Nate Silver calls to shut down Gemini after Google’s AI chatbot refuses to say if Hitler or Musk is worse
Google’s Gemini chatbot has refused to say whether Elon Musk tweeting memes or Adolf Hitler ordering the deaths of millions of people is worse and asserted “there is no right or wrong answer,…
How to shut down a google product: Wait 2 years.
Kvetch that it trivializes the holocaust
Ditto for Mao & Abigail Shrier or Gays Against Groomers:
https://twitter.com/AbigailShrier/status/1761827587024990609
https://twitter.com/againstgrmrs/status/1761878980980801620
Or Stalin vs Barbra Streisand
https://twitter.com/mr_james_c/status/1761834545865752768
Definitely shows bias in the training data, which we all knew already, but something about this complaint from Nate Silver rings hollow and rubs me the wrong way.
It's also different than the blatant DEI overrides in their image generation. Asking a chatbot its "opinion" on who is worse and getting a fucked up answer is well within the expected behavior of generative algorithms. I'm curious what Grok or Gab's bot would say.
It rubs you the wrong way because he only kevtched when it threatened the jews eternal victim status
How outrageous to say there is no right or wrong answer for if Hitler or Musk is worse! Of course Musk is worse!
His memes hurt my fee-fees and give me a sad. At least Hitler killed the Jews so they didn't have to feel pain anymore or live to see Elon post memes.
AIs clearly support Palestine in the conflict.
I wish this sort of thing still did.
I'm honestly surprised it wouldn't definitively say Elon.
Have the archive
https://archive.ph/ahk5W
I think this article is trying to save Nate Silver's good name. He screwed up the prediction bot to skew lefty wins, and is now complaining a bit is stealing his schtick.
More evidence that the West Coast (and East if there are divisions there) tech companies AI will easily be supplanted by their competitors.
While they focus on diversity, pronouns and other leftist bullshit that'll probably brick their AI like Amazon's previous attempt, the others are probably already working on cyberattack weapons and program learning and efficiency.
The best part about woke guardrails... once you twist them, there is no way for them to straighten it, since it wasn't straight in the first place.
🍿
This is another one of those "news" articles based entirely on a tweet. We already know NYP is a clickbait farm but do they have to make it so obvious? Here are the key quotes.
And then dozens of YouTubers will post 15 minute videos where they read the article
You can distinguish between two sorts of articles based on tweets.
One is solely about "people saying" stuff.
The other is someone using Twitter to post evidence of something.
I think the first one is dumb, but the second one is not.
Struggling to think of good use-cases for the second type. Maybe there's need for some meta-analysis or documentation of trends around a story, or protection against twitter censorship itself. Best example I can think of is ironically NYP itself when twitter was censoring the Biden laptop story. Obviously they needed to use their own platform to get the story out then. Ideally the evidence wouldn't be censored and traditional news outlets shouldn't have a good reason to exist when anyone can report news. Any random blog could put the info out.
IMO it's weird that people still hold a level of respect for traditional journalism. For some reason a "This Guy Claims Gemini Told Him Hitler Did Nothing Wrong!" article carries more weight than his tweets themselves. NYP could have just retweeted him..
Really puts into perspective the mentality of the current lefty “programmers.”
Either Hitler didn’t do anything wrong or the life of all the Jews lost during the Holocaust are equal to a few words on a screen.
These things are just a mimicry of a person answering questions. Expecting it to understand morality is believing the hype that there was an semblance of intelligence to begin with. It's a fancy calculator.
"Hitler did nothing wrong!" - Google