AI: the most important news of the week (November 24-30)

Gemini 3 breaks every record, Claude 4.5 dominates coding, and Big Tech moves $100 billion in a week. Meanwhile, autonomous agents arrive and top researchers fl

Every Monday we select and analyze the 5 most significant news stories from the world of artificial intelligence. Not just a simple summary, but a critical reading of the developments that are truly changing the industry. No hype, no unnecessary technicalities.

Why 5 news stories? Because they are enough to stay updated without being overwhelmed by information.

1. Gemini 3: Google Rewrites the Rules of the Game

Google has launched Gemini 3, the first model to surpass the 1500 Elo mark in standardized benchmarks.

🔍 What happened: The new model from Mountain View excels particularly in coding and application development, also integrating advanced image generation capabilities. Gemini 3 is already being implemented in Google Search (AI Mode) and the Gemini app.

💡 Why it's important:
This isn't just another more powerful model. Surpassing the 1500 Elo threshold represents a qualitative leap in complex reasoning capability. But the real news is the immediate integration into consumer products: Google is transforming search from a passive tool into an active conversation. This radically changes the way billions of people will access information in the coming months.

🎯 Our take: Google has learned its lesson: it's not enough to have the best technology, you have to put it in users' hands before the competitors. Deployment speed is the new metric that matters.

Source: LinkedIn – PA Media Weekly Digest

Also read: Quantum Computers and AI: The Next Technological Revolution and ChatGPT 4.5 and the Turing Test: When AI Makes Us Doubt

2. Record Investments: Over $100 Billion in One Week

Microsoft and Nvidia have invested $15 billion in Anthropic, bringing the startup's valuation over $300 billion. OpenAI has signed a $38 billion partnership with AWS.

🔍 What happened: In parallel, Nvidia recorded Q3 revenue of $57 billion (+62% year over year), while Anthropic purchased $30 billion in computational capacity from Microsoft data centers. Dell also raised its AI server forecasts, while HP announced massive personnel cuts.

💡 Why it's important:
These numbers mark a paradigm shift: AI is no longer a technological experiment but a mature industry moving capital comparable to sectors like automotive or pharmaceuticals. The "multi-cloud" era has officially begun, with big tech betting on distributed infrastructure instead of vertical monopolies. But there is a dark side: HP's cuts show that AI creates new winners but also new losers.

🎯 Our take: When you see investments on this scale, you understand we're not talking about hype but a deep industrial transformation. Nvidia's CEO denies the "AI bubble," and looking at these numbers it's hard to disagree. But beware: concentrating so much capital in a few companies creates systemic risks we don't yet fully understand.

Sources: LinkedIn – PA Media, HumAI Blog

Also read: Predictive Economics: If AI Could Anticipate a Financial Crisis and Algorithmic Micro-Financing

3. Claude Sonnet 4.5: The Undisputed King of Coding

Anthropic has released Claude Sonnet 4.5, which dominates programming benchmarks with a 54% improvement on complex tasks thanks to "extended thinking" and integrated tool use.

🔍 What happened: The new model introduces extended reasoning capabilities that allow it to tackle multi-step coding problems with unprecedented precision. Tool use functions are now native, no longer external add-ons. Cognizant has already begun deploying Claude at enterprise scale for complex code refactoring.

💡 Why it's important:
While everyone was watching OpenAI and Google, Anthropic has quietly built the best assistant for developers. The +54% is not a marginal increase: it means Claude can now autonomously handle complex code refactoring and advanced debugging. We are witnessing the birth of a new category: AI as a senior developer colleague, not just as intelligent autocomplete.

🎯 Our take: The model battle is shifting from generic capabilities to vertical specialization. Claude is winning in coding, other models will dominate in other areas. The era of the "perfect universal model" is over before it even began.

Source: HumAI Blog

Also read: AI Tools for Freelancers: 5 Tools That Save You Hours Every Week and How to Automate Your Daily Workflow

4. The Era of Autonomous AI Agents Has Arrived

Google has launched Anti-Gravity, Microsoft has announced Agent 365: the week saw the simultaneous launch of "agentic AI" systems that can execute complex tasks without continuous supervision.

🔍 What happened: These are not improved chatbots but systems that can plan, execute, and verify chains of actions autonomously. Samsung presented a tiny model that beats giant LLMs on specific reasoning tasks. The first cyberattack orchestrated by autonomous AI also emerged.

💡 Why it's important:
"Agentic AI" represents the shift from AI as a command tool to systems that operate with a degree of real autonomy. When Google and Microsoft launch similar products in the same week, it's not a coincidence: we are entering a new phase. But the announcement of the first cyberattack orchestrated by AI should make us reflect: more autonomy also means new risk vectors that we still don't know how to manage completely.

🎯 Our take: AI agents will truly transform office work, but not in the way we thought. They won't "replace" entire roles, but will completely redesign workflows and responsibilities. Companies that prepare today will have an unbridgeable competitive advantage.

Sources: HumAI Blog, Radical Data Science

Also read: AI and Cybersecurity: Protecting Digital Systems from the Future and Future War? The Unsettling Shadow of Smart Weapons

5. Brain Drain: 20+ Top Researchers Leave Big Tech for Startups

Over 20 senior researchers left OpenAI, Meta, and Google this week to found or join emerging AI startups. Meanwhile, OpenAI launched Sora 2 with synchronized audio.

🔍 What happened: The exodus includes heavyweight names who contributed to the most significant breakthroughs in recent years. In parallel, OpenAI released the second generation of Sora, now with synchronized audio-video capabilities that look like they came from a professional production studio.

💡 Why it's important:
When the best talents leave big tech in the midst of the AI boom, it means two things: first, they believe real innovation now happens in startups, not in giants; second, they've seen enough from the inside to know where they don't want to be. The launch of Sora 2 shows that OpenAI maintains leadership in generative video, but the loss of key talent could slow down the next breakthroughs. The history of technology teaches us that real revolutions are often born in garages, not on corporate campuses.

🎯 Our take: 2026 will be the year of AI startups "born" from ex-big tech. We are witnessing the formation of the next generation of unicorns. It's a shame that the concentration of capital (see news #2) makes it increasingly difficult for small realities to compete without mega-funding.

Sources: HumAI Blog

Also read: Artificial intelligence and continuous learning: learning at any age and AI and digital skills: what to learn to not fall behind

📊 What these developments really tell us

Let's stop for a moment. Let's breathe. And let's try to look beyond the numbers and the bombastic announcements.

This week tells us a story that goes far beyond the $100 billion in investments or the technological benchmarks. It speaks of a world that is changing faster than we can metabolize, and of choices that will define our relationship with technology for the coming decades.

Let's start with an uncomfortable question: what does it really mean when Microsoft and Nvidia invest $15 billion in a startup? It's not just a matter of money. It's that we are witnessing the birth of a new kind of power: control over computational infrastructure. Whoever owns the data centers that run AI, whoever has access to that computing power, becomes a sort of new gatekeeper of innovation. And this, let's admit it, should make us reflect on who really controls the future of artificial intelligence.

But there is another side of the coin, a more human one. When you see 20+ top researchers leave big tech at the same time, you can't help but wonder: what have they seen? What have they understood that pushed them to leave right now, in the golden moment of AI? Perhaps they understood that real innovation is not born in corporate meetings, but in the garage, in the basement, in spaces where there is still room to dream without having to justify every decision to a board.

And then there's the elephant in the room: autonomous agents. When we read that the first cyberattack orchestrated by AI has occurred, the temptation is to think "it was inevitable." But let's stop for a second. Are we really prepared for a world where machines can act autonomously? Where the difference between a human action and a machine action becomes blurred? It's not science fiction, it's last week.

What strikes me the most, however, is the speed. Google launches Gemini 3 and immediately integrates it into the search we use every day. There is no longer time for "beta launch," for "testing period." Technology arrives directly in our lives, ready or not. And this forces us to a difficult question: are we really choosing how we want to coexist with AI, or are we simply accepting what we are given?

There is, however, also a note of hope in all this. Claude Sonnet 4.5 is not just a more powerful model, it is an example of how AI can amplify our abilities instead of replacing them. A developer working with Claude is not replaced, they become more effective. It's a subtle but fundamental difference, that between replacement and amplification.

And while the big players divide up billions, small startups continue to be born, fueled by the brain drain from big tech. This tells us something important: innovation cannot be bought, it can only be created. And it is created better when you have the freedom to fail, to experiment, to follow intuitions that an Excel spreadsheet could never justify.

The real question is not "will AI steal our jobs?" or "will AI save us?". The real question is: what kind of relationship do we want to build with this technology? Do we want to be passive users who suffer the decisions of a few tech giants, or do we want to be active participants who shape the future?

This week has shown us that we are at a crossroads. The $100 billion invested, the autonomous agents, the talent exodus: they are not isolated events, they are symptoms of a deep transformation. And like all great transformations, it requires us to make choices.

The good news? We are still in time to make them. But time is running, and fast.

💬 Your opinion matters: Which of these news stories do you think will have the greatest impact in the coming months? Are we really at the beginning of the era of autonomous agents or is it still early? Do you feel prepared for a world where AI acts autonomously? Write to us or share on social media.

📧 Don't miss the next AI News: Subscribe to the La Bussola dell'IA newsletter to receive every Monday the analysis of the most important news on artificial intelligence, without hype and without unnecessary technicalities.

This article is part of the weekly "AI News" column by La Bussola dell'IA. For in-depth analysis and detailed insights on these topics, visit our website www.labussoladellia.com