nravic

I think even easier in fact - what's happening behind the scenes w/ an LLM is far more opaque

This is super interesting! We do something similar I think by taking a checkpoint after model initialization. I'm curious what you think about our approach, here's some benchmarks: https://docs.cedana.ai/articles/performance-of-cedanas-gpu-i...

We do some on-the-fly optimizations as well (like compiling into CUDA graphs or fusing together calls) which ends up resulting (for some inference engines) faster token throughput too.

How were you handling GPU state w/ pytorch? We added some custom code around CRIU to enable GPU checkpointing fwiw: https://docs.cedana.ai/setup/gpu-checkpointing/

We do this using CRIU right now! https://github.com/cedana/cedana

In fact one of our customer's use cases is exactly what you describe, allowing users to "hibernate" container workspaces.

I'm probably the cofounder of the guy you spoke with! Here's our repo: https://github.com/cedana/cedana

what kind of pannier bags do you use to haul groceries? All the ones I've used have been too small for my purposes.

IIRC (and in my experience) KL divergence doesn't account for double counting. Wrote a paper where I ended up having to use a custom metric instead: https://digitalcommons.usu.edu/cgi/viewcontent.cgi?article=4...

This is so cool! Out of curiosity, could you walk through your decisionmaking around using Nix and how it fits in with Rivet?

how often are these batch jobs run? I'm curious to know what the absolute maximum sync frequency can be.

Self plug: run llama.cpp as an inference server on a spot instance anywhere: https://cedana.readthedocs.io/en/latest/examples.html#runnin...

GrapheneOS uses WebUSB to flash your phone.

Pretty wild first experience doing it for me - I'm used to the hassle of sideloading a bootloader and then flashing.

Hey! I'd love to learn more about the position - the hook @ jpl email on your website is bouncing though. Do you mind sending me contact details so we can chat?

Thanks :)

Do people out of college even do Leetcode anymore? Surely there's a better use of time once you're in a paying job.

If you're looking at aerospace/defense startups, this is usually the case.

It's not unheard of for them to employ people who's primary role is grant writing to try and get (for example) SBIR funding

Zotero is my go-to

Julia's REPL is beautiful

Yeah, spiking neural networks are a tough nut to crack. Check out Nengo if you're interested in learning more

What are Protonmail's problems?

very visceral reaction there

I've followed the same path, out of curiosity do you mind if I PM you asking a few questions? I'm a couple months out of undergrad and I'd like to pick your brain a bit.

This is really cool, what are you training?

The population argument is fundamentally flawed. It places the burden of climate change onto the global south, which is far from the actual truth

yes

pleasantly surprised to see julia used in this

There's a bunch of Eigen analogues in Rust, which is slightly frustrating. ndarray is pretty great though

looks like that's a "fuck you"

Hello, I emailed you a while ago and I got a response back regarding a phone call; however nothing after that. Would you like me to follow up on that?

I sent you an email two months ago, and asked about a follow up last month (after which I sent an email). So here's another follow up (with which I'll also send another email) in the hopes that you see it. Would love to talk to you about the kind of work you do, specifically with ML and how you use actual clinical data

are all your positions subject to ITAR restrictions?

Hello,

Are all your roles subject to ITAR restrictions?