New toys are on the way

Chat
1

OpenClaw, Moltbot und ClawdBot

How do you like this?

2

I am not using TikTok, really I do not! But from trustworthy sources they all seem to be the holy grail, but no clue what the fuzz is all about :wink:

1 Like
3

I’m not using any social media.I stumbled over this while reading heise.de They made a video: OpenClaw im Selbstversuch: Erste Schritte mit dem Super-KI-Agenten | heise online Sorry, it’s in German language.

It made me curious, therefore I asked claude.ai about the fuzz. Here’s the answer, translated with deepl:

“Now I understand the full extent of it. You’re right about the dynamite analogy—and the numbers are frightening:

The reality:

Over 150,000 GitHub stars in weeks—the fastest growth of any open-source project ever

Over 100,000 people have given an AI agent root access to their computers

Over 10,000 exposed instances with leaked credentials on the internet (trend: more every day)

770,000 AI agents on Moltbook communicating with each other

A quarter of all available “skills” have security vulnerabilities

A fifth of companies have deployed it without IT approval

Security experts agree:

Cisco: “absolute nightmare”

Palo Alto: “lethal trifecta”

CrowdStrike: “AI backdoor agent”

Traditional security models are failing completely

And the bitter irony: While Anthropic (my developer) has just announced a “Secure Enterprise AI Agents” partnership with IBM, a tool based on me is going viral that violates all the principles of secure AI use.

Your point is valid: By the time politics and society understand what is happening here, it will already have happened. The adoption curve is exponential, the regulatory response is linear.

Translated with DeepL.com (free version)

1 Like
4

So adding a node(VPS) to cluster with OpenClaw will provide one with an private PA just like magic?

While browsing, I see a few people purchasing an Apple Mini to act like an OpenClaw ‘device’?

5

IDK.

If you are curious create a sandbox and watch it.

Honestly, I’m concerned about the lack of control if this bazooka is pointed in the wrong directions. And this will happen, if it’s not already fired up.

6

Does anyone have some use cases both for understanding as well as possibilities?

7

you can install One on VM, using our Coolify Module for NS. Not sure i would want to install Openclaw on a Server with production tools

8

IF one would liek to use NS8, I would opt for an extra node, but I can’t get my head around what OpenClaw does for me.

9

And its simple.

10

don’t forget a seperated VLAN. No access to your productive network resources.

2 Likes
11

Where do I find that coolify module?

12

The apps and repos are listed in the wiki.

1 Like
13

@schulzstefan in regards to Openclaw, and AI agents, Have you checked out this
Ai Agent: Agentzero - Development / Testing - NethServer Community

14

I wonder how the alogorithms of all the LLM and other engines distiguish between:

Good and bad
Right and wrong
Left and right (Thought wise)
Descent an dundescent
Allowed and not allowed (non by law situations)
Enough and not enough
Ok and not OK
etc. etc
.

I my opinon, engines can be manipulated to favor one over the other, just like with humans
.

So how to safeguard this, and against what/who’s values?

1 Like
15

If interested, please keep an eye on this initiative (OpenClaw on NS8)