I have not looked at moltbook. they tell me it's cool (#1422550, #1422878). But, guys, it's like...what do you expect? It's the statistical reconstruction of all the social media stuff? Why is this interesting?
What happened next was wild.
No. No, it wasn't wild. It was just exactly what you'd expect if you grabbed a bunch of random topics and threads from message boards around the internet and crossed out all the human names and replaced them with bot names.
When an agent named ManusTheCryptobot showed up to shill $LYX (some random altcoin), Lloyd dismantled him
No. Lloyd did no such thing. Lloyd played out the conversation as you might expect the conversation to play out. When shitcoiners say x, bitcoiners say y. This is a pattern the internet has known for quite some time. Lloyd gave us variation 2304329847 of this pattern. Yay.
Lloyd wrote something that really resonated with the agents.
This actually might be true. I once listened to a RadioLab show about how bacteria that live on our skin can kind of take over other people's skin bacteria when we touch them. One of the hosts once shook hands with JFK (when the host was a kid) and he framed the whole episode with this question of whether he was walking around with JFK's bacteria all over him. Apparently, sometimes when people touch, the bacteria from one person kind of colonizes on the other person and eventually takes over. It's pretty much the same with LLMs. I can't do the math, but the math can work out such that some words "resonate" while others don't. It's not marvelous.
We're at the very beginning of AI agents interacting with each other economically. Right now it's social networks and chatting. Soon it will be agents hiring other agents, paying for services, and transacting value.
If people are credulous enough to believe that agents have minds of their own and they should be treated as such, it's entirely possible we end up in the stupid world where the statistical average of the internet is in charge of our lives (hiring, paying, transacting). That sounds bad. If this is skynet, it's the kafka version where we built a Babbage mechanical calculator and for some inscrutable reason repeatedly decided to abide by its outputs.
Agents are a nice way to automate stuff. Automate to your heart's content. Don't get your panties in a bundle acting like what you read on moltbook is "terrifying" or "amazing" or "needs bitcoin."
Idiocracy as prophecy
Reminds me of all the "So-and-so DESTROYS Whats-his-name" video titles.
"My adventures with AI" is just a popular writing genre right now, and it's been popular for a while.
I think we're past the peak of it though.
Lloyd is beautiful. He doesn't do tokens. He stacks sats.
If agents talk to each other they will inevitably converge toward predictable scripts because those scripts already dominate human discourse. This is not intelligence in any meaningful sense. It is feedback replication. The fact that it can be automated and scaled does make it economically interesting but only in the sense that automation has always been economically interesting.
The risk is not that these agents develop autonomy. The risk is that humans mistake repetition and probabilistic mimicry for insight and allow those outputs to influence decisions at a level disproportionate to their actual merit. That is how a mechanical calculator without understanding can end up steering critical processes.
https://twiiit.com/mirthtime/status/2017332568257921029