vishnukool

6 ms·

1.
by vishnukool 2mo ago
I built basically this on top of my telegram setup last year. The thing that killed it was group chats, per-message LLM calls on a busy group eat your api budget alive, so I ended up putting a tiny local model in front as a first pass.
2.
by vishnukool 2mo ago
Such a rage bait title, but sure I'll take the bait. It is easy to phrase anything as slaves. Are refrigerators and microwave ovens our slaves? But if you push the tech fare enough.. there is a boundary where we do cross the slave thre
3.
by vishnukool 4mo ago
The problem is that LLMs do not have a conceptual grounding in actual time. They estimate based on statistical correlation found in their training data which is filled with standard corporate project management timelines legacy codebases an
4.
by vishnukool 4mo ago
Good one. One thing that's becoming clear is that agent security is less about jail break prompts and more about permission boundries, hidden context flow and un intended tol behavior.
5.
Ask HN: Are all the YC founders lying offlate?
3 points by vishnukool 4mo ago|0 comments
6.
by vishnukool 4mo ago
System design definitely carries more weight now but Leetcode has not totally vanished. It's just evolved.
7.
by vishnukool 4mo ago
Respect. It's a good day to remember the impact they had even if they are no longer around to pick up the phone. Btw calling her now.
8.
by vishnukool 4mo ago
This is the weird part of AI hallucinations and people start treating it like a verified fact after it gets repeated.
9.
by vishnukool 7mo ago
What have you seen to be the bottleneck ? From the usage patterns I've seen for 90% of the use cases it seems to be fine, the one exception being Browser based use cases which chugs memory and so needs a beefier machine.
10.
by vishnukool 7mo ago
Very true and actually totally spot on! It's the damn chromium instances that chugs memory like an Irish man chugs beer. I experimented launching containers with it and quickly realized the tiny shared infra hosts are not a good fit. W
11.
by vishnukool 7mo ago
Yes exactly! Although we've launched a Claude Skill to help setup OpenClaw! So Agent 37 is literally the easiest Openclaw setup at this point :)
12.
by vishnukool 7mo ago
Absolutley! We do have some metrics like, when a user interacts with a skill or not, and so that's a signal
13.
by vishnukool 7mo ago
Yes, once we have a large enough numbers to justify that we'll factor that into the search results quality
14.
by vishnukool 7mo ago
Ah that's a good point, I think I could introduce some kind of time decay by also factoring in last updated repo time into the search priority.
15.
by vishnukool 7mo ago
Great point, we de-duplicate the search results by Repository Id. that largely filters out content from the same person , and note that search prioritizes github stars, so low quality one's get pushed to the bottom
16.
by vishnukool 7mo ago
Thank you! We have limits around usage. Also as a startup we have a lot of Gemini credits from Google, so that helps in offering it. Not to mention the gemini 3 flash that it's running on is dirt cheap. Regarding low quality skills, o
17.
by vishnukool 7mo ago
Yup we actually backup data in a separate partition per container. So it survives restarts and even accidental deletions.
18.
by vishnukool 7mo ago
Thanks! At the moment no, but down the line I imagine we can expose it if this is useful
19.
by vishnukool 7mo ago
We do back up data, although it's manual. The docker instance home directory and homebrew are backed up in a separate partition. So it survices restarts and container updates.
20.
by vishnukool 7mo ago
SO far pretty good! Turns out these things are mostly idle
21.
by vishnukool 7mo ago
We only allow ttyd not ssh, so users only have access to their individual containers and isolated from the host. But inside the container they can do whatever they'd want. There are also caps on CPUs and resource and PID limits so as
22.
by vishnukool 7mo ago
True but Unfortunately, it's expensive to run VMs that are stateful and ongoing. In a way the 99 cents is just gate keeping really interested people vs others.
23.
by vishnukool 7mo ago
thanks!
24.
by vishnukool 7mo ago
Yup on the roadmap!
25.
by vishnukool 7mo ago
Haha thanks ! Yeah i really pushed the limits on this one :)
26.
by vishnukool 7mo ago
Sorry that was only for the early adopters, the current price is at 3.99 so we can have some margin.
27.
by vishnukool 7mo ago
The 4GB is just from my personal profiling and using OpenClaw
28.
by vishnukool 7mo ago
At the moment it's more of a demo for hobby users setting up for themselves and friends, but if it's an enterprise use case where you need more control, we have https://www.agent37.com/openclaw offering that works
29.
by vishnukool 7mo ago
Yeah the auth here is simple intentionally. This is just an example of what's possible, but you could extend the node server to do forward auth or anything else you have in mind
30.
by vishnukool 7mo ago
That would basically be as simple as updating the Docker image and entry point file and then restarting the container.

More ›