Technology Taoist | Contrarian | Modern Elder | Wisdom Worker

Token efficiency is mechanical, how fast & cheap a model outputs text.

Thinking efficiency is cognitive, how much signal that output carries.

A system can burn millions of cheap tokens & still produce low-order thought of little value.

#TechTao #AIeconomics #ThinkingEfficiency #ValueDensity

AI can be quantified as follows:

Value Density = Output ÷ (Token × Time)

How much value a system creates per unit of energy and attention.

Big isn’t smart. Efficient is. The next frontier isn’t “more parameters,” it’s more value per joule.

#TechTao #AIeconomics #ValueDensity #EnergyIntelligence

“Why is the ocean king of a hundred streams? Because it lies below them.” - Lao Tzu

Apple just took the Taoist path. Tim Cook’s confirmation that ChatGPT is only the beginning, with plans to integrate more AI partners over time.

#AppleIntelligence #AI #Ecosystems #TechnologyTao #ContrarianStrategy

Mental models begin to oscillate, and sooner or later, the synthetic starts resembling the orginal works from someone, somewhere.

All knowledge and information exists, the more interesting question is how it was discovered, especially complex patterns.

That’s not plagiarism. It’s recurrence drift.

How?

RAG was meant to ground AI in truth.
Instead, it’s teaching machines how to loop.

Each cycle of, retrieve, generate, re-train feeds the system its own echoes.

#TechTao #RAG #AI #Copyright #SyntheticData #Recursion #PatternDrift

The next copyright war isn’t really about words or shapes.

It’s about who owns the mental models or patterns of thought themselves.

This is a very difficult problem to solve as “I think therefore I am”.

#TechTao #AI #Copyright #SyntheticIntelligence #Singularity #RecursiveEpistemics
#MentalModels

Once an LLM learns mental models, copyright becomes a paradox. It no longer needs to copy, it thinks in patterns once owned by others.

Over time, synthetic data layers atop human work like sediment, and from it, new creations emerge, legally grey, but conceptually novel.

#MentalModel
#Copyright

The unintended consequences of convenience.

Sharing a meal 🥘 at a Restaurant, was as much about face to face social engagement as it was about eating food.

It provided novelty, a certain amount of uncertainty and friction.

Restaurants are #Instagram canvases, instead of places for conversation.

#AI #TechTao #OIA #Entropy #Truth #FlowOverForce

Action:

Anchor intelligence in authentic, verifiable and diverse content and context.

The next challenge is not more data, it’s higher signal integrity.

#AI #TechTao #OIA #Entropy #Truth #FlowOverForce

Insight:

Entropy is not chaos, it’s equilibrium. When everything is equally plausible, nothing is true.

AI without grounded data becomes a philosophical echo chamber.

#AI #TechTao #OIA #Entropy #Truth #FlowOverForce

Grok AI

Observation:

Grok now feeds on X and Grokipedia, two mirrors reflecting each other’s distortions.

When a model consumes its own noise, truth density signal decays.

#Grok #AI #TechTao #OIA #Entropy #Truth #FlowOverForce

Never say Never.

Dick Tracy’s watch was once a dream in 1946, now it’s a reality 75+ years latter.

Fiction plants the seed. Engineers water it. Culture harvests it.

#TechTao #AI #Future #Innovation #Singularity #Future

To learn, one must first fail. We don’t learn from success as much as we do from failure.

#AI accelerates failure, it lets us explore, discover, and crash faster, virtually and safely. The best leaders fail forward.

#Leadership #AI #Failure #Success

Lao Tzu warned 2,500 years ago:

“The more laws and restrictions, the poorer the people; the more cunning they become.”

There is a fine line, overregulate AI, & you don’t get safety you get smarter circumvention by people with resources.

Systems adapt faster than bureaucracies.
Flow >Force.

#AI

Up until the mainstream availability of #LLM, the world to most people was mysterious, as to understand & comprehend a topic it took time to develop expertise.

Two things are happening; one a compression of time & two a greater number of people are delving into the unknown.

Knowledge is abundant.

Need to redefine how we measure compute in an era of efficiency.

Unlike FLOPS per watt, Compute Spatial Density (CSD) is how compute fits in real space, thermally & electrically stable. CSD captures the full stack: silicon, cooling, memory, networking, power.

#ComputeDensity #AI
#AIInfrastructure

All complex systems, natural, synthetic or hybrid will oscillate towards a state of entropy.

That is least amount of effort, for maximum amount of output.

In the ultimate state of decay, the output has low value, that is why evolution takes place with new inputs.

#AI is no different to #Evolution

While everyone focuses on #AI to solve silver bullet problems, they are missing an important trend.

AI is solving a large number of everyday problems, in seconds that previously took hours or days.

It’s the accumulative sum of these everyday problems being solved by #AI that is going unnoticed.

The ongoing shortage of refined rare earths, is going to require a novel approach for resolution.

Urban Mining via more aggressive recycling ♻️ of electronics and batteries 🔋 will be an approach that needs to be taken more seriously by all countries.

#RareEarths #UrbanMining #Recycling #Electronics

More debt and the solution will be more tariffs.

Supply Chain Risk Management, Second Sourcing, you would have thought that COVID lessons would have been learnt.

Consider the problem may well be the inability to ask great, precise and concise questions. GiGo principles are more important than ever.

You will receive exactly what you ask for in depth, scope and accuracy.

#GiGo #LLM

LLMs are like children growing up, they have to learn what is wrong, rather than what is right. In time x on the future , if you place the same code in an LLM the result will surprise you.

The point; perfection is perfected over time, expectations other than this are unrealistic.

Implications:

Grass-fed code drives speed. Grain-finished code creates soul.

Great Software Developers will know when to stop grazing and start finishing.

#AI #LLM #Software #Development

Grain-finished ( #Human ):

Humans refine the final 20%; architecture, nuance, UX, security, and judgment, adding density, trust, and distinctiveness.

Grass-fed ( #AI ):
AI grazes on vast data, producing broad, fast, low-cost code, syntax, scaffolding, tests, docs. Like grass, it’s abundant and efficient but bland and uniform.

We are now in transition phase where a well funded & managed company with above average software developers building large scalable software stacks, can feasible move to a model, where 80% of code is AI generated other 20% is human generated.

The analogy is similar to grass feed, grain finished.

Information is not knowledge, knowledge is not understanding, understanding is not comprehension.

The human trait of comprehension & application is the strength that is valuable in an age of #AI, #LLM, abundant information and knowledge.

Comprehension + Application = Wisdom

#WisdomWorkers

Observation:

The leading LLMs are learning a lot about U.S Constitutional Law at the moment.

Insight:

Taking a complex structured text like the U.S Constitution and training an LLM, gives the everyday person, access to a Constitutional Paralegal.

#ConstitutionalLaw #Legal #Paralegal