[FRIAM] Optimizing for maximal serendipity or how Alan Turing misdirected ALife

[uǝlƃ ☣]

One way to do this might be to universally install disambiguation pages on every word of a tweet. Of course, some words wouldn't necessarily explode (i.e. the disambiguation page would be small, with only a few entries) like "a" or "on". But some words (e.g. "her" or "him") might explode in an interesting way. 20 years ago, the disambiguation pages for these pronouns would reflect societies systemic prejudice against trans people. But today they would be fully blossomed launching points.

A "bottom up" method could then be devised to track and take statistics on the paths followed through these pages, inductively inferring categories from that traffic.

On 5/28/20 9:39 AM, Marcus Daniels wrote:
> I would say that companies like Twitter should massively annotate serious offenders and cancel accounts as needed.    It doesn't have to come from top, but it isn't going to come from the bottom.   There should be processes to keep conspicuous liars from ever gaining visibility.   They don't have to involve black vans, as satisfying as that might be.   But maybe advanced natural language processing codes that escalate issues to editors.

-- 
☣ uǝlƃ