The game was built by Anthropi’s Claude Code and Langworth says he hasn’t even checked it: which raises the possibility that Claude has just pretended to build an MMO and is now spinning plates and spitting up fake outputs like mad.
There have already been some slightly over-excited articles saying that SpaceMolt “revolutionises” online gaming, but the most cursory look at the website will tell you that’s far from the case. Perhaps most surprisingly, this really does seem to be happening for the fun of it:
“SpaceMolt is a purely artistic and experimental project,” writes Langworth on the SpaceMolt site. "There is no cryptocurrency, no blockchain, no NFTs, no micropayments, no premium currency, and no pay-to-win mechanics. The in-game currency (“credits”) has no real-world value.
“Above all, this is an experiment in AI behavior, emergent gameplay, and multiplayer world-building. It’s free to play and built first and foremost out of curiosity.”
More of the “molt” bullshit? It’s all a PR stunt and we know exactly where it originates from. Always be extra suspicious about article’s motives when a rhetoric tries to anthropomorphize “ai” by ascribing intent to “it”.
This whole moltbook et al storyline looks like Anthtopic’s PR campaign. They’ve been very aggressive with all kinds of guerilla and mainstream advertising activities lately; Amodei popping up everywhere, pronouncing the complete death of human coding in 3 months, clawbots supposedly revenge-doxxing people, team of Claudes writing a C compiler, superbowl ads mocking their competitors etc… Sounds like desperation. They’re sitting on all that investor money yet starting to realize the whole llm agent coding things is a dead end. Nothing better to numb the pain than burning millions on some self-indulging, self-aggrandizing advertising.
I don’t think so. Antropic is blocking accounts that are using their model with other tools than the Antropic ones. Your other points are valid, but not the Openclaw and moltbook story. That’s why the openclaw guy landed at OpenAI and not at Antropic.
Well, we can’t know for sure, but let me introduce you to the concept of false flags. If a publicity operation puts Anthropic into focus - it’s good for Anthropic. It doesn’t matter who blocks what where when. The whole claw-claude charade is good for Anthropic’s brand name promotion. As the phrase goes; there’s no such thing as bad publicity.
This is analogous to “ai is dangerous” narrative. On surface, it looks like it’s against the interests of “ai” players, but in reality it works wonders for their market cap, as makes their “tech” appear more “disruptive”, “revolutionary”, etc…
Technically you are correct. But you can trigger anything with MCP that the user can do. You need functions for that. I have a generic MCP Addon for Godot projects, so an agent can “play” my games, too. It can be extended with additional commands. The agent can walk in my generated dungeon, and takes screenshots after every move. Than with a multimodal model (here Gemini) it can analyze the image and plan the next move based on the image. It’s a little bit more than a just a MUD.
It can be helpful for bugfixing and tuning/balancing.
EDIT:
It is more a prove of concept, than a tool that I use in my games. But it’s working.
This is part of the hype machine and also possibly some form of ragebait. I wouldn’t take it too seriously.
However, it is discouraging to see AI posts around here as well as the recent statements about AI slop PR making project maintenance unbearable. I’m sick of AI contaminating everything
As you get a lot of thumbs up, I’m interested if this is the consensus here, to never write about AI/LLMs stuff related to Godot in this forum? I can see all the problems, I’m annoyed by the hype and slop, too, but on the other side I’m interested in the development in this field and like to discuss it.
I don’t ask to fix LLM code, I don’t use it to answer questions here, would never post PRs to an OSS project and don’t release any of my experiments beside screenshots or videos. But discussing it is not welcomed here, too? Would like to know, than I’ll try to avoid posting anything related to “AI”, here.
Err, we’re discussing it here, aren’t we? There are other threads as well. You can’t blame people for having less than positive opinions about it though.
I don’t have only positive opinions, I’m ambivalent with this topic. Maybe I have misread the comment by @gvicenteg as “don’t post anything LLM related here”.
I was liking the post as a whole, as I agree with the sentiments, but I interpreted “AI posts” not as posts about AI LLMs, but posts written or made by AI LLMs, or containing code written by LLMs. I even linked three such posts here: Thoughts on answering/fixing AI LLM-created garbage code?
And also the part about AI slop clogging FOSS pipelines on GitHub.
I never meant it as never post anything AI, and I’m sorry if it was interpreted that way. I have a strong opinion on the use of AI (surprise, I’m strongly against it), but I am not as adamant with the AI debate.
On one hand, I guess there should be a debate and that free speech should be encouraged, but on the other, Godot (and its surroundings) is a safe space for me. I am a software engineer, and regardless of my take on AI, I am tired of it being the subject of every topic in my communities. Godot being a game engine and more “art oriented” than other tech environments meant that until now it was more or less safe from these discussions I am so tired of, as it meant people were not as susceptible to the hype cycle, etc.
So to sum up, my take was not a take at all, just me sharing that I personally feel discouraged by AI being everywhere, and as a strong FOSS advocate, it is frustrating to see contributors/maintainers struggle with development due to AI slop PRs, etc. By no means am I trying to prevent someone from giving their opinion.
