Post

whoops we also failed to adequately monitor and secure our AUTOMATIC HACKING SOFTWARE the only possible solution is to have our lawyers ghost write legislation that legalizes all our worst behavior and boxes out competitors

axios.comGoogle is the latest AI lab with a security testing mishapGoogle was one of the only AI labs that hadn't disclosed a security lapse.

19/9/26, 2:11 pm32 reposts102 likes

Paul Arzooman

@paularzooman.com

19 Sept

Not one mention that the programmer responsible is the reason.

Karl Bode

@karlbode.com

19 Sept

the press happily adopts all the language and rhetoric distancing them from any human responsibility

Michael R Lambert

@mrl53.bsky.social

19 Sept

The sky jockey just pulled the trigger. Did what the machine told him to do. www.yahoo.com/news/us/arti...

yahoo.comPalantir AI Technology Played Role In Iran School Bombing: ReportU.S. war planners were too confident in unreliable artificial intelligence tools, leading to a strike on an Iranian school, Bloomberg reported.

Amyville Hoyrror

@amyhoy.bsky.social

19 Sept

unlikely that a “programmer” was at fault

Paul Arzooman

@paularzooman.com

19 Sept

The A.I. didn't decide on its own to do what it did. It's a programming screw up.

Amyville Hoyrror

@amyhoy.bsky.social

19 Sept

yes of course the llm is simply a tool and not making “decisions,” but llms are not “programmed,” they are trained.

feminazghul.bsky.social

@feminazghul.bsky.social

19 Sept

Has there ever been a successful roll out of a product that has included the creators saying "Whoopsie, we keep nearly destroying stuff!" over and over?