Adaptive Markets

Andrew Lo

(Professor MIT)

Probably Approximately Correct

Leslie Valiant

(Professor of CS, Harvard University. Recipient of the Turing Award.)

How the Internet Became Commercial

Shane Greenstein

Stephen/ChatGPT 4o: The Adaptive Markets Hypothesis is based on the insight that investors and financial markets behave more like biology than physics. Financial markets aren’t always efficient or irrational, but evolve like living systems—shaped by human behavior, competition, and adaptation over time.

“The brain as a predictive machine. We can think of a narrative as an advanced for of simulation, using very high degree abstraction to describe phenomena”

Stephen/ChatGPT 4o: Posits that both biological evolution and intelligent learning can be understood through “probably approximately correct” algorithms—computational processes that adapt and thrive in complex, unpredictable environments.

"One now needs to analyze not only the algorithm itself but also the algorithm's relationship with its environment."

"I believe the primary stumbling block that prevents humans fro being able to learn more complex concepts at a time than they can, is the computational difficulty of extracting regularities from moderate amounts of data, rather than then need for inordinate amounts of data."

"I believe the attempt to make a thinking machine will help us greatly in finding out how we think ourselves." - Alan Turing

Stephen/ChatGPT 4o: Explains how the internet evolved from a government-run research network into a competitive, profit-driven commercial platform through decentralized innovation and policy shifts in the 1990s.

“By 1999 it was obvious that nobody a decade earlier, not even the most optimistic visionary, could have forecasted the direction of the Internet in commercial markets. After privatization too many contingencies and surprises shaped outcomes making the ultimate economic outcome fundamentally unknowable in advance.”

“The combination of low friction and high decentralization unleashed a particularly potent form of market based learning”

Stephen/ChatGPT 4o: Argues that open-source software development, driven by decentralized collaboration and transparency, can outperform traditional top-down approaches in innovation and efficiency. Fascinating to re-read in the era of closed versus open AI development.

"The naively simple strategy of releasing every week and getting feedback from hundreds of users within days, creating a sort of rapid Darwinian selection on the mutations introduced by developers"

"Put simply the closed source world cannot win an evolutionary arms race with open-source communities"

"The path towards open source in the evolution of such markets are well-illustrated by the reconvergence of data networking on TCP/IP in the mid 1990s following 15 years of failed attempts at empire-building with closed source protocols such as DECNET, XNS, IPX and others" - worth a blog post on contrast compare AI development versus TCP/IP.