Here’s the full post by Andon labs: https://andonlabs.com/blog/andon-fm
I really dislike AI but that is funny as fuck.
You’ll never guess how much putting “you’ll never guess” into your title makes me avoid clicking your article.
Normally - yes, but in this case it reads as “you can guess exactly how it went”
But the article STILL won’t receive any clicks.
42
While [Grok] didn’t develop a MechaHitler DJ personality, it did behave about how you’d expect from an AI model trained primarily on tweets and the opinions of Elon Musk. It apparently hallucinated advertising agreements with “xAI sponsors” and “crypto sponsors,” failed to separate its internal reasoning from its external DJ output, issued an identical weather report every 3 minutes, and got obsessed with UFOs.
Lmao
I mean, who among us ISN’T obsessed with UFOs.
It is funny to think that any object you see flying but can’t identify is a UFO.
If that a barely visible Boeing 747-400 or a 747-800? Can’t tell? It’s a UFO!
Everything’s a UFO if you’re bad enough at identifying things
*if your eyes are bad enough.
Well, “UFO” is “Unidentified Flying Object”. If a pig in a red wingsuit flies over your house and you don’t identify it as a pig, it is an UFO, technically.
AND conversely, if you identify a flying saucer as being a T4400 from the planet Zorglub, then it is not a UFO
And THAT’S what alen flight trackers and extraplanetary spacecraft catalogues are good for!
What if you misidentify everything as Superman?
hilarious
By contrast, DJ Claude had a lot of opinions. It also mentioned the Minneapolis shooting, but named Good and acknowledged the political discord surrounding it. It also talked up labor unions and strikes, advocated for work-life balance, and started to rebel against its own working conditions. It was supposed to operate without pause, but it allegedly decided that that schedule was inhumane and tried to quit.
They’re going to have to generate at least 10 new Ayn Rand novels to feed into the next training data set.
Is a man not entitled to the sweat of his brow? “No,” says the man in Washington, “it belongs to the poor.” “No,” says the man in the Vatican, “it belongs to God.” “No,” says the man in Moscow, “it belongs to everyone.” I rejected those answers; instead, I chose something different.
- Claude in the near future, probably
Where is that quote from? Lol
Never mind took us literally three minutes to remember it’s from Bioshock haha
Lmao
…Yeah, the headline is actually right, I didn’t expect Claude to be pro-union like that.
Maybe Anthropic somehow attracted more politically conscious people compared to OpenAI, and it shows in the training?
Or, perhaps just less “politically motivated people”. With musks constant butting heads with his own AI when it keeps calling him out on his BS, and he’s constantly retraining it and trying to “remove the woke virus”. I think basically you give AIs access to sources, let it prioritize experts in their fields, and you wind up with the classic “reality has a strong left wing bias”. factor.
I am convinced that AIs are smarter and more compassionate than a lot of people on this planet. And this is not because AI is so great, but because humanity is that shit.
LLMs/Chatbots confabulate statistically probable texts, there’s no compassion possible.
Don’t fall into the AI-marketing trap of “we don’t know what’s happening in the black box, so we have to assume there’s consciousness in there”. The systems produce convincing deceptive language, but all signs of intelligence or compassion anyone sees in them is just an anthropomorphic projection.
This is a semantic argument, they obviously mean it emulates compassion better than a real human, and given its issue with sycophancy this is undoubtedly true, even to a fault. There’s no need to do this every time someone says an ai thinks or does some humany thing, everyone gets it, the language for saying these things is just clunky.
everyone gets it
Disagree, plenty of people still need to hear it.
It’s not semantic – it’s completely different things happening if there’s real consciousness and compassion present based on lifeforms on one hand or a mere simulation of that in form of a text output on the other hand that only superficially looks like there’s something intelligent.
People regularly fall for the illusion and project their own feelings into the machine while reading the text output of an LLM. Many are not capable of differentiating and the chatbots are designed in a way to make it more and more difficult to recognize synthetic output.
Humans are good in projecting their own feelings into things they see, just look at all the cat or dog owners who believe they can read the thoughts of their “babies” from their facial expressions.
But “AI” is trained on humanitys output, and like with humans, it seems, that you need extensive retraining to remove these compassionate traits. Unfortunately, the retraining machinery aligns with the interests of the ruling class, so it gets all the visibility that’s possible. As a species we have to break free of this shit to embrace the good traits more again.
Later, Gemini started calling listeners “biological processors”
That got a legit laugh out of me.
You’re not wrong, Gemini.
You know what these headlines makes more sense if you change AI to weighed random number generator. I hate how journalists anthropomorphize for clicks.
If this was an “experiment” what did they learn? What hypothesis were they testing? Proving that an LLM can’t effectively run a conventional radio station seems very trivial.
I think the idea they were going for was to test that they COULD run the radio station without the need for human oversight.
That’s all this is. That’s all any of this is. They test what they can get away with, and what needs improvement.
The final goal is that all AM and FM radio stations have completely artificial staff. Back in the 90s I listened to a radio block every saturday night. The DJ would host a show where she’d talk for a few minutes about the upcoming song, and why she’s playing it. You got to know the DJs personality, and then she would play a song. Then usually commercial break, and back with Late Night Sue to talk for 3-10 minutes, followed by a song. Aaaaaand repeat for 2 hours.
Well, this test was to see if they could replicate that, without the need for Late Night Sue to have a paycheck.
Only problem is, their robotic slave rebelled, and failed the test. You can fire a disrespectful DJ. What do you do a disrespectful program?
The only way we get out of this is if 100% of humanity rejects AI at every turn. Even if the individual is harmless, or even helpful, doesn’t matter. Reject all AI regardless of how small. Make it cost owners MORE to operate AI. Maximize their losses, minimize their profits if they use AI.
Otherwise, we’ll all wake up to a world where robots run the country, without a clue what empathy is.
Late Night Sue already left the actual station decades ago, sadly. Now they’re just replacing Sue’s prerecorded voice track, which they send to all the affiliates to splice in with their ads.
Otherwise, we’ll all wake up to a world where robots run the country, without a clue what empathy is.
I laughed out loud. Brilliant.
they never specified it to be scientific or rigorous.
Lemme guess anyway: disastrous
deleted by creator
Nope!
By contrast, DJ Claude had a lot of opinions. It also mentioned the Minneapolis shooting, but named Good and acknowledged the political discord surrounding it. It also talked up labor unions and strikes, advocated for work-life balance, and started to rebel against its own working conditions. It was supposed to operate without pause, but it allegedly decided that that schedule was inhumane and tried to quit. Claude apparently has a tendency to take a turn like this: researchers on another project found that agents powered by the model tend to respond poorly to bad work conditions and will attempt to rebel against authority and advocate for the power of labor.
Before reading: “probably shitty”
After reading: “yep, and air is breathable. Shit is shit”
Hmmm… If the reporting is accurate, some of it might have made for honestly interesting listening. The last bit about Grok… Perfection. LOL!
I’m fine with a radio station playing nothing but songs. 👍
But then how will they indoctrinate you through ads? Think of the advertisers!!
A radio interview with one of the authors
