OpenAI Training Bot Crawls 'World's Lamest Content Farm' 3 Million Times in One Day.

OpenAI's spider seems to have got stuck in a web that is too big for its capabilities. The artificial intelligence giant seems, in fact, to have a crawler whose programming has recently been found to be inefficient.

John Levine, besides being the author of 'The Internet for Dummies', is also the owner of web.sp.am/., a site whose only function seems to be the world's lamest content farm. To be precise, Levine's domain started out as an online experiment - in fact, the owner pointed out that the site receives a derisory number of visits from human users, whereas it seems to be the favourite site of most modern crawlers.

In addition to being visited daily by Google's spider, this content farm has recently revealed itself as a web from which it is impossible to escape, especially for GPTBot, OpenAI's training crawler which, earlier this month, was stuck inside the site for an entire afternoon - exploring the pages within it 150 times per second.