vishnukool

6 ms·

1. 31.▲
   Show HN: Open source setup to self-host OpenClaw instances on any host(github.com)

   7 points by vishnukool 8mo ago|9 comments
2. 32.▲
   by vishnukool 8mo ago

   I wanted to let a few friends get a hosted OpenClaw setup without each of them dealing with Docker, reverse proxy, SSL certs, etc. So I built a small toolkit that turns one Linux box into a multi-tenant hosting setup. Each person gets their
3. 33.▲
   by vishnukool 8mo ago

   Love it!
4. 34.▲
   by vishnukool 8mo ago

   Thanks!
5. 35.▲
   by vishnukool 8mo ago

   I think github stars is a reasonable signal, but it would be nice to augment it with a signal of usage of this skill on a sandbox to rank certain one's higher.
6. 36.▲
   by vishnukool 8mo ago

   Yeah, vercel does something similar with install data and ranks them higher based on it. I currently don't have such data but maybe down the line, we can introduce metrics around, quality signals from the users of this app.
7. 37.▲
   by vishnukool 8mo ago

   Well, we rely on github stars as a signal of quality on top of that, we don't consider exactness match since it can be pretty misleading. A better signal is that we vector embed the name and description of a skill and find proximity ma
8. 38.▲
   by vishnukool 8mo ago

   Thank you, let me know of any feedback once you try it out :)
9. 39.▲
   Show HN: YC Advisor – AI grounded in 434 YC essays, interviews, and lectures(agent37.com)

   2 points by vishnukool 9mo ago|0 comments
10. 40.▲
    by vishnukool 10mo ago

    So far the creators I've worked with are all non-technical. They really just wanted to solve a problem in their niche. In the process of building it for them I kind of naturally biased towards using skills + Claude Agent SDK, to solve
11. 41.▲
    by vishnukool 10mo ago

    Well, my hypothesis is, that this is a significant enough pain point. And in a way the distribution problem is amplified by the fact that currently there's no easy way to share Claude skills to your audience without a lot of setup fric
12. 42.▲
    Show HN: Agent37 – Monetize your Claude skills with shareable links(agent37.com)

    5 points by vishnukool 10mo ago|3 comments
13. 43.▲
    by vishnukool 5y ago

    Deep learning is expensive and even if you have AWS/GCP credits, you'll quickly run out of it. We built the tool to make training cheaper and to have these credits longer. SpotML is a command line tool that automatically manages M
14. 44.▲
    Show HN: SpotML – Pip Package for Managed ML Training on AWS Spot Instances(docs.spotml.io)

    3 points by vishnukool 5y ago|1 comments
15. 45.▲
    by vishnukool 5y ago

    SpotML is a command line tool that automatically manages ML training on AWS spot instances. It lets you handle spot interruptions by resuming training using the latest checkpoint. Documentation link to try it out: https://docs.sp
16. 46.▲
    Show HN: Seamless ML training on AWS Spot instances(docs.spotml.io)

    6 points by vishnukool 5y ago|3 comments
17. 47.▲
    by vishnukool 5y ago

    Haven't used cortex.dev, but looking at the docs, I'd say primarily simplicity and ease of getting started quickly (<3 mins) to get it up and running. Also with cortex, It's not clear to me yet if spot instance fails, cort
18. 48.▲
    by vishnukool 5y ago

    We'll be updating the documentation with examples in a few weeks, once we release it. No, it doesn't support distributed training right now.
19. 49.▲
    by vishnukool 5y ago

    Thanks, Makes sense. It doesn't use any "code" from nimbo. The documentation and the design simplicity of the tool were the things that was appealing to us and adopted. The project itself was forked from spotty which has an
20. 50.▲
    by vishnukool 5y ago

    It uses a mounted EBS(Elastic Block Store) so all the checkpoints, data etc. is already in the persistence storage. This is simply be re-attached to the next spot/onDemand instance after interruption.
21. 51.▲
    by vishnukool 5y ago

    Interesting, thanks, we weren't aware of Metaflow. I've read through the docs, the one difference that comes to my mind is the automatic fallback to on-Demand and resume back to spot when available. I can't readily see a way
22. 52.▲
    by vishnukool 5y ago

    Thanks for the suggestion, yes makes sense.
23. 53.▲
    by vishnukool 5y ago

    Thanks for pointing it out, We realize our mistake here. We also should've done proper attribution. Will be correcting this.
24. 54.▲
    by vishnukool 5y ago

    Hmm.. makes sense, yes. thanks!
25. 55.▲
    by vishnukool 5y ago

    Thanks for the feedback, yeah at this stage we really just wanted to find out if we can build something useful for the community. Agreed on the pricing suggestion. Also, interesting point about inference. I'm not sure though how common
26. 56.▲
    by vishnukool 5y ago

    Curious, did you eventually start using Sagemaker with spot instances or did you give up on it ? Also what would you say were the biggest pain points with Sagemaker?
27. 57.▲
    by vishnukool 5y ago

    Thank you!
28. 58.▲
    by vishnukool 5y ago

    Thanks, yeah. At this stage we really just want to validate if this was a real problem in the ML community. Down the line I suppose as we scale to handle multi instance training and other use cases, we could probably charge more, say a %
29. 59.▲
    by vishnukool 5y ago

    Good question, I've heard other engineers ask for this. The MVP version doesn't yet handle multiple instances, but it will be on our roadmap.
30. 60.▲
    by vishnukool 5y ago

    That's right it will need aws credentials with access to create EBS volumes, S3 Bucket, spawn instances, etc. In addition to the "cli" a cloud service constantly monitors the progress of the Jobs(by registering the pid when l