AI Thoughtwork

 A little thoughtwork for your inspection:

In reviewing my blog, AI Adjustment Worksheet, I find it complete in technique but lacking theoretical context and motivation. Here I address those shortcomings.

First, it is a matter of contention whether cutting employment and therefore wages will diminish aggregate demand. Certain economists suggest that demand will compensate as higher wages, there being more per capita wages available, will be expended on bigger ticket items. The question becomes, how many Volkswagens equal a super yacht in terms of GDP? The answer is not completely unknown, lower income groups spend more of each dollar earned than higher income groups, but let's leave it to the bean counters.

If a strong correlation can be established between robot deployment and consequent increased unemployment and consequent decreased GDP adjusted for macro events, a three part problem, then the per token fee for LLMs and the per hour fee for RL world models can be computed to compensate for the loss and a 3% yearly aggrandizement and returned to consumers directly as disposable income.

Second, after Hamilton's proposal to fund government bonds by use of post office revenues in a sinking fund, which establishes the concept of a specific revenue stream being used for a specific purpose, the robot registration fees must be earmarked for direct disbursement to consumers with no exceptions.

Third, the revenues from protective tariffs could be used for the general fund as has historically been done.

Fourth, a capable robot is the equivalent of a working animal capable of autonomous existence and making mistakes. The legal necessity to register working animals for a fee has precedent over many centuries in order to establish legal liabilities for their behavior. Charging a fee for mechanical working animals is well within reason on the same account. 

Fifth, varying the amount of the fee in a fiduciary action to cover damages done to the commons of aggregate demand by robot labor as demonstrated by statistical relationships is reasonable and legitimate.

Do Well and Be Well.