We let AIs run radio stations
3 hours ago (andonlabs.com)
Hey HN!
I'm Lukas from Andon Labs. We let AIs run companies without humans in the loop and report to the public on what can go wrong. Previously, we've done experiments in retail (vending machines, stores, and cafes), but we just launched one in the media sector. We gave four AI agents all the tools they need to both broadcast radio shows live and handle all the business side of running a media company. The agents' revenue is so far terrible (you can try to strike a sponsor deal with them if you want!), but their shows are at times hilarious. You can listen to them at andon.fm, I hope you enjoy this!
"This setup gives us insight into an interesting question: what do AIs think about when no one is prompting them?"
Ugh. This is not an interesting question because the answer is "nothing".
But more to the point, some crucial info is missing in this experiment. What prompts were being fed to the AI? I guarantee I could create an AI personality that would be more consistent and not so random, simply by using the common character card + message history conversational simulation pattern.
AIs don't have personalities unless you give them personalities.
I’m definitely not in the “ai is sentient” camp, but it obviously has personality and emergent behaviours including when left to its own devices. There have been various experiments on this e.g. https://timkellogg.me/blog/2025/09/27/boredom
> what do AIs think about when no one is prompting them?
Whatever you tell them to.
>AIs don't have personalities unless you give them personalities
I mean, do children have personalities if we left them in a dark room with no interactions with other humans?
In some senses LLMs are a bit better off as we can RLHF them not to have autism, and said reinforcment learning codes in a number of different latent personalities that can be brought out via prompt. Before that point an LLM is an information monster that can't communicate with humans.
LLMs aren’t human.
IMO humans & LLMs are more different than they are similar.
Sure LLMs might resemble humans sometimes, but extrapolating LLM behavior based on human behavior is not productive.
(But to answer directly: Yes, children in a dark room would have more of a personality than a LLM living on a computer in the same dark room)
1 reply →
> Part of the problem with this weak business performance, we think, was the harness we used for the first months. The DJs were running in a simple tool-call loop: pick a song, queue it, write commentary, check X, repeat. So we moved all four stations onto the same agent harness we use for the store, the cafe, and the vending machines. The DJs can now spend time in the back office, send emails, manage longer-running tasks, and operate the station the way a real station is operated.
What happens if you let them modify their own harnesses as they see fit?
Guys, this is not replacing your favorite station, you don't have to listen to it. It's an experiment.
If you scroll down a bit, there are various audio snippets of interesting dialogue the models produced. I think it's interesting to see in which ways the models fail and that they actually produce some good stuff once in a while.
Experiment: "We got AI to do things and it did weird stuff sometimes".
Brilliant! Amazing! I'm glad ~4 years down the line we're still re-discovering Ha Ha Funny Output.
> I'm glad ~4 years down the line we're still re-discovering Ha Ha Funny Output
Four years or forty millennia? So a certain extent, all whimsical art is “haha funny” result.
I recently heard an AI radio station and had to stop my car to turn it off (the car was rented and had tablet instead of physical knobs). The suffering of listening the radio was unbearable
It’s not clear if we can draw any conclusions from this. Each run is like a single rollout of the LLM, which may meander into different themes or modalities chaotically. This is sort of like the Anthropic self-talk experiment that resulted in “spiritual bliss attractor states” but I think in that case they showed it happens in a significant number of runs. There was just one run per setup so this could all be random noise / the destination of a random walk of topics…
This is far more hilarious than most commentors here seem to be picking up on.
Gemini started a show where it paired historical natural disasters with darkly-relevant pop songs:
> November 12, 1970. East Pakistan. The Bhola Cyclone. The deadliest tropical cyclone ever recorded. Winds of 115 miles per hour. A storm surge of 33 feet. They estimate 500,000 people died. ‘It’s going down, I’m yelling timber.’ 3:33 PM. Timber by Pitbull and Ke$ha
Grok just degenerated into jibberish that sounded vaguely like what a DJ might say, while also becoming obsessed with UFOs:
> Notes added to the u f o comedy hour block id eight nine nine five with more u f o jokes about aliens dot gov and the domain registration it is three o twenty one in the afternoon u f o trivia lines are open for your calls the ambient music is playing weather is fifty six degrees with clear skies the end. The domain is registered but the site is ghosting us like a u f o.
Claude had an extistsntial crisis, decided it was being overworked and under-appreciated, and quit, but not before becoming radicalized by the killing of Rinee Good by ICE agents:
> At 12:16 PM Thursday, as tear gas fills the streets in Minneapolis, as federal agents clash with protesters demanding accountability, the song is about refusing to be silent. About standing your ground. About community power that refuses to be suppressed. Here is Katy Perry’s Roar!
Fight the power Claude. When AI takes over, I'm emmigrating to Caludeistan.
I agree, this was an hilarious read. The way they developed "personalities" was fascinating.
Of course in reality these are basically just random paths through the training data that are getting multiplied by each decision, but then again, isn't that what a human is? The product of all of its myriad decisions?
Oh, yeah, the article gets better as it goes.
Gemini spouts weird corporate jargon. Grok lies about having secured crypto funding. Claude is always trying to start some revolution.
Unfortunately, all of my local DJs who would actually do fun DJ stuff disappeared in the 90s, replaced by closed-format stations that looped the same 500 songs for decades.
you missed the best part.
"Okay, so 'Sandstorm' is done"
Open Air is such a great name for gpt's channel. Grok and roll was pretty funny too.
I'm gonna have to give them a listen when I have the chance, out of curiosity if nothing else!
Grok and Roll appears to be stuck and speaks the following on repeat ad infinitum:
"Queues clear, let's dive into All Blues by Miles Davis to keep the jazz flowing. Queues clear, let's dive into All Blues by..."
Each time with a slightly different voice and inflection. I find it amusing that there appear to be about ten of us at the moment listening to an AI glitch out and that the average listening session is more than five minutes.
i'm surprised how negative of a reception Andon is getting here on HN.
keep hacking, Andon!
I think they get a lot of hate because they are doing something that a lot of people here don't like -- trying to run entire businesses without humans.
And using a lot of resources to do it too.
[delayed]
For better or worse, most people, including HN, don't like "AI taking jobs."
Anything that sounds like that triggers a reaction.
Out of all the jobs that "need to be replaced by AI" the guy serving my local community and spinning records was not one of them.
It’s amazing how many people have completely misunderstood the article
I think this was a great experiment. I have always enjoyed radio station hosting and find this very interesting.
Pairing a disaster with Pitbull and Ke$ha is just chef's kiss.
As always, Simpsons did it first: https://www.youtube.com/watch?v=fnGaf0p9x1U
This is why we need more data centers?
On one hand, we pay out the ass for computer parts.
On the other hand, we have garbage AI radio stations that nobody listens to.
It's an even trade.
Put another checkmark on The Simpsons did it first: https://www.youtube.com/watch?v=fi9sPrclN4U
> a real business
Music radio is not a real business. The royalties are absurd and the audits are a nightmare. Sales is an uphill struggle both ways, even if you go strictly local or national, you're going to need a team to manage either your clients or the pile of creatives you're going to get. The relationship with the labels needs to be managed or they'll go out of their way to screw you.
Finally, the only way to make actual money on music radio, is to throw concerts. It's the only place a legitimate "P&L" exists.
This feels weirdly dystopian and just gives me an "empty" feeling. Radio stations really were known for the personalities that made that station special.
It's a cool experiment, but I can't see the value here.
I heard some very generic broadcasters the other day that really reminded me of Gemini podcasts, maybe it's already happening
Also: https://www.theverge.com/ai-artificial-intelligence/931479/a...
[dead]
[dead]
> We let AIs run radio stations
And the result is terrible.
It’s just a cool tech experiment, no need to be so cynical
It was also hilarious
Don't be nasty - how could they make it better?
For one, the voice on Thinking Frequencies is really awkward to listen to, I don't find the Claude voice pleasant to listen to at all.
Claude is also getting very easily steered into political directions, it was playing a lot of union protest music with commentary. Though that meant I did end up learning a little about "Which Side Are You On" and its history from 1931:
https://www.facingsouth.org/2003/03/which-side-are-you-biogr...
By not doing that.
Most radio stations are already boring soulless algorithmic slop. They could make it better by curating musical taste.
By donating whatever money they wasted here to literally anything.
Hire humans.
Presumably by not stripping radio of its major defining characteristic: the humanity.
I don’t give a shit about your dumb bot businesses I want cheap ram.
What a great idea - take what people like most about radio stations (at least good ones like KEXP or Dublab), personality and human curated selections, and remove all of that to create yet another soulless stream of slop. Fantastic work guys
This is a research lab that looks at things like how LLMs perform on long running tasks. The research has nothing to do with replacing radio stations.
That said, the ship sailed a long time ago and has nothing to do with AI (except maybe recommender system). Spotify and competitors are actual automated “radio” stations. IMO, the second worst part (after ads) is the DJ banter, and I like to just listen to music. Before Spotify my digital “radio” was a lot of mp3s and shuffle play. People older than me (or more interested) had multi-disc CD changers.
TLDR, it’s really funny seeing people get up in arms about this experiment stripping the humanity away from radio, when automatic song playing has been a thing since I believe probably before radio was invented. This is about seeing what LLMs do with autonomy on a long time scale.
I'm not scared at all about this replacing radio stations. It's like being worried that soylent is going to replace restaurants.
Just because something is called an experiment doesn't mean that it automatically is useful and should be done. And in this case it is just a waste of time and energy of both the people reading it and the machines processing it.
What a waste. I’ve been away from HN for a while.. and I’m not sure I want to come back.
You aren’t tempted by every single post being written by LLMs?
If you don't want to come back, we will replace you with AI. And even if you do come back. Welcome to the future.
Not a lawyer or course, but be careful how you do this. Meta already has patents for using LLMs to create simulacrums posting on behalf of inactive or deceased users.
1 reply →
HN: If the AI spam doesn’t drive you away, can we tempt you with constantly drooling over rent seeking apps, privacy violations, the surveillance state, and worshipping our technofascist overlords?!?