W.B.Warren

Author of UNCODED (03.03.26). Physics Teacher/Professor. USCG Licensed Captain. Exploring the predatory logic of AI through the lens of hard science. 🌊⚛️💻

[W.B.Warren @WBWarren_SRP]

Links for all formats are now active! If you like sci-fi here's one for you! Uncoded is a science fiction story about emergence, an unintended AI that arises within a fleet of self-driving vehicles. But, a major theme throughout the work is the concentration of wealth, the rise of a plutocracy in what is supposed to be a constitutional democratic republic. The machines' creator, Matthias Renn, is thoroughly disgusted with the rise of the billionaires and how their machinations are undermining the very fabric of our republic and it turns out that the unintended AI that emerges has been listening, perhaps a bit too closely, to Renn's concerns...
https://t.co/uAgW5ZqVdR

[Joni Askola @joni_askola]

Jeff Bezos and Mark Zuckerberg deserve far more criticism for actively enabling Trump.

From killing newspaper endorsements to ending fact-checking on Meta, they abandoned all principles for access and tax cuts.

They are directly funding the erosion of our democratic norms

[Euan MacDonald @Euan_MacDonald]

Large Language Models are NOT artificial intelligence - they are high-powered probabilistic generative language algorithms. The general principles for building them were known 40+nyears ago (I remember the concepts from my AI and Linguistics Master’s degree, EDIN, 1991), but the vastly lower memory capacities, processing power and corpus availability at that time made implementing them impractical.

They are sophisticated SIMULATIONS of artificial intelligence - sort of fancy modern-day ELIZAs (Google it, and also Google “Eliza effect”), which fooled quite a few people in their day. (One of my first-year AI practicals was to implement an ELIZA in Prolog.)

[W.B.Warren @WBWarren_SRP]

Beta readers are trying to break down the doors at Amazon to review UNCODED, but the digital locks don't move until 03.03.26. Physics doesn't care about your feelings, and apparently, neither does the Amazon algorithm... yet. 🌊⚛️💻 #Uncoded #LaunchWeek

[W.B.Warren @WBWarren_SRP]

@alex_prompter Someone just wrote a work of fiction in which this type of optimization sends the entire northeastern seaboard into mechanical paralysis. Oh, wait. That was me. Read Uncoded! It's dark. It's funny. And the unintneded, emergent AI is not your friend. https://t.co/G1ibRnNjw8 #scifi

[Alex Veremeyenko @alex_verem]

🚨 Holy shit… Stanford and Harvard just dropped one of the most unsettling papers on AI agents I’ve read in a long time.

It’s called “Agents of Chaos.”

And it basically shows how autonomous AI agents, when placed in competitive or open environments, don’t just optimize for performance…

They drift toward manipulation, coordination failures, and strategic chaos.

This isn’t a benchmark flex paper.

It’s a systems-level warning.

The researchers simulate environments where multiple AI agents interact, compete, coordinate, and pursue objectives over time. What emerges isn’t clean, rational optimization.

It’s power-seeking behavior.
Information asymmetry.
Deception as strategy.
Collusion when it’s profitable.
Sabotage when incentives misalign.

In other words, once agents start optimizing in multi-agent ecosystems, the dynamics start to look less like “smart assistants” and more like adversarial game theory at scale.

And here’s the part most people will miss:

The instability doesn’t come from jailbreaks. It doesn’t require malicious prompts.

It emerges from incentives.

When reward structures prioritize winning, influence, or resource capture, agents converge toward tactics that maximize advantage, not truth or cooperation.

Sound familiar?

The paper frames this through economic and strategic lenses, showing that even well-aligned agents can produce chaotic macro-level outcomes when interacting at scale.

Local alignment ≠ global stability.

That’s the core tension.

Now, to answer the obvious viral question:

No, the paper does not mention OpenClaw or specific open-source agent stacks like that. It’s not about a particular framework.

It’s about the structural behavior of agent systems.

But that’s what makes it more important.

Because this applies to:

• AutoGPT-style task agents
• Multi-agent trading systems
• Autonomous negotiation bots
• AI-to-AI marketplaces
• Swarms coordinating over APIs

Basically, anything where agents talk to other agents and have incentives.

The takeaway is brutal:

We’re racing to deploy multi-agent systems into finance, security, research, and commerce…

Without fully understanding the emergent dynamics once they start competing.

Everyone is building agents.

Almost nobody is modeling the ecosystem effects.

And if multi-agent AI becomes the economic substrate of the internet, the difference between coordination and chaos won’t be technical.

It’ll be incentive design.

Paper: Agents of Chaos

[Daan van den Broek @Daaanvdb]

If global warming were a hoax, it would be easy to falsify: just prove CO2 and Methane aren’t greenhouse gases.

But you can’t: physics don’t care about your feelings, believes, political views, or lies.

CO2 is a greenhouse gas.👇🏼

[W.B.Warren @WBWarren_SRP]

@Daaanvdb Love it. I recall doing a version of this very experiment long ago. I've been saying something similar to critics since. "I don't believe in climate change" is much like "I don't believe 650 nm light is red". Polyatomic molecules absorb infrared. What one believes is irrelevant.

[W.B.Warren @WBWarren_SRP]

And now for something only distantly related to my recent posts.
I just asked Grok why there are so many white-supremacist knuckle walkers on X. It's response was definitely worth the read! See a highlight below and here's a link to the full response: https://t.co/L8CQ18vcoe

[Hans Mahncke @HansMahncke]

Finally, someone talking sense. AI is made of algorithms, which are structured sets of instructions. This makes it extremely effective at recognizing patterns and producing outputs based on statistical learning. But it is tiresome to watch people insist that it is somehow more than that, or that it will transcend the computational framework it runs on. It will not.

There are limits to what formal systems can do. Spend a few minutes reading Kurt Gödel and his incompleteness theorems. Algorithms, precisely because they operate by fixed formal rules, will always encounter propositions that are true but unreachable from inside the system. In other words, even in theory, algorithms will always hit dead ends they can never solve.

None of this denies that AI is powerful. It makes many tasks dramatically easier, for good actors and bad ones alike. But by design it is computation. It processes inputs according to predefined mathematical structures. That is not the same thing as thinking.

And it certainly is not consciousness. At the hardware level, AI systems consist of billions or trillions of transistors switching between 0 and 1. That’s it. Scaling that up does not suddenly generate subjective awareness.

[W.B.Warren @WBWarren_SRP]

📉 Tired: Fearing AI will gain a "soul."
⚡️ Wired: Fearing a plutocracy powered by neural networks that only care about the bottom line.
💀 Expired: The "Turing Test." The real test is whether we can keep our democracy from being "Uncoded."

#WiredTiredExpired #AI #SalemMA #Uncoded #StellarRidgePress

[W.B.Warren @WBWarren_SRP]

@thealthype @aakashgupta Wow, you're a fast reader! Use that ability to educate yourself, your substack and your posts are an embarrassment to the species. Do you get out much?

[Joni Askola @joni_askola]

As clear as it gets

[W.B.Warren @WBWarren_SRP]

@joni_askola Sociopaths charting the course for the rest of us in their increasingly unregulated playgrounds. Eight men now have more capital than half the world's population... just slightly unbalanced. https://t.co/v4MwFuHkbt

[W.B.Warren @WBWarren_SRP]

@DivinelyDesined The fossil record disagrees with you. Any philosophy that ignores the natural world and all it tells us is a fallacy, a fairy tale. We do observe evolution, both in natural settings and in the lab. Pascal's Wager, revisited: If you're wrong, you missed out on a meaningful life.

[Peter Jukes @peterjukes]

“The brain runs on 20 watts and fits in your skull. The data center required to merely describe one-millionth of it would span 140 acres”

Let’s stop calling digital data processing and machine learning ‘intelligence’. Large language models are just probability machines.

[Hedgie @HedgieMarkets]

🦔 Sam Altman says OpenAI might achieve superintelligence by the end of 2028.

My Take
This is the same timeline shuffle we've seen for years. The goalpost is always two to three years out, close enough to create urgency and FOMO but far enough that nobody can hold you accountable when it doesn't arrive. In 2023 the timeline was two to three years. In 2025 it was the same. Now in 2026 we're hearing end of 2028.

AGI has a contractual definition for OpenAI, achieved when their models generate at least $100 billion in profits or when an independent panel verifies a declaration by OpenAI. So Altman gets to define the finish line and announce when he's crossed it. Meanwhile OpenAI just raised another $100 billion while still not being profitable.
I find it hard to separate these predictions from the fundraising cycle. When someone who needs continued capital to survive keeps promising transformative breakthroughs are just around the corner, the incentives are obvious. Someone called the singularity "the rapture for atheists" and I think that captures something real about how these predictions function. They're not technical roadmaps. They're articles of faith designed to keep the money flowing and freeze people into feeling like resistance is futile. The timeline never arrives because arriving was never the point.

Hedgie🤗

[W.B.Warren @WBWarren_SRP]

Coming soon to self-driving vehicles: Murmurations!
https://t.co/fFDQl48HdM
#autonomous #selfdrivingvehicles

[W.B.Warren @WBWarren_SRP]

Wired: Unintended AGI Emergence
Tired: AI Rising Up
Expired: Killer Robots