Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Nvidia NemoClaw

github.com

Post reply on HN
Print view View original

171–180 of 291 posts

Re: Nvidia NemoClaw

#171

Post by tokioyoyo »

jesse_dot_id wrote: post #88

The fully autonomous agentic ecosystem makes me feel a little crazy — like all common sense has escaped. It feels like there is a lot of engineering effort being exhausted to harden the engine room on the Titanic against flooding. It's going to look really secure... buried in debris at the bottom of the ocean. When a state sponsored threat actor discovers a zero day prompt injection attack, it will not matter how iso…

I’m still not sure why there’s this general idea that people care about security/privacy. For critical systems, sure. But over the last decade, we’ve seen that an average person will always choose fun and convenience over security.

Even the analogy to free love is interesting, because sex in itself during that era was fun. Frankly it’s the same nowadays as well, we just figured out a way out of most of the diseases.

Re: Nvidia NemoClaw

#172

Post by rcarmo »

observationist wrote: post #133
Netcob wrote: post #65

Am I missing something? Why is everyone talking about sandboxes when it comes to OpenClaw? To me it's like giving your dog a stack of important documents, then being worried he might eat them, so you put the dog in a crate, together with the documents. I thought the whole problem with that idea was that in order for the agent to be useful, you have to connect it to your calendar, your e-mail provider and other servic…

What makes it even better is that these dogs are like Malinois. If they want to get into something, they will; people have had their entire network compromised by bots they left running overnight, and any important information like account logins and so on runs the risk of being misused. It's one thing to sandbox, maybe give the bot a temporary, limited $100 card or account to go perform a specific task, but there's…

I beg to differ. I took one, defanged it (well, I let it keep the claw in the name), and turned it into a damn useful self-modifiable IDE: https://github.com/rcarmo/piclaw

Yes, it has cron and will do searches for me and checks on things and does indeed have credentials to manage VMs in my Proxmox homelab, but it won't go off the rails in the way you surmise because it has no agency other than replying to me (and only me) and cron.

Letting it loose on random inputs, though... I'll leave that to folk who have more money (and tokens) than sense.

Re: Nvidia NemoClaw

#173

Post by ghxst »

tgtweak wrote: post #136

I think the more useful tool would be an LLM prompt proxy/firewall that puts meaningful boundaries in place to prevent both exfiltration of sensitive data and instructions that can be destructive. Using the same context loop for your conversational/coding workflow makes the task at hand and the security of that task very hard to differentiate. Sending POST?DEL requests? risky. Sending context back to a cloud LLM with…

The proxy you suggested sounds similar to a WAF, I don't doubt there's use for it but I would assume it comes with similar downsides.

Re: Nvidia NemoClaw

#174

Post by worldsayshi »

Netcob wrote: post #65

Am I missing something? Why is everyone talking about sandboxes when it comes to OpenClaw? To me it's like giving your dog a stack of important documents, then being worried he might eat them, so you put the dog in a crate, together with the documents. I thought the whole problem with that idea was that in order for the agent to be useful, you have to connect it to your calendar, your e-mail provider and other servic…

Why isn't users of openclaw "just" giving it its own identity? Give it its own mailbox, calendar and other accounts. Like an assistant.

Sure it takes away part of the point but only the part that is completely unhinged.

Re: Nvidia NemoClaw

#175

Post by esperent »

rcarmo wrote: post #172
observationist wrote: post #133

Earlier quoted context omitted.

What makes it even better is that these dogs are like Malinois. If they want to get into something, they will; people have had their entire network compromised by bots they left running overnight, and any important information like account logins and so on runs the risk of being misused. It's one thing to sandbox, maybe give the bot a temporary, limited $100 card or account to go perform a specific task, but there's…

I beg to differ. I took one, defanged it (well, I let it keep the claw in the name), and turned it into a damn useful self-modifiable IDE: https://github.com/rcarmo/piclaw Yes, it has cron and will do searches for me and checks on things and does indeed have credentials to manage VMs in my Proxmox homelab, but it won't go off the rails in the way you surmise because it has no agency other than replying to me (and onl…

Besides the web ui, what can it do that pi agent in a terminal can do?

Re: Nvidia NemoClaw

#176

Post by TheDong »

observationist wrote: post #133
Netcob wrote: post #65

Am I missing something? Why is everyone talking about sandboxes when it comes to OpenClaw? To me it's like giving your dog a stack of important documents, then being worried he might eat them, so you put the dog in a crate, together with the documents. I thought the whole problem with that idea was that in order for the agent to be useful, you have to connect it to your calendar, your e-mail provider and other servic…

What makes it even better is that these dogs are like Malinois. If they want to get into something, they will; people have had their entire network compromised by bots they left running overnight, and any important information like account logins and so on runs the risk of being misused. It's one thing to sandbox, maybe give the bot a temporary, limited $100 card or account to go perform a specific task, but there's…

> people have had their entire network compromised by bots they left running overnight

I'm curious if you have references to this happening with OpenClaw using one of the modern Opus/Sonnet 4.6 models.

Those models are a bit harder to fool, so I'm curious for specific examples of this happening so I can do a red-team on my claw. I've already tried all sorts of prompt injections against my claw (emails, github issues, telling it to browse pages I put a prompt injection in), and I haven't managed to fool it yet, so I'm curious for examples I can try to mimic, and to hopefully understand what combination of circumstances make it more risky

Re: Nvidia NemoClaw

#177

Post by worldsayshi »

angry_octet wrote: post #140
hector_vasquez wrote: post #109

Earlier quoted context omitted.

> Am I missing something? You are indeed missing a TON. A lot of Open Claw users don't give it everything. We give it specific access to a group of things it needs to do the things we want. If I want an agent to sit there 24/7 maximizing uptime of my service, I give it access to certain data, the GitHub repo with PR privileges, and maybe even permissions to restart the service. All of this has to be very thoughtful a…

The problem is boundary enforcement fatigue. People become lazy, creating tight permission scopes is tedious work. People will use an LLM to manage the scopes given to another LLM, and so on.

> creating tight permission scopes is tedious work

I have a feeling this kind of boundary configuration is the bread and butter of the current AI software landscape.

Once we figure out how to make this tedious work easier a lot of new use cases will get unlocked.

Re: Nvidia NemoClaw

#178

Post by dchichkov »

observationist wrote: post #133
Netcob wrote: post #65

Am I missing something? Why is everyone talking about sandboxes when it comes to OpenClaw? To me it's like giving your dog a stack of important documents, then being worried he might eat them, so you put the dog in a crate, together with the documents. I thought the whole problem with that idea was that in order for the agent to be useful, you have to connect it to your calendar, your e-mail provider and other servic…

What makes it even better is that these dogs are like Malinois. If they want to get into something, they will; people have had their entire network compromised by bots they left running overnight, and any important information like account logins and so on runs the risk of being misused. It's one thing to sandbox, maybe give the bot a temporary, limited $100 card or account to go perform a specific task, but there's…

[dead]

Re: Nvidia NemoClaw

#179

Post by Vexs »

post-it wrote: post #90
dolebirchwood wrote: post #85

Earlier quoted context omitted.

All of which can already be done programmatically without OpenClaw.

Not with a single prompt.

You could pretty reasonably vibe code that in a single prompt odds are.

Additionally, there are browser extensions that can do this- check on a timer, see if some page content is there, and then notify.

Post reply on HN