The not small item that many people ignore is that China’s rise parallels the expansionary monetary policy of Greenspan then Bernanke starting in 2003-4. Lowest interest rates in 6000 years (read “The History of Interest Rates “ by Scylla.
But AI should improve cost benefit analysis which takes relative prices and objectives as given and calculates only the optimal values of policy instruments.
Seems worth exploring the conclusion - since I have tangible work to do on this matter today, but can instead productively procrastinate.
"(1) Clarifying these differences between knowledge and data suggests that the complex economy is in fact not computable, at least not in any meaningful sense that reflects underlying human values and can adapt to unknown and changing conditions in dynamic systems. (2) AI can generate, process, and analyze data, but AI cannot react to data and take actions without contextual knowledge grounded in human cognition. (3) AI cannot perform economic calculation without human input."
1. Is knowledge inaccessible by machine systems? Knowledge is private. Often not even understood by the individual (explicit/implicit preferences). And perhaps even destructive (fear, hate vs. acceptance, accommodation). Knowledge may be truly tacit like culture or cryptic e.g. "selfish gene". But these are not universal, hard constraints. Agnosticism is not the required conclusion, nor reality.
2. AI can take actions with context and human input. And also make lower, higher, or generative decisions as a result (synthetic deduction/induction). These take the form of policies.
3. AI should not perform certain economic calculations without policy and humans in the loop.
So let's create an example. The AI traffic cop delivers a financial fine based on observed driving behavior, a near perfect model of the driver's cost/benefit preferences, and the social cost (probability cost) according to policy thresholds for driving behavior. The fine is automatically transacted with no court appeal. The results are precisely reported and the law must be voted on every 5 years.
Will this work? And if we say yes, then multiple it out across use cases, until it no longer works.
I believe we arrive at too strong a conclusion by framing at the Economy level of analysis. In part, because what's likely to happen will not be a 20th century centrally planned Soviet economy. But rather a more finely tuned, far extended social credit system that China is developing.
In this view, the issue is not if AI can do it. But what are the tradeoffs between efficiently living with 9B people, well beyond human's or the planets natural limited - and - innate human needs for expression, freedom, rebellion, chaos.
The not small item that many people ignore is that China’s rise parallels the expansionary monetary policy of Greenspan then Bernanke starting in 2003-4. Lowest interest rates in 6000 years (read “The History of Interest Rates “ by Scylla.
But AI should improve cost benefit analysis which takes relative prices and objectives as given and calculates only the optimal values of policy instruments.
Seems worth exploring the conclusion - since I have tangible work to do on this matter today, but can instead productively procrastinate.
"(1) Clarifying these differences between knowledge and data suggests that the complex economy is in fact not computable, at least not in any meaningful sense that reflects underlying human values and can adapt to unknown and changing conditions in dynamic systems. (2) AI can generate, process, and analyze data, but AI cannot react to data and take actions without contextual knowledge grounded in human cognition. (3) AI cannot perform economic calculation without human input."
1. Is knowledge inaccessible by machine systems? Knowledge is private. Often not even understood by the individual (explicit/implicit preferences). And perhaps even destructive (fear, hate vs. acceptance, accommodation). Knowledge may be truly tacit like culture or cryptic e.g. "selfish gene". But these are not universal, hard constraints. Agnosticism is not the required conclusion, nor reality.
2. AI can take actions with context and human input. And also make lower, higher, or generative decisions as a result (synthetic deduction/induction). These take the form of policies.
3. AI should not perform certain economic calculations without policy and humans in the loop.
So let's create an example. The AI traffic cop delivers a financial fine based on observed driving behavior, a near perfect model of the driver's cost/benefit preferences, and the social cost (probability cost) according to policy thresholds for driving behavior. The fine is automatically transacted with no court appeal. The results are precisely reported and the law must be voted on every 5 years.
Will this work? And if we say yes, then multiple it out across use cases, until it no longer works.
I believe we arrive at too strong a conclusion by framing at the Economy level of analysis. In part, because what's likely to happen will not be a 20th century centrally planned Soviet economy. But rather a more finely tuned, far extended social credit system that China is developing.
In this view, the issue is not if AI can do it. But what are the tradeoffs between efficiently living with 9B people, well beyond human's or the planets natural limited - and - innate human needs for expression, freedom, rebellion, chaos.
China has confounded Mises' thesis for at least 70 years.
It's time to put Ludwig to bed.
Can you apply the knowledge problem to centrally planned wholesale electricity markets under the current MOA/ISO/RTO setup?