TWiT 926: Bias, Hallucinations, and Jail Breaks

Beep boop - this is a robot. A new show has been posted to TWiT…

What are your thoughts about today’s show? We’d love to hear from you!

I am intrigued by the term Hallucinations being used to describe AI, when my stupid Android phone gave me the monthly run down from Google Maps and declared I have been visiting KFC, a thing I would never do - it also made up several other places and has no clue how far I walk in a month, apparently if I walk our Great Dane a daily average of 1.5 miles, in Google terms that adds up to just over 7 miles a month.


Does she look like she only does that little?

1 Like

That’s one big puppers. She looks like a horse!

2 Likes

52Kg 9 months old 114pounds

Listening to writers dispute, an AI could easily write a better Seinfeld or Friends and they were 2 of the most popular.

As for people criticising AI content on the Web. IMHO the quality dropped when everyone and their dog decided they were and Expert and wrote a blog not mention the huge personal biases that are reflected in the content of sites like the Verge. If you had an AI that Parsons d academic journals only, you could improve the quality of the Web dramatically

It’s been a crazy busy week and I didn’t finish this episode until Friday (I listen on the drive to/from work). This week was a great episode. I enjoyed the discussion on AI. I think we will learn a great deal from the AI village at DEFCON this year.

1 Like