Advertisement · 728 × 90

Posts by Alok Rajiv

Daniel Lowd (@lowd.bsky.social) Since this platform is finally attracting a critical mass of ML researchers, here's our recent work on prompt-based vulnerabilities of coding assistants: arxiv.org/abs/2407.11072 TL;DR — An attacker can convince your favorite LLM to suggest vulnerable code with just a minor change to the prompt! https://arxiv.org/abs/2407.11072

bsky.app/profile/lowd.bsky.social...

I think it makes sense given the open source philosophy.

1 year ago 0 0 0 0

I'm just not used to a Social Media platform with this level of transparency. Open Source protocol and apps with free choice on the Algorithm. Haha..I think this is going to be a fun journey!

1 year ago 1 0 1 0

I think they are leaning into the fact that, they are still "by popular opinion" the best at code. And, so they want to make it jump to showing off.

(Now that I think about it) Sounds like a lot of humans I know :D !

1 year ago 1 0 0 0

(DOMAIN: SUPPLY CHAIN) Was super confused when at first, I heard this being thrown around.

So, after half a decade, my current (unsure) definition aligns as: a realtime data warehouse that predicts downstream impacts of events.

So in short, a glorified simulator with realtime operational feeds.

1 year ago 0 0 0 0

Thanks for that! You sent me down a (good) rabbithole for a bit!

I am not really sure where the ranking code is. The basic one seems to be just chronology based is it?

github.com/bluesky-soci...

1 year ago 0 0 0 0

Honestly, love the un-hype! This should be a trend!

Helps me appreciate the paper even more!

1 year ago 1 0 0 0

What I love more now - is introducing it to people, who haven't and then checking in - a month after! It just becomes part of their flow - so seamlessly!

1 year ago 1 0 0 0
Advertisement