Good morning,

The AI revolution is in full swing.

Every single company - from the smallest to the largest is starting to use it in some way. It’s fascinating. I am seeing people “waking up” left and right.

BUT. Let me spoil the party.

They might actually be waking up inside of a nightmare.

I have to be honest: I am deeply worried about where things are going. And I’m not talking about the increasing costs these AI labs are already pushing on us. That might deserve its own article.

Because that is also something worth noting. It’s trick number one in the arsenal of any drug dealer. First you give the product for free - then you make someone completely dependent on it, then you increase the price.

ANYWAY …

I’m here to completely ruin your day in another way.

I’ve spent the last few years staring into the guts of the machine for my research, and the view is increasingly predatory. We are witnessing the quiet installation of a new world order. One where the "invisible hand" of the market has been replaced by the iron grip of the algorithm, and I'm not sure people realize how little time we have left to grab the wheel.

I have bad news, and I have worse news. The bad news is that we are currently being drafted into the digital serfdom of Technofeudalism. The worse news? This outcome is "overdetermined"-a fancy way of saying the deck is stacked, the house always wins, and the gravity of greed is pulling us into a black hole whether we like it or not.

To the uninitiated, Technofeudalism is the death of the free market as we know it. We’ve moved past the era where companies simply sold you a widget for a profit. We are now living in a world of digital fiefdoms. Big Tech "Lords" own the platforms-the digital land-and we, the users, are the serfs. We don't own our tools, our data, or even our digital identities; we merely rent them, paying a "tax" of attention and information to the cloud-based masters who own the servers.

Believe it or not, but I carry an actual Masters diploma in History (this isn’t relevant but I like to inflate my ego once in a while).

History is a meat grinder with a short memory. In ancient China, peasants were fed into the corvée labor system like logs into a furnace to build the Great Wall and the Grand Canal. They died in droves because the fertile river valleys produced an annual crop of humans-a renewable resource for warlords to spend.

But even those tyrants needed those humans. Their power was built on the muscle of the masses.

Under a fully automated economy-where robotic legionnaires hold the perimeter and AI agents balance the ledgers-your physical body is no longer a resource. It is a net negative. You are a mouth to feed that offers no mechanical ROI.

In online circles, they call this the "permanent underclass." Yuval Noah Harari calls it the "useless class." But even those terms are too polite. They suggest you still have a "class," which implies a civic standing.

In the eyes of the silicon gods, you aren't a citizen. You are becoming Redundant Biomass.

This isn't a conspiracy theory; it’s a structural reality driven by three unstoppable forces:

We live under the Enlightenment delusion that "natural rights" are inherent. They aren't. Historically, rights are legal fictions written in the afterglow of bloodshed. They are concessions extracted through coercive force. The American Revolution wasn't a polite debate; it was a violent withdrawal of consent.

But how do you extract concessions when you have no leverage? How do you strike when a GPT-7 instance does your job for the cost of a lightbulb?

The darkest possibility is a "self-sealed system." The elites secure every step of production-from the mines to the farms-with AI and robotic security. They monitor us with automated surveillance and simply wall us off. They won't need to "oppress" us; they will simply ignore us.

Right now, the owners of capital still need our labor, which means we still have a collective veto. We can still throw a wrench in the gears. But we are losing that veto power inch by inch. The window to demand a seat in the future is closing, and the "Lords" are already starting to change the locks.

Welcome to the Blacklynx Brief

Google I/O 2026: Gemini 3.5 Family, Omni, Spark, and the Biggest Search Overhaul in Years
Google unveiled a sweeping AI wave at I/O: Gemini Omni handles text, image, audio, and video inputs to produce video output, while Gemini 3.5 Flash matches competing frontier models at roughly four times the speed. Gemini Spark debuts as a 24/7 personal agent running on Google Cloud VMs, and Search received its most significant redesign in a generation, adding cross-modal inputs and a generative UI layer. (Google Blog)

Jury Unanimously Dismisses Musk's $100B+ Lawsuit Against OpenAI
After a three-week trial, a jury ruled that Elon Musk's lawsuit against OpenAI, Sam Altman, Greg Brockman, and Microsoft was filed years too late, dismissing all claims on statute-of-limitations grounds. The Microsoft claims were separately dismissed, and Musk announced plans to appeal. (CNN)

Google Confirms World's First AI-Authored Zero-Day Exploit Deployed in the Wild
Google's Threat Intelligence Group disclosed that on May 11, 2026, an AI-generated exploit was used in a live attack -- a Python script that bypassed two-factor authentication on a widely used open-source admin tool. The malware was identified as AI-authored in part because it contained a hallucinated CVSS score, a fingerprint no human attacker would leave behind. (Google Cloud Blog)

Andrej Karpathy Joins Anthropic to Lead AI Research Automation
OpenAI co-founder Andrej Karpathy has left to join Anthropic, where he will head an internal group focused on automating AI research itself, working under pre-training lead Nick Joseph. The move is widely seen as a significant talent gain for Anthropic and a notable loss for OpenAI. (Andrej Karpathy on X)

Microsoft's 100-Agent Security System Tops UC Berkeley Cybersecurity Benchmark
Microsoft's Multi-Model Agentic Scanning Harness scored 88.45% on UC Berkeley's CyberGym benchmark, ahead of Claude Mythos at 83.1% and GPT-5.5 at 81.8%, by coordinating more than 100 specialized AI agents in a staged pipeline. The system has already found 16 Windows vulnerabilities in production, including four critical remote code execution flaws. (Microsoft Security Blog)

Cursor Composer 2.5 Delivers Near-Frontier Coding at a Fraction of the Price
Cursor released Composer 2.5, built on Moonshot's Kimi K2.5, with benchmark performance approaching Claude 4.7 and GPT-5.5 on coding tasks at under $1 per average CursorBench task, compared with up to $11 for equivalent OpenAI model runs. The company also confirmed it is training a proprietary SpaceXAI-backed model using ten times its current compute. (Cursor Blog)

Meta Begins Largest Layoffs Since 2023 in Zuckerberg's AI-First Push
Meta started cutting up to 8,000 employees this week and simultaneously pulled listings for roughly 6,000 open roles, marking the company's deepest workforce reduction in three years. The restructuring reflects Zuckerberg's explicit shift to an AI-first organizational model, redirecting headcount budget toward infrastructure and model development. (CNBC)

That’s it for us this week. Please like and subscribe 🙂