggestion from Musk has triggered major debate about whether Apple should outsource its core AI assistant technology. If it ever happened, it would dramatically reshape the AI ecosystem. Follow us (👉@artificialintelligenceee) for everything latest from the AI world. Source: Moneycontrol

Oracle’s massive commitment to become a core infrastructure provider for OpenAI has pushed the company into one of the most aggressive investment cycles in its history, creating a level of financial strain that markets have reacted to sharply. The scale of its data center expansion requires years of heavy spending, and the reliance on borrowing adds pressure as interest costs rise and cash generation trails far behind the pace of construction. What concerns analysts most is how closely Oracle’s future is now tied to OpenAI’s trajectory, because any slowdown in model demand or commercial momentum would leave Oracle carrying the full weight of its long term financing obligations. Investors are questioning whether today’s AI partnerships still deliver the stock market boost they did a year ago, especially as multiple tech giants have seen muted reactions to billion-dollar deals in the current environment. Follow us (👉@artificialintelligenceee) for everything latest from the AI world. Source: CNBC

Gemini 3 is quickly becoming one of the most practical AI models for real creative work as people across the internet are using it to generate games, build simulations and prototype interactive tools with a level of speed that feels entirely new. Developers are showing how the model can plan full structures, write efficient code, assemble user interfaces and animate scenes by following simple instructions while keeping track of logic, physics and spatial constraints without losing coherence. These examples also highlight how the model handles multi step processes like building 3D environments, adjusting gameplay systems, shaping animations and creating visual tools that respond instantly to user changes. The range of projects being shared online suggests that Gemini 3 is opening the door to fast experimentation by letting people sketch ideas in plain language and turn those ideas into working digital experiences almost immediately. Follow us (👉@artificialintelligenceee) for everything latest from the AI world. Source: minchoi/X Credits: skirano/X sebkrier/X

Nano Banana 2 leaked For roughly two hours, an unlimited, uncensored build of Nano Banana 2 went live on Higgsfield, and a handful of users slipped in. The outputs were shocking - raw, unfiltered, and frankly crazy capable compared to any prior public demo. The window is temporarily closed. @Higgsfield.ai says anyone who subscribes today will get the earliest, cheapest access when Nano Banana 2 launches. How did they release it before Google? #higgsfieldpartner

Microsoft and Nvidia have teamed up to invest in Anthropic, marking one of the biggest alliances in the AI industry this year. As part of the deal, Anthropic will spend $30 billion on Microsoft’s Azure cloud, tightening the competition around frontier AI models. Nvidia is expected to invest up to $10 billion, while Microsoft will put in up to $5 billion. Both companies are joining Anthropic’s next funding round, a move analysts say is designed to reduce their dependence on OpenAI. Satya Nadella said Microsoft will continue working with OpenAI, but the company does not want to rely on a single frontier model provider. The partnership also deepens Nvidia’s position in the AI ecosystem. The chipmaker will work with Anthropic on next-generation hardware, including efforts tied to its Grace Blackwell and Vera Rubin platforms. Anthropic has committed up to 1 gigawatt of compute using Nvidia systems. Industry executives estimate this setup could cost more than $20 billion. Anthropic is now one of the fastest-growing AI companies, valued at around $183 billion and powered by strong enterprise demand. The company expects its revenue run rate to climb to nearly $26 billion next year, supported by more than 300,000 business customers. Follow us (👉@artificialintelligenceee) for everything latest from the AI world. Source: CNBC