The Network Effect Paradox: Why AI Won’t Disrupt Established Winners

The biggest levers of generating wealth are all tied to “who you know”.

How is AI supposed to disrupt that market and grow GDP?

You’re either Coca-Cola or McDonalds with 2 billion users already and established brand, or a VP of Fortune 500, or own a social network, have access to hundreds of millions of patient records, run a large Hampton-style community, etc.

Outside of tech (which was non-existent until 30-40 years ago), most successful companies are 100 years old. Takes forever to build the network, and when you do, it’s hard to go out of business if the unit economics are just “good enough” and you aren’t placing 30 FTEs where 2 people in a startup can deliver the same output.

High-ticket, high-volume, or massive data moat. Strong trust factor.

I haven’t found a single use of AI disrupting this, besides OpenAI or Google which are simply selling shovels to the gold diggers.

LLMs may be leading to “internal efficiencies” or improving operational effectiveness in certain cases – similarly to how the Internet helped optimize communications and logistics, machine learning got integrated into bank credit ratings, and generic automation was bundled in BPM systems and ERPs/CRMs.

In terms of revenue generation, if anything, it may shrink ARR and global revenue in favor of faster turnaround and fewer cycles.