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

181–190 of 291 posts

Re: Nvidia NemoClaw

#181

Post by _--__--__ »

mjr00 wrote: post #86
chill_ai_guy wrote: post #74

Earlier quoted context omitted.

Lol you think these github repos just materialize as is? They probably did all the iteration and development internally and then ported it over to a github repo and made it public afterwards

No they didn't. You can see all the commits as this was built iteratively[0]. This project started development on Saturday morning and now it's here. This is pretty common now, people love to rapidly throw together stuff and show it off a few days later. The only thing different about this from your average Show HN sloppa is that it's living under the NVIDIA Github org, though that also has 700+ repositories[1] in it…

There was some kind of public knowledge of this project over a week ago because people were trying to domain squat them and submit it to HN: https://news.ycombinator.com/from?site=nemoclaw.bot

Re: Nvidia NemoClaw

#182

Post by yibers »

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…

We are in the middle of a gold rush. Nvidia makes the shovels.

Re: Nvidia NemoClaw

#183

Post by ljm »

noosphr wrote: post #166
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…

The first well known example of long running agents taking to each other was shilling a goatse based crytpo: > Truth Terminal had become obsessed with the Goatse meme after being put inside the Claude Backrooms server with two Claude 3 chatbots that imagined a Goatse religion, inspiring Truth Terminal to spread Goatse memes. After an X user shared their newly created GOAT coin, Truth Terminal promoted it and pumped t…

If Infinite Jest was real I think this would be it, human and AI alike rendered catatonic by an abyssal rectum

Re: Nvidia NemoClaw

#184

Post by Nolxy »

hmokiguess wrote: post #84
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…

Yes, although what I think is different in this setup here is the OpenShell gateway override, as they mention: > NemoClaw installs the NVIDIA OpenShell runtime and Nemotron models, then uses a versioned blueprint to create a sandboxed environment where every network request, file access, and inference call is governed by declarative policy. The nemoclaw CLI orchestrates the full stack: OpenShell gateway, sandbox, inf…

Right, the gateway layer is the genuinely interesting part. Intercepting every outbound network call before it leaves the sandbox gives you a real enforcement surface, not just "trust the app to behave". The problem is the threat model is still inverted for the security critics in this thread: the agent is the client, so the dangerous calls are the ones going out to your authenticated services (Gmail, Slack, whatever), and a gateway that filters those is only as good as your policy definitions. One misconfigured rule and ure back to square one. The GPU rental angle makes total sense too. This is basically Nvidia saying "don't buy Mac Mini, rent ours" wrapped in enough infrastructure glue to make it feel like a platform.

Re: Nvidia NemoClaw

#185

Post by mjr00 »

elif wrote: post #163
mjr00 wrote: post #105

Earlier quoted context omitted.

> Sorry to be the one to inform you that we edit history in git. Who is "we"? Do you work for NVidia? > There has been reporting on nemoclaw for the last couple weeks. The earliest reporting I've seen was yesterday. Can you link something from prior to March 14? edit: I did find some articles from before March 14[0] which says NVidia was "prepping" this. Which is extremely funny, because it means they were hyping up…

alright so the git history goes back 4 days. I learned about nemoclaw 5 days ago here: https://www.youtube.com/watch?v=fL2lMpLjxWA but it was reported 8 days ago here: https://www.youtube.com/watch?v=345GsxnrHHg I am not anyone special. I don't know anything about nvidia. I just know that the "4 day history" you think matters, is not a reasonable belief given that random youtubers have been reporting on it. and by "w…

There's nothing clean about the history. You think commits like [0], with the commit message "improve", count as "clean"? What do you think the motivation for the author would be to modify git history to make it appear that this was written over a weekend, including separating each feature/commit by a few hours, which corresponds to a reasonable amount of time that it may have taken to write that feature? Including a break on Mar 15 at 1:18 AM PDT before continuing to commit at Mar 15 at 12:43 PM PDT. Hey, isn't there a normal human behaviour that occurs around this time every day which takes 6-10 hours?

I'm fully aware you can rewrite git history to whatever you want, but this is an occam's razor situation here. You'd only think this wasn't a weekend project if you desperately wanted to believe that this was some major initiative for some reason.

[0] https://github.com/NVIDIA/NemoClaw/commit/b9382d27d13b160dcf...

Re: Nvidia NemoClaw

#186

Post by ex-aws-dude »

"NVIDIA NemoClaw installs the NVIDIA OpenShell runtime, part of NVIDIA Agent Toolkit, for inference through NVIDIA cloud."

After that I eat an NVIDIA sandwich from my NVIDIA fridge and drive my NVIDIA car to the NVIDIA store NVIDIA NVIDIA NVIDIA

Re: Nvidia NemoClaw

#188

Post by aavci »

How does this compare the building your own bot that has access to these tools: - web plugin - api access to messaging - access to a job scheduler

Re: Nvidia NemoClaw

#189

Post by macNchz »

TheDong wrote: post #176
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…

> 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,…

No maliciousness or injection required, even the newest and most resistant models can start doing weird stuff on their own, particularly when they encounter something failing that they want to work.

Just today I had Opus 4.6 in Claude Code run into a login screen while building and testing a web app via Playwright MCP. When the login popped up (in a self-contained Chromium instance) I tried to just log in myself with my local dev creds so Claude would have access, but they didn't work. When I flipped back to the terminal, it turned out Claude had run code to query superadmin users in the database, picked the first one, and changed the password to `password123` so it could log in on its own.

This was a sandboxed local dev environment, so it was not a big deal (and the only reason I was letting it run code like that without approval), but it was a good reminder to be careful with these things.

Re: Nvidia NemoClaw

#190

Post by chrishare »

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…

Yeah, but atleast the dog is going to eat your documents only, and not crap on your rug
Post reply on HN