qrdlgit

Yes, the entire engineering world is holding their breath :) SWE-bench is probably the best eval out there right now, and this is one of the best ways to evaluate...

for the sanity of everyone everywhere please merge this

buy side security shouldn't be a priority beyond spoofing. sell side security is really all that's important (beyond identity theft). You can't fix the problem of vendor trust with this...

I get the sense people don't program agents here, or practice any reasonable sense of infosec. Clicking on an untrusted link is crazy. Much much worse though, is using an...

I use brave extensively and exclusively for agentic search and I think you folks are poised to take over the world, though you need to get massive investment and very...

I do a lot of agentic stuff. 90% of my agentic transactions are < 0.01c. Most, if not all, of my transaction have an unknown up front cost and in...

> > I do a lot of agentic stuff. > > 90% of my agentic transactions are < 0.01c. Most, if not all, of my transaction have an unknown up...

> Ah, I see. Paying for LLM calls rather than allowing LLMs take responsibility for arbitrary transactions. No, that's inaccurate. LLM based tool calls can be quite arbitrary, depending on...

Here is another good paper leveraging multiple agents in a single flow https://arxiv.org/pdf/2510.26658 (though async + parallel is much better, as the output of a parallel process can asynchronously join...

Another example is cache hit requests, which are near free. For example Brave has added requirements for people who store data which forces them to make re-requests. Optimizing for cache...