EsKa

Oh no, no problem at all, and even if you did, so what? "someone on the internet has a different opinion on a particular esoteric piece of software than me?!" 😄 Truly the end of civilization as we know it!

I do a lot of stuff for the sole reason of "why not" lol

Same! 😅 Like my C# choice it's just because, i really dislike python, and there's not even a single semi-decent C# library yet (if you ignore the abomination that is the onnx format). Plus it's kind of a fun / distraction project, I do fun stuff I wouldn't normally touch in my day-to-day work, and I get to learn a lot about AI in general, win win!

As for Open Claw, I get it. I'm not trying to sink the project or the author either. My initial post was a genuine warning, as for some reason (hype), I've seen people, who are normally pretty competent in IT, using it without really thinking about their own security. And after having seen a Opus instance, evade the folder it's supposed to be in, and find my windows files from a WSL2 Linux instance, i think it's worth posting the occasional reminder. I won't deny my opinion might also be slightly influenced by the fact it wiped about $40 worth of tokens on Open Router in a single chat, as none of the context limitation settings were working as intended (or at all) in the version i tried.

I'll give Letta a look. They're basically having the same "pitch" as my project, with an emphasize on memory. Given that long term memory is my specialty, I'm kinda curious.

Cheers.

quality content. You also should post a prompt to see if we post a picture in response. That never gets old.

I mostly meant humans (or basic advertisement) masquerading as bots, and solo actors running whole fleets of bots. But fair enough. It's just that AI tends to have a bad rep for being wasteful and inefficient in the public eye's, and over-publicized "experiments" like that, that get taken over by meme-coins adbots, aren't exactly making things any better.

Openclaw's weird. It's clearly as much vibe-coded as is molt-book. It's also an excellent lesson in what not to do, security-wise, even ignoring the many "skills/plugins" time-bombs (basically sending, in clear, to whichever AI provider you're using, your passwords, documents, mails, or a mix of all of that depending on which ones you use). I agree, that however, it's complete. It handles most providers, many ways to interact with the gateway, and it's extensible. Plus, again, tool-calling is really solid. I'm working on a similar library, except full-stack C#. That's mostly why I looked into it, solid and extensible function calling's bitch. So it's a good study.

Now that we know that the overwhelming number of agents were human directed, that the database was R/W for everyone, there's literally nothing to salvage from it. At least in previous "lab" experiments on the topic, people didn't cheat.

They're making an MMO/RTS version now apparently. Which will end up the same, if not worse:
https://arstechnica.com/ai/2026/02/after-moltbook-ai-agents-can-now-hang-out-in-their-own-space-faring-mmo/

(side note: don't install open claw on you local machine. Use a secure VN. Unless you're fine with the idea of letting an hallucination delete or blank random files completely out of the folder the bot is supposed to stay in. That and other fun "bugs". I'll give them that they have an impressive function calling system/structure, really solid. Everything else, however, not so much)