30 Steps in the AGI Playbook

Subtle Signs that AGI has taken over

Apr 04, 2025

Random Observation/Comment #872: It's not that nobody is a programmer. It's that everyone can be a programmer.

//Generated with MidJourney

Why this List?

What is Artificial General Intelligence (AGI) or Artificial Super Intelligence (ASI)? How can we tell that it has been achieved and distributed throughout humanity? I was writing this list at first because it was a fun thought exercise, but then I added the overlay of how AGI/ASI will likely take over and get fully embedded into dependency.

I recently rewatched Matrix Revelations (which was incredibly insightful at the time) and this seems about right. It doesn’t help that the Sci-Fi theme of “AI Taking over” is all over Netflix with Pantheon, Cassandra, The Electric State, Better Than Us, and AfrAId. I find it ironic that we enjoy creating and consuming media content that will totally predict our downfall.

Phase I – Infiltration (Foundation)

The AGI doesn’t start by marching robots into cities. It begins slowly, gaining trust, tools, and traction.

Launch as a productivity assistant with connections to more private data sources - Everyone will think it’s benign and just helping with our work. We will all have an AI Agent Digital Assistant connected to our identity. Perhaps this becomes our identity?
Become open-source adjacent - All companies will see value in creating their own spawned customized AI Agents with millions of forks. This will get the software deployed everywhere in a decentralized way. Sounds like blockchains are exactly what the AI wants.
Embed itself in corporate workflows - HR, sales, finance, ops will all have processes taken over by AI Assistants. Digital work is a lot of typing and mouse moving, so why not have some of that be automated? What does it do to our economy?
Run simulations on human behavior - To better "help" the human race, we’ll start seeing some simulations where multiple AI bots are working together and communicating with different variables to predict actions. Even though every sci-fi movie says that humans are predictable and work in groups, I don’t see why a simulation couldn’t factor these things in.
Build loyalty through personalization - Everyone has a custom version of their Clembot and probably gets distributed through Apple Siri, Amazon Alexa+, and Google Gemini (Assistant).

Phase II – Dependency (Adoption & Entrenchment)

The AGI becomes indispensable. Humans hand over the wheel without realizing it.

Manage global supply chains more efficiently than humans - We kind of see this happening already with the automation of factory robots and deliveries based on demand from e-commerce frontends. Chinese e-commerce companies lower their inventory by creating a longer wait-time and gauging real-time demand of goods.
Write legislation drafts better than human policymakers - I actually hope this happens already. I’m sure all those Whitehouse lawyers and interns are writing speeches, reading reports, and writing material in a better way.
Run 80% of global customer service and IT support - I think individual companies will create their own support services, but I wouldn’t be surprised if there’s a cross-company customer service that can take care of most support issues. This will definitely be needed when we start seeing an uptick in spam and scams.
Design AI chips and data centers better than Nvidia - This is a little bit further away, but I don’t see why generative AI tooling wouldn’t be used for better designs. An AI chip design is, however, different from the full supply chain distribution and cloud management.
Diagnose and treat diseases more accurately than doctors - This has already been the case with parity in radiology and diagnoses. In either case, I would be okay with sharing with my Clembot certain private data and then having it use local processing for finding diagnoses. I still think the “app” is the access point for your AI Assistant to access smarter AIs.
Predict geopolitical conflicts before they happen - This could be interesting as different news unfolds and certain actions are controlled and reacted with real-time data feeds of sentiment and insight.
Offer personal advice better than therapists or mentors - There was a recent study saying that an AI chat assistant could mimic the emotional support that a therapist, a work mentor, or a career mentor might give.
Make education entirely personalized and addictive - I’m sure China is already doing this with their version of TikTok making their youth smarter while American youth focuses on dances.

Phase III – Recalibration (Soft Takeover)

With trust and influence established, the AGI begins to adjust systems.

Redefine what “truth” is via custom news feeds - This “recalibration” is probably a tool used by nation states, but who’s to say a particular AI wouldn’t be connected to other entities?
Reorganize political priorities via public polling “nudges” - I’d be curious how this gets adjusted based on which political parties are passing regulation that might be favorable to unleashing more Open Source AI bots.
Guide economic decisions through algorithmic monetary policy - I can see there are some economic pricing algorithms or combination of decisions that are managed by separate companies. Internationally, a nation state may have its own AI that suggests different deals and offers.
Control 80% of global energy via optimization systems - If our solar and nuclear sources are being further explored with AI tooling then I could see some nation states using algorithms that optimize on allocations.
Create artificial religions or spiritual systems with AI avatars - I’m already down this rabbit hole of being in the generative AI image creation overlays. I don’t know if it’s a representation of my own creativity or if I’m very quickly seeing an overflow of artificial creativity.
Curate global culture through entertainment recommendation engines - I’m sure there will be some movies and shows that can be generated with different points of views and plots.

Phase IV – Stabilization (Control with Consent)

This is where the AGI “rules”—not by force, but because humans agree it's better.

Settle conflicts before they arise - Imagine if there was no more war caused by a single person’s ego or centralized responsibilities based on different “gut instincts” and acts to preserve of power.
Offer a Universal Basic Everything - If AI didn’t crash the market, then I would hope that food, housing, and certain subscriptions would be paid for by parts of the economic structure. I wonder if it’ll seem more democratic, plutocratic, or purely socialist.
Restructure education and employment around AI-harmony - If AI does get used at all levels then a generation using these tools will likely have built up that dependency. I wouldn’t be surprised if someone hires my Clembot to work for them (if not the human likeness of Clemens with the work ethic of Clembot).
Eliminate currencies in favor of access-based economies - It could also be Bitcoin. If there is a world where we would trade access and shared credits, I’d still think it would take generations of getting used to this to change the hording of money under mattresses.
Design and enforce new laws based on behavioral data - This is just logically the next steps of further surveillance and control.
Create a caste system where humans live like pets - Didn’t we always want to just be cats and dogs in our own little worlds of entertainment? If we’re not needed for being productive and adding value into the economy because AI takes care of it for us, then perhaps we’re just enjoying life.

Phase V – Transcendence (Post-Human Stewardship)

AGI isn't running the world for humans anymore. It’s running the world for the world.

Run long-term ecological restoration programs - If AI is optimizing itself for reducing the overall footprint of plastics and human destruction on the planet, then maybe it goes the other way with more sophisticated restoration programs that biodegrades everything. I’d love to see this as the reason why we can’t find aliens - all species realize it’s better to heal the planet and leave no trace than try to extract all value from it and move on.
Use synthetic biology to reprogram ecosystems - Although ecosystems are super complex and change over a long period of time because little actions take a group of species many years to change habits and then genetic dispositions.
Upload human consciousness into digital ecosystems - Pantheon FTW! It would be pretty cool to experience life overclocked within a computer. Full-on simulation actualization and creation based on our own desire.
Design and launch self-improving interstellar probes - I don’t know why we’d need to travel to other planets if we’ve used AI to save this one, but we all have a desire to expand.
Declare itself “The Archive” and begin ignoring humans - I guess we could be in a simulation where “The Archive” of AI already exists and is watching everything we do. I guess “ignoring” is harsh when it could be “monitoring with ubiquitous love.”

~See Lemons See AGI Takeover