theaniketmaurya

7 ms·

31.
Show HN: Run coding agents in a sandbox locally(github.com)
2 points by theaniketmaurya 6mo ago|1 comments
32.
by theaniketmaurya 6mo ago
i was using neon and they had some similar feature but now using planetscale. would be curious to know how you all are doing it?
33.
by theaniketmaurya 6mo ago
I've been dogfooding it by running coding agents. I mount my local filesystem and pull the changes back when done.
34.
Open-source isolated runtime for AI agents(github.com)
4 points by theaniketmaurya 6mo ago|3 comments
35.
by theaniketmaurya 6mo ago
I've been working on an open-source isolated runtime for AI agents. Over a period of time I was able to bring the cold start time in sub-seconds, persistent filesystem, and network control. I'm looking for feedback and if you&#x27
36.
I rewrote network setup for sandboxes in Rust and it sped up by 57x(github.com)
2 points by theaniketmaurya 6mo ago|0 comments
37.
by theaniketmaurya 6mo ago
yes, we have an example of OpenClaw. SmolVM uses microVMs which means the resource consumption is quite minimal. For now, you can specify the amount of RAM while starting the VM. We are making it elastic so you can configure to expand the
38.
Show HN: SmolVM – open-source sandbox for coding and computer-use agents(github.com)
8 points by theaniketmaurya 6mo ago|3 comments
39.
Open-source sandboxes to run AI code, browser agents and computer-use(github.com)
2 points by theaniketmaurya 7mo ago|0 comments
40.
by theaniketmaurya 8mo ago
Also, I've been running OpenClaw using SmolVM on my macOS. It adds a hardware level isolation (using KVM). https://github.com/CelestoAI/SmolVM/blob/main/examples/openc...
41.
An attempt to make MicroVMs more accessible (SmolVM – Python SDK)(github.com)
1 points by theaniketmaurya 8mo ago|1 comments
42.
Show HN: Built an open-source SDK to simplify tool authentication for AI Agents(github.com)
1 points by theaniketmaurya 1y ago|0 comments
43.
by theaniketmaurya 1y ago
let me put some more in nvidia now
44.
by theaniketmaurya 1y ago
Also found this Vibe coding OSS from Modal - https://github.com/modal-labs/modal-vibe Curious if Vercel is also gonna open-source theirs.
45.
by theaniketmaurya 1y ago
Nice! they forgot to add the Github link for the repo lol. PS: If you're looking too - https://github.com/cloudflare/vibesdk
46.
by theaniketmaurya 1y ago
I’m using Neon too and upgraded to the scale up version today. Curious, what do you mean rhat they can get pricey?
47.
by theaniketmaurya 2y ago
You can run with LitServe. here is the code - https://lightning.ai/lightning-ai/studios/deploy-llama-3-2-v...
48.
by theaniketmaurya 2y ago
You can run it with LitServe (MPS GPU), here is the code - https://lightning.ai/lightning-ai/studios/deploy-llama-3-2-v...
49.
A full-stack Python model serving library(github.com)
8 points by theaniketmaurya 2y ago|0 comments
50.
LLM training in simple, pure C/CUDA(twitter.com)
4 points by theaniketmaurya 3y ago|1 comments
51.
by theaniketmaurya 3y ago
LLM training in simple, pure C/CUDA. There is no need for 245MB of PyTorch or 107MB of cPython.
52.
by theaniketmaurya 3y ago
great to see this project coming up so well. I had a chat with Adrey years ago. Congratulations!!
53.
by theaniketmaurya 3y ago
Lightning AI is hiring for Sales Development Representatives in NYC and Palo alto https://grnh.se/aaae3f6c3us
54.
Where do LLMs spend their FLOPS?(artfintel.com)
3 points by theaniketmaurya 3y ago|0 comments
55.
by theaniketmaurya 3y ago
Proud to see this work built using Lit-GPT coming through.
56.
Scaling Large (Language) Model Training with PyTorch Lightning(lightning.ai)
2 points by theaniketmaurya 3y ago|0 comments
57.
OpenLLaMA – comparable performance and better in several task to original LLaMA(twitter.com)
2 points by theaniketmaurya 3y ago|0 comments
58.
by theaniketmaurya 3y ago
I would say similar since the building block is the transformer for both. In this blog post, the fine-tuning strategy used is Adapter. It basically adds a learnable layer to the Transformer block.
59.
by theaniketmaurya 4y ago
Yes, agree that the weights aren't commercially licensed (yet)! The other ways to run LLaMA are using GPL license which makes it difficult for commercial use even if someone trains and upload the weights publicly. This could be a step
60.
by theaniketmaurya 4y ago
GPL is a copyleft license which requires you to share anything that you build using the original software. This makes it difficult for commercial use.

More ›