1. EachPod

Lies, Damned Lies, and Siri

Author
cj little and Jeff Parker
Published
Thu 09 May 2024
Episode Link
https://podcasters.spotify.com/pod/show/nnsnspdcst/episodes/Lies--Damned-Lies--and-Siri-e2ic0nt

(0:00) Pre-Show


(1:10) Open


(1:25) Special Day: Magnetic Letter Day!


(2:50) cj's week: A Surprise Visit From Beck


(4:17) Jeff's Week: I’m Not a Robot!


(5:47) Headline: FCC fines wireless carriers millions for sharing user locations without consent 


(10:26) Headline: Bluesky confirms Jack Dorsey is no longer on its board


(11:27) Headline: TikTok sues US to block law that could ban the social media platform


(15:39) Segment: The Siri is a Liar


(43:06) cj’s recommendation: Stanley Cup Hockey & The National Anthem


(44:50) Jeff’s Recommendation: Long Shot on Netflix



You might want to reconsider what you ask your personal, virtual assistant to “just do” for you.  As the boys discuss in this week’s episode, sometimes your well-meaning Siri can just get it plain wrong - or in cj’s case, plane wrong!  We suspect this “Siri poisoning” will continue to be a thing, just like the DNS poisoning of yesteryear, and the more current AI poisoning plaguing training data today.


But before Jeff gets a chance to revel in cj’s naivety, the boys discuss more FCC fines (yay!), Dorsey leaving Bluesky (meh), and TikTok suing the US to prevent their sale-or-ban future (boo!).


And ferrous friendly letters, whatever that means!




Beck joins Cage The Elephant on stage to perform Trouble




What really happens when you click 'I'm not a robot'




Headlines


FCC fines wireless carriers millions for sharing user locations without consent | CNN Business




Bluesky confirms Jack Dorsey is no longer on its board - The Verge




TikTok sues US to block law that could ban the social media platform




Segment


Domain Name System - Wikiwand




DNS spoofing - Wikiwand




What Is Data Poisoning? - CrowdStrike




This new data poisoning tool lets artists fight back against generative AI | MIT Technology Review




How 'sleeper agent' AI assistants can sabotage code • The Register




AI poisoning could turn models into destructive “sleeper agents,” says Anthropic | Ars Technica




Trojan Puzzle attack trains AI assistants into suggesting malicious code




Close


Vegas Golden Knights vs Dallas Stars 4/27/24 Round 1 Game 3




Watch Long Shot | Netflix Official Site





Send us a ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠voice message⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠!

Share to: