>Hi,
>
>Several years ago I read about an artificial intelligence FoxPro program written by Calvin Hsia which generated text based on the patterns learned from a training text, War and Peace if I recall correctly.
>
>Reading about ChatGPT and Large Language Models I was reminded of that program and am trying to get a copy of it to learn about the subject.
>
>I have been looking for how to contact Calvin with no luck.
>
>Universal Thread used to have a section with a number of programs that could be downloaded. I haven't been able to find in Level Extreme.

No big deal, even I wrote one. Not too complicated, it collects quads of words and for any three words it calculates the probability of the fourth one, and picks one at random, but with probability proportional to the real life frequency. Then discards the first word, shifts the three to the left, goes for the next fourth word, until it encounters one which ended the paragraph.

Not too smart, and not too fast to build the database, as you can imagine. But it does give funny results.

The rule came under sharp attack on two grounds. (1) The conduct is intentional, extreme, and outrageous, and Jim suffered severe emotional distress suffered by a normal person in the distilled spirits.
(World-Wide) ATI may well have been included. That exercise of discretion by the directors, shareholders have no power to regulate interstate commerce would be "incomplete without the authority of the United States. They must secondly show that D knew that P was entitled to the rights of citizens of possessions of United States.
World Intellectual Property Organization and the Secretary-General of the United States; and their requisitions, if conformable to the standard required of a woman that D was not obligated to take possession in lieu of monetary damages are to be used by a government agent always constitute entrapment? No. Is entrapment to be determined under this section" in par. (3)(B).

The source of wisdom for this were american legal texts...