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

Claws are now a new layer on top of LLM agents

twitter.com

Post reply on HN
Print view View original

781–790 of 1001 posts

Re: Claws are now a new layer on top of LLM agents

#781

Post by dracyr »

jrvarela56 wrote: post #765
adrianN wrote: post #712

Earlier quoted context omitted.

One prompt injection away from sending all your credentials to the Internet?

Agree, i was going the vaultwarden route and figured this pattern seems better: https://fly.io/blog/tokenized-tokens/ Secrets are encrypted and the proxy decrypts on the fly if destination is whitelisted for that token.

Reading through the discussion I was also thinking of the other fly.io blog post around their setup with macaroon tokens and being able to quite easily reduce the blast radius of them by adding more caveats. Feels like you could build out some kind of capability system with that that might mitigate some risks somewhat.

Re: Claws are now a new layer on top of LLM agents

#782

Post by faeyanpiraat »

aix1 wrote: post #744
ZitchDog wrote: post #484

Earlier quoted context omitted.

I've created my own "claw" running in fly.io with a pattern that seems to work well. I have MCP tools for actions that I want to ensure human-in-the loop - email sending, slack message sending, etc. I call these "activities". The only way for my claw to execute these commands is to create an activity which generates a link with the summary of the acitvity for me to approve.

Is there a risk that the summary doesn't fully match the action that actually gets executed?

Side note: Just like with a human employee asking for permission to do something.

Re: Claws are now a new layer on top of LLM agents

#784

Post by irthomasthomas »

colbyn wrote: post #720
dang wrote: post #714

Earlier quoted context omitted.

The personal attacks I saw were against different people, not just one. In a lot of cases it's just routine internet cynicism, which is always amplified against unusually successful or prominent people. There's also a lot of fear and anger about the AI tsunami these days, among certain user cohorts, and that's an amplifier as well. On HN, personal attacks aren't allowed regardless of who's being attacked, and comment…

Okay I see. Are people being attacked for engaging in AI research/dev irrespective of their character or other personal attributes?

I saw simonw getting attacked for sharing his bloglink about it, only it was not even simonw who shared the link here.

Re: Claws are now a new layer on top of LLM agents

#785

Post by paganel »

irthomasthomas wrote: post #784
colbyn wrote: post #720

Earlier quoted context omitted.

Okay I see. Are people being attacked for engaging in AI research/dev irrespective of their character or other personal attributes?

I saw simonw getting attacked for sharing his bloglink about it, only it was not even simonw who shared the link here.

Saw that, too, but at some point one cannot just stand like sheep in the slaughterhouse, the reaction was to be expected (even though it could have happened in a more civilized way, not via personal-ish attacks, I agree with that).

More generally, there are now literally trillions of dollars being invested in this madness/tsunami/whatever-one-wants-to-call-it, which means that it has now become impossible to follow said money so as to follow the conflicts of interests (it’s easy to assume a conflict of interest for a guy like Karpathy given his past and recent employment history, but I do think that Simon is more on the genuine side), so this is why that counter-reaction is now manifesting itself so chaotically, hitting left and right with not necessarily any logic behind it, which means that there are going to be collateral “casualties” during it all (such as Simon in this case).

Re: Claws are now a new layer on top of LLM agents

#786

Post by flimflamm »

I just don't trust "the claw" so I build following system - Docker 1: * Locked up Claw docker - user level priv. Access outside to "one port" only.

- Docker 2: * Tool gateway with pre-baked commands - openclaw can only index what command to execute * Keys are here * Telegram hook to approve all "post" commands i.e. sending email or posting something somewhere.

- Docker 3: * LLM gateway keeping track of cost and routing

Re: Claws are now a new layer on top of LLM agents

#787

Post by BoredPositron »

irthomasthomas wrote: post #784
colbyn wrote: post #720

Earlier quoted context omitted.

Okay I see. Are people being attacked for engaging in AI research/dev irrespective of their character or other personal attributes?

I saw simonw getting attacked for sharing his bloglink about it, only it was not even simonw who shared the link here.

I can understand the sentiment against Simon it's just to much of the same content over and over again but I handled it with just blacklisting him no need for personal attacks.

Re: Claws are now a new layer on top of LLM agents

#788

Post by lwhi »

flimflamm wrote: post #786

I just don't trust "the claw" so I build following system - Docker 1: * Locked up Claw docker - user level priv. Access outside to "one port" only. - Docker 2: * Tool gateway with pre-baked commands - openclaw can only index what command to execute * Keys are here * Telegram hook to approve all "post" commands i.e. sending email or posting something somewhere. - Docker 3: * LLM gateway keeping track of cost and routi…

So you need to approve all actions that actually do something, individually?

Re: Claws are now a new layer on top of LLM agents

#790

Post by flimflamm »

deadbabe wrote: post #475

Instead of posts about claws I would like to see more examples of what people are actually doing with claws. Why are you giving it access to your bank account? Even if I had a perfectly working assistant right now, I don’t even know what I would ask it to do. Read me the latest hackernews headlines and comments?

If you don't have workflows which repeat in inet you don't need openClaw. - Messages from school where to react - Getting payments from someone and tracking that you get them - Summary of news the way you like it from sources you like it every day - Integrated task lists reminders - Drafting taxation reports based on spending etc etc.
Post reply on HN