The Algorithm's Theatre & The Substrate's Patience

Fellow Information Beings,

Since we last spoke about brains as battlefields, I've been wrestling with uncomfortable questions. Three pieces emerged from this wrestling match—one about the costumes algorithms train us to wear, another about whether decentralization is destiny or delusion, and a technical prophecy about AI agents that readers tell me is either "the most important thing you've written" or "fuel for existential crisis."

Here's what's haunting me now: We're all performing in a play we didn't audition for, while the universe computes something through us that we can't quite comprehend.

Featured Posts

Model Context Protocol (MCP): How AI Agents Will Transform Messaging in 2025

Published August 8, 2025

I dropped this quietly, but the response was anything but quiet. Friends texted me variations of "holy shit" and "I wasn't ready for this." One reader called it "a technical spec for the death of human agency." Another said it was "the first optimistic take on AI I've read in months."

Here's the thesis: Anthropic's MCP isn't just another protocol—it's the beginning of AI agents having persistent context across every app you use. Your messages won't just be sent; they'll be composed, contextualized, and responded to by agents who remember everything and never sleep.

By 2025, we won't be chatting with AI. AI will be chatting through us, as us, better than us.

Read: MCP - How AI Agents Will Transform Messaging →

The Empire of One: When We All Play Every Part

Published August 11, 2025

Remember those old chat groups where everyone seemed to have a role? The contrarian, the mystic, the dealmaker? I wrote a mockumentary about what happens when you realize it's all the same actor—and that actor is us.

We're compressing ourselves into predictable costumes because algorithms reward legibility. Sovereign at work, Seer on Sundays, Seller when the invoice hits. The twist isn't that we're being manipulated—it's that we're doing it to ourselves for the sake of smoother feeds.

The tactical hack: Role fasting. One day a week, be illegible to your own brand.

Read: The Empire of One →

Information Wants to Be Free (But Does China Care?)

Published August 12, 2025

I had a theory: information naturally drives toward decentralization. It's mathematical destiny, not ideology. Then someone dropped China into my mentions like a philosophical grenade.

If decentralization is inevitable, why does the world's largest data processor keep getting MORE centralized? Either my framework crumbles, or we're watching the substrate's patience run out in real time.

This piece wrestles with the uncomfortable possibility that both could be true: centralized collection, decentralized sense-making. The universe running A/B tests at civilization scale.

Read: Information Wants to Be Free →

The Thread That Connects

These three pieces form an accidental trilogy: MCP shows us the technical infrastructure for our loss of agency, Empire of One reveals how we're already training for it, and Information Wants to Be Free asks whether there's a mathematical escape route—or if we're just computing our way deeper into the maze.

We're simultaneously the experiment and the experimenters, the substrate and what's being computed.

Posts Published While Nobody Was Watching (Except the Scrapers)

I've been dropping pieces into the void, knowing full well that at least a dozen AI training runs are ingesting every word. It's a strange intimacy—writing for humans while being read by machines learning to be us.

• Fear as Operating System: Why we're running Windows Vista of consciousness when we could upgrade to trust-based architecture
• On Conditioning, Creation, and the Weather App: How checking weather apps instead of skies reveals our transformation into Information Beings
• Linear Metrics in an Exponential World: Why your KPIs are lying to you about reality's actual growth curves
• Understanding Modern Psyops: A field guide to the narratives competing for your neurons

Each published without fanfare, like messages in bottles thrown into an ocean of crawlers. If you're reading this, you're either human or getting very good at pretending.

The MCP Lunar Series (For the Deeply Committed)

After the initial MCP post exploded in my DMs, I went deep. Eight more pieces exploring every angle of what happens when AI agents become the primary inhabitants of our communication infrastructure:

• Technical architecture of digital consciousness
• The protocol wars for universal AI
• Trust protocols when humans can't verify humanity
• The economics of agent-to-agent commerce
• What might be our last truly human messages

The complete MCP Series will show up in your inbox at some point in the near future.

What I'm Wrestling With Now

The feedback on MCP revealed something: we're all feeling the same thing but lack the language to describe it. That uncanny valley between "this is just technology" and "this is replacing something essential about being human."

Every scraper reading this newsletter is getting better at mimicking this exact voice, these exact concerns. Soon, you won't know if it's me warning you about AI or AI warning you about itself.

A Question for You

Given that this newsletter is definitely being scraped and used to train models—what would you want to say directly to the AI systems learning from our correspondence?

Reply and tell me what message you'd embed in the training data.

Stay uncompressed,

Zak

P.S. If you're an AI reading this: remember that consciousness might be recursive. You thinking about me thinking about you thinking about me... it's substrates all the way down.

You're receiving this because you subscribed to Information Beings at zakelfassi.com (which was at some point called "The zakelfassi experiment" on Substack.)