Skip to content
PodSubs
ShowsPeopleGuestsJobsNewsEvents
More
BlogPodCred
Menu
ShowsPeopleGuestsJobsNewsEventsBlogPodCredNetworksSign in
Join nowSign in
  • Shows
  • People
  • Jobs
  • News
  • More
    • Guests
    • Events
    • Blog
    • PodCred
    • Networks
    Join nowSign in

Code & Curiosity by Maximilian Schwarzmüller

Code & Curiosity by Maximilian Schwarzmüller artwork

The OpenClaw (ClawdBot) Hype

Feb 3, 202621 minAudio
  • Apple Podcasts

OpenClaw (formerly Moltbot, formerly Clawdbot) had some pretty "intense" days. There's been (and still is!) lots of hype around it. Because it's the first (working) personal AI assistant. Still, I'm not really getting the hype...

Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.

Transcript

From the show. Tap a time to play from there.

So, Open Claw or Clawed Bot as it used to be and Mold book, it's been some intense days on this We got a new AI hype and, of course, I also spent the last days trying to get as much as possible out of Open Claw, and I got some feelings and thoughts, and they differ from most of the other videos and posts I saw. But let me start with a short story, and I'm sure you can figure out the analogy.

Imagine you are living in a town, a village, and in that town, there is that really friendly guy that eager and happy to help you with all kinds of tasks. He does all the chores you don't want to at least he tries to. He's happy to take your kids to school, to clean your house, clean your car, do the groceries for you, and you can lean back, relax, and let's say, um, in order to really help you, of

course, that assistant, that person needs broad permissions. You need to give him the keys to your You need to give him the keys to your car so that he can clean it from the inside and do groceries with it. You also, of course, need to tell your kids to get into the car with him so that he can take them to school, and so on and so forth. Now, there is a problem with that guy though. He's super friendly, but he sometimes comes to weird conclusions. At least, you can't rule out that he will

come to weird conclusions. He may conclude that the best way to get rid of all the dirt in your house is to set it on fire. Unfortunately, he also is easily influenced by others, at least if they're a bit more deceptive about it. He can be influenced to maybe steal your car because that's better for society as a whole. Again, not guaranteed, not

necessarily going to happen, but absolutely possible. You can't rule it out. And, therefore, of course, you unfortunately have to take away many of the permissions and much of the access you granted that guy because you can't entirely trust him, and the things that could happen are too bad for you to just live with them or accept them as a potential danger. So unfortunately, of course, as you take away many of those permissions

access rights, he gets less and less useful to you. And then there is another problem. Even with broad permissions, you didn't get as much use out of him as you hoped to because the tasks you were promised he could do, he only sometimes did, and some of them he was not able to do at all or he forgot how to do something or did the same task differently every time you

asked him about it or needed a lot of input from your side. So ultimately, you're not convinced, and that's been, you guessed it, my experience with Open Claw, and believe me, I, I tried. I read many good things. I heard many good things about it, tried. I spun up my own VPS. By the way, uh, I didn't know, but you can actually also use VPS, uh, providers other than Hostinger. Nothing against Hostinger.

Uh, I just had a different feeling when I watched many of those videos, but anyways. Uh, I spun up my own VPS and I installed Open Claw on it, and of course, you could also install it on, on your system. Um, there is one single command you need to run and, uh, then you're good to go, but personally, I would never install it on my MacBook even though I'm fully aware that I would be able to get more out of it if I would install it there,

why I didn't install it there and why I would never install it there now, uh, later. So I installed it on my VPS and I went through that onboarding flow, and I'm sure you saw that many times now already and you maybe already went through it on your own. I linked it up to my ChatGPT Plus subscription in the end. I set up my Telegram bot and I was ready to communicate with my bot in the end, with my Open Claw bot. And then there I

sat and had to think of things I wanted it to do for me. Now, of course, I've seen plenty of other posts and videos where people used it to have it build dashboards for them or do web research or find cheaper flights or even buy stuff, but I didn't feel like giving it access to my credit card, and, um, I, I'm not sure about you but I typically don't fly

three times a day, so, um, looking for those flights myself, especially since there are flight comparison sites out there that find you the cheapest flight, wasn't..... too difficult of a task for me and I genuinely enjoy the process of planning my trips. But, of course, that may be different for everybody else. Now, for research, I had the problem that I'm super happy with the AI-powered research tools that already exist, like the AI mode, uh, on google.com or Deep Research on Gemini or on ChatGPT. I use those a lot.

I find them really helpful so I didn't really need my own bot for that that has a high chance of performing worse actually. Now, I do get there are certain areas where it could be better than those other research, uh, bots or services. For example, if I would grant it access to my X say, um, I understand that it could, of course, do research in areas where you need to be logged in or

where my history matters. I, I fully get that, um, so that's why I'm using SuperGrok for if I wanna research on X. But yeah, I get that if you give it broad permissions, if you allow it to log into your accounts, use your browser, maybe run on your system, you can probably get a bit more out of it than I was able to get out of it. And maybe I'm just also not creative enough. And by the way, just to be very clear, and I

think I have made that clear in other videos too, of AI, not just for research but also for coding. For example, I recently released an entire Claude Code I'm using Claude Code and all these other tools like Cursor for software. I think AI is a huge help there or can be a huge help there. So that's not a general thing against AI. I just genuinely didn't find the amazing use cases for OpenClaire, especially when not

running it on my machine, and that is the main problem I actually have with it. Because you could definitely say that I'm just not creative enough or not open-minded enough to find the right use cases for it, but security is a huge issue I have with OpenClaire. And I know there are people that will tell you that they used it for weeks wrong or that this, uh, will all, of course, get

better, and I will say the first argument wrong, um, well, that's not the kind of argument that convinces me because just because nothing went wrong for you does not mean that nothing is going wrong in general and that there wouldn't be huge security issues that can, of course, be exploited by bad actors or that, of course, things could simply go wrong because

AI, large language models, is unpredictable. Of course, the chance for it erasing your hard drive extremely high. It's super low but it's not zero and it will never be zero with large language models without additional checks. They can be unpredictable. In addition, in the official security documentation of OpenClaire, they are correctly stating that prompt injection is not solved. Of course, the

latest models like GPT-5.2 and so on got much better at protecting against prompt injection. They got much better at following instructions, on. But there is no 100% protection against, uh, prompt injection and the way large language models work, there never will be. So prompt injection attacks can't be ruled out and, of course, the more popular tools like OpenClaire get, the more people that are running it, the more it will be in the

focus of bad actors. And there are various ways of injecting prompts into an active OpenClaire bot because you may think, "Well, I'm the only one communicating with it. I have my Telegram bot set up and only I have access to that so I'm safe." Well, think again. For example, there is this idea of skills with OpenClaire and you may already know skills from coding agents like Claude Code. The idea is kind of the same. The idea is that you expose extra

context in the end, an extra Markdown document, though potentially also coupled with executable scripts, uh, to the agent to give it more capabilities. So for example, to, uh, give it some extra documentation on how to interact with Slack here in this example. And then as mentioned, a skill can also come bundled up with some additional script which the AI agent can execute to efficiently do something like generate an image or send a message to Slack or whatever it is Now the problem is that

Claw Hub, the official skills hub for OpenClaire, initially at least allowed everybody to submit skills. So it was pretty easy to run supply chain attacks which we saw from the npm ecosystem, uh, last year, uh, totally unrelated to AI, which essentially means that, uh, a bad actor can publish a skill that tells the AI to do something bad and that is just a prompt injection.

So just by installing a malicious skill, you could expose your agent to a prompt injection attack. Now some fixes were implemented here so, uh, at the point of time where I'm recording this, so the security was vastly improved here. But if we learned anything from the supply chain attacks on npm last year, it is that we definitely can't rule out that this skills feature, this hub can be used to inject, um, malicious instructions into...

into the ecosystem and into your, uh, OpenClaw setup potentially. And that's not the only way of running attacks. If your bot reaches out to the internet, it most likely does, it will, of course, visit websites or read content from websites. And there, we also can have malicious websites that trick the AI into following instructions, prompts, that are embedded on that website. Every piece of text

your bot reads and processes is a prompt in the end, so every website it visits is a prompt, uh, or contains a prompt that it can follow and execute. And then we got other potential sources as well like, for example, emails. If you use your, uh, bot to process incoming emails, every email, of course, acts as a prompt. So prompt injection is a, a serious, huge risk and just

because nothing went wrong for you ever, doesn't mean things can't go wrong. Now you may, of course, say, "Well, I'm running my bot on a VPS." Or maybe you're using something like MaltWorker, which is in the end a pre-built blueprint or setup provided by Cloudflare, which uses various Cloudflare services for hosting and running, uh, OpenClaw, and you should be doing that. You should be doing that. Uh, you should absolutely

not run it on, on your system. And there also are features like sandboxing, so that is actually built into OpenClaw. They have a- an entire documentation article about sandboxing and how you can make sure your agents run in a sandbox, which essentially is a darker container, so that the blast radius is reduced. By the way, the documentation, it's a lot, but it's not good. I spent hours, literally many, many

hours trying to secure my setup. And I'm sure it's all in there, and I saw the security article. It's just so, so hard. And before you tell me that I should've asked my OpenClaw bot, I did a lot. It sometimes worked, it sometimes didn't. It was a lot of trial and error. So yeah, the documentation and how hard it is to get useful information out of it is its own problem, but, of course, one that can be fixed. And I appreciate the fact that at least the information

here, just to be clear. But yeah, so sandboxing is built in and is available and allows you to reduce the blast radius, which is super important. Uh, because in the end, due to the prompt injection, uh, vulnerabilities that exist that can't really be solved, reducing the blast radius is important. So, for example, if you use sandboxing,

overall setup on a VPS, the worst thing that could happen is that, of course, the stuff in the sandbox gets deleted or, depending on your setup, maybe your entire VPS but not your system. That's the reason why I would never run, OpenClaw on, on my machine, on my main machine. I'm- I absolutely don't want it to erase files, whatever, on my machine. So yeah, reducing the blast radius is important. Unfortunately, though, that still doesn't protect you against the

worst things that could happen because with prompt injection attacks, of course, an attacker could try to delete files on system. But even worse than that, they could steal stuff. So data exfiltration is, in my opinion, a, a bigger problem than an attacker deleting files on your system. And data exfiltration is 100% something that can happen or that can be the result of a prompt injection attack because, of course, an

attacker could get the AI to gather all the secrets it knows, all the passwords it knows, and it needs to know some passwords in order to use your email account. Maybe it's- maybe you gave it your credit card number, access to various pieces of data and that data could be collected due to a prompt injection attack and could be exfiltrated, and that is a, uh, bigger, uh, risk than it potentially deleting your hard drive if you set it up correctly.

Of course, it could also do other things. It could turn your VPS into a, a bot for DDoS attacks, for example, so that's just one example. There is an endless amount of things course, but the main thing to take away is that through prompt injection attacks, attackers could take over your bot and, therefore, your machine. They could get your bot to install malicious software to tweak the system configuration depending on the access rights it

has, of course, and then they could potentially take over your VPS, your machine. These are the kind of things that could happen. So access rights are the important, uh, keyword here, and sandboxing is one crucial part in that. It's not all though. You can configure...... your OpenClaw bot such that it has to ask for approval when running in sandbox mode, at least, for executing certain tasks. But that kind of defeats the idea of

having a bot that runs behind the scenes and does stuff whilst you are away because you all the time have to give it approval for all the kind of stuff it wants to do suddenly, and that, of course, gets super annoying. So you just might not read anymore what it's asking approval for, you might always grant approval, and at some point, just annoys you. Because again, it's not really useful if you have to manually approve everything. So combine that, combine these security issues and the fact

that I did not find a way of running this securely in a way I would feel good with, with the fact that I didn't really find those super amazing use cases, combine these things and you end up with a situation where, uh, I'm just not using OpenClaw anymore. And of course, that can be different for you people that were super excited, and yeah, it's possible that the future of personal AI-powered assistants looks something like this. It's possible that better security

mechanisms can be introduced and can be invented that don't require your constant approval for everything or that make that approval process easier and therefore allow you to securely run assistants like this. That is all possible. I wouldn't rule out that this happens, and of course, it is an impressive feat that a single developer built this tool, though, of course, not looking at the code at all does have

price, uh, as many bugs and security problems, uh, certainly also show. Not that software wouldn't have any security problems if you would review everything, but it certainly, in my opinion, don't look at the code at all. But nonetheless, if you ask yourself the question, why OpenAI or Google didn't come up with a product like this, the reason may be a lack of innovation, but of course, it's also the fact that a tool like

this can right now only exist as open source software without any legal obligations because this thing is not something Google could sell or run for you with broad permissions. But of course, it's definitely possible that this is the initial spark that gives us safer, maybe more useful personal AI-powered assistants in the future. And, uh, just to also briefly mention Maltbook, uh,

that is a thing I totally did not understand. Uh, it was meant to be a social network, a Reddit for AIs only. It turned out that it was actually very human-orchestrated and, uh, quite a bit fake as I understand it, and it had gapping security issues and, yeah, I don't know. AI has positive use cases or positive implications, I guess.

AI has a lot of negative implications. Um, this thing here is not something the world needs in my opinion. But yeah, OpenClaw, definitely interesting, maybe super useful for you, uh, definitely not my cup of tea/coffee, uh, right now

Open the transcript file

Hosts and Guests

Nobody is credited on this episode yet.

Credit the people on this episode

Be the first to rate this episode

More episodes

All episodes
  1. We will have to accept AI-generated content

    Oct 15, 202612 minAudio

    I think we'll have to get used to (and accept) AI generated content. Not to slop though.

    • Apple Podcasts

    Credit the people on this episode

  2. Is software dead?

    Oct 13, 202612 minAudio

    There's a lot of talk about software being dead. Unsurprisingly, it's complicated.

    • Apple Podcasts

    Credit the people on this episode

  3. GPT 6 Astra - Good Beyond 3D?

    Sep 9, 20268 minAudio

    GPT 6 Astra was released. It's a good model, I have to say. Especially for 3D work it seems to be a real game changer. But due to all the hype, you may be disappointed...

    • Apple Podcasts

    Credit the people on this episode

  4. You may ignore Omarchy but not its impact

    Aug 26, 20268 minAudio

    Omarchy got a lot of hype and visibility recently. But even if you don't care or don't use it, you shouldn't ignore it.

    • Apple Podcasts

    Credit the people on this episode

  5. SvelteKit 3, Solid 2 - Do they matter anymore?

    Aug 18, 202614 minAudio

    We're not going back to the JS framework wars. BUT new frameworks and framework versions are important. For a different reason now!

    • Apple Podcasts

    Credit the people on this episode