Pay structure

In school we “learned” a list of alleged propaganda methods. None of them were real. The list itself was propaganda to distract us from real propaganda.

One effective REAL method is distorting verb aspect, treating a permanent or habitual process as something new and unknown. Weather becomes CLIMATE EMERGENCY! Flu season becomes VIRUS PLAGUE! Corruption becomes TRUMP!

Advertisers can also distort aspect. Soap becomes NEW BODY WASH! Databases become NEW BITCOIN! The advocates of AI-based coding are playing this way. Claude is the FIRST CODE HELPER IN HISTORY! Nonsense. Computer programming IS code helpers.

Many complicated machines are programmable without language helpers. A clock with chimes and alarms has a quarter-hour loop, a one-hour loop, and a user-controllable alarm loop. A manual typewriter has a margin alarm loop with user-controllable limits and a controllable number of linefeeds per return.

The first Hollerith computer was programmable in the same way as a typewriter or cash register, by turning dials and setting switches to different modes. Punch cards were only for data, not control.

The BIG step happened in the 50s when punch cards expanded to set up the loops and alarms and choices. At that point the commands were visible as text. After that BIG step, the rest of the stages evolved naturally.

Much later, experienced programmers developed their own helpers, sometimes a specialized language with its own keywords, usually a set of automatic routines and macros for common activities like converting graphics to SVG or converting text to HTML. These helpers are programs writing programs.

Claude operates at this thoroughly familiar level, except that it hears ordinary sentences and answers in ordinary sentences.

AI takes one GIANT STEP beyond the previous sequence of language-ification. Before LLMs, each layer was developed by one inventor or team or company. We can identify the startpoint. Most importantly, the single startpoint was PAID for the work, usually salary from the goverment, IBM, DEC, or Bell Labs. When we use languages or operating systems we are paying the developers directly or indirectly through a traceable chain of pay-for-value licensing transactions.

Open source was the first breakdown of the pay chain. Open source takes advantage of “volunteers” who enjoy building code. AI goes to the next level, treating everyone as unpaid “volunteers.”

Summing up: The PROBLEM with Claude is its pay structure, not its grammar or math. Previous layers of language-ification were created by identifiable people or teams who were paid and credited for their work. Claude steals from everyone and doesn’t pay or give credit.

= = = = =

NOW I SEE sidenote: Math didn’t evolve! Math as taught in school still uses mysterious wildly inconsistent symbols and NON-symbols, not readable words and sentences. The ABSENCE of a symbol means multiplication except when it doesn’t. Math didn’t yield to the printing press, the typewriter or the computer!

Programming languages gained ground when they stopped trying to look like math and started trying to be UNDERSTANDABLE. Cobol went all the way, writing out ADD, MULTIPLY, SUBTRACT, DIVIDE in an effort to be auditable by non-programmers. Most used the symbols available on a typewriter with a few adjustments like * instead of x for times.